1
Object Relative Clauses in Archaic Chinese
Edith Aldridge University of Washington
DRAFT September 22 2011
1 Introduction
This paper examines headless object relative clauses in late Archaic Chinese of the 5th to
3rd centuries BCE1 and proposes that they are reduced relative clauses with genitive
subjects Accordingly they should be analyzed as mixed projections in the sense of
Abney (1987) Borsley and Kornfilt (2000) Baker (2005 2011) Grohmann and
Panagiotidis (2009) and others I argue specifically that they are structurally composed
of a vP dominated by a nominal layer similar to approaches put forth by Krause (2001)
Miyagawa (2008 2011) Aygen (2011) and others
Late Archaic Chinese employed different strategies for relativizing on subject
position as opposed to VP-internal positions In the former type the functional
morpheme zhe appears following the clause as in (1a) When the gap is VP-internal the
functional morpheme suo appears to the left of the VP
1 I follow Wang (2004) and Zhou (1963) in identifying the Warring States period of the 5th ndash 3rd centuries
BCE as a distinct period in the history of Chinese which they term Late Archaic Chinese I also agree with
them that the language of the few hundred years following this period exhibits its own characteristics They
label this period Pre-Medieval Chinese I refer to this period roughly corresponding to the Han period (2nd
century BCE ndash 2nd century CE) as Early Middle Chinese
2
(1) a 欲 戰 者 (Zuozhuan Cheng 6)
yu zhan zhe
desire fight ZHE
lsquo(those) who desire to fightrsquo
b 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
Aldridge (2009) argues that ZHE is a determiner based in part on the fact that ZHE
could also select an NP complement as in (2) Additionally the fact that the ZHE
constituent in (2) occurs with a demonstrative suggests that ZHE is probably not the head
of DP but a lower functional head
(2) a 夫 三 子 者 之 言 何 如
[Fu [san zi zhe]] zhi yan he ru
DEM 3 gentlman ZHE GEN word what like
lsquoHow what those three gentlemen saidrsquo (Analects Xianjin)
3
b 夫 二 人 者 魯國 社稷 之 臣 也
[Fu [er ren zhe]] Luguo sheji zhi chen ye
DEM two person ZHE Lu nation GEN minister COP
lsquoThese two men (will become) ministers of Lu at the national levelrsquo
(Zuozhuan Cheng 16)
(3) clearly shows that ZHE occupies a position below D since the constituent
projected by ZHE can be possessed
(3) 庾公之斯 衛子 之 善 射 者 也 (Mencius 8)
Yugongzhisi [Weizi zhi [shan she zhe]] ye
Yugongzhisi Weizi GEN well shoot ZHE COP
lsquoYugongzhisi is a skilled archer of Weirsquo
Aldridge proposes that ZHE is n a determiner on the DP spine but structurally lower
than the D position itself She follows Williamson (1987) Kayne (1994) and others in
assuming that a determiner can take a relative clause2 as its complement She further
assumes with Basilico (1996) that the determiner can serve as the binder for the gap to
form the relative clause The head position in the relative clause is limited to VP-external
2 Aldridge (2009241 fn 2) assumes that the relative clause is a TP and not a CP This is primarily because
Archaic Chinese lacked finite embedded clauses these all show evidence of nominalization On this
analysis then the phase head for the embedded clause is D rather than C
4
positions by the Phase Impenetrability Condition of Chomsky (2004) since the VP is no
longer accessible to the computational system by the time ZHE is merged
(4) DP D nP
TP ZHEi proi hellip
In this paper I propose that SUO forms relative clauses on VP-internal positions by
acting as the binder for an operator in the edge of vP This restricts SUO relativization to
VP-internal positions since the subject is licensed outside of vP and is not c-commanded
by SUO This leads to the question of why the operator resides in the edge of vP rather
than CP I propose that this is because there is no CP layer in these relative clauses The
TP headed by SUO and dominating vP is nominalized The subject moves to the specifier
of this nominalized T However since this T is not finite and therefore unable to license
nominative case the subject must value genitive case with D SUO relatives are thus
reduced relative clauses with genitive subjects in the sense of Krause (2001) Miyagawa
(2008 2011) Aygen (2011) and others
5
(5) DP D[Gen] TP
DP[Gen] Trsquo TNom vP SUOi OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
In this paper I argue in turn for each component of the analysis in (5) I begin by
showing that SUO occupies a clause-medial position and is not associated with the CP
layer I next offer indirect evidence for movement of the operator to [Spec vP] and
confirm that SUO is not this operator but rather must be analyzed as occupying a position
external to vP Finally I argue for the reduced nature of the clause by showing that the
subject receives genitive case However this case is valued under Agree with D the
subject remains in the specifier of the nominalized T and does not move to [Spec DP]
2 Position of SUO
This section argues that SUO occupies a clause-medial position between the genitive
subject and vP Crucially SUO is not associated with the CP layer or other TP-external
position
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
2
(1) a 欲 戰 者 (Zuozhuan Cheng 6)
yu zhan zhe
desire fight ZHE
lsquo(those) who desire to fightrsquo
b 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
Aldridge (2009) argues that ZHE is a determiner based in part on the fact that ZHE
could also select an NP complement as in (2) Additionally the fact that the ZHE
constituent in (2) occurs with a demonstrative suggests that ZHE is probably not the head
of DP but a lower functional head
(2) a 夫 三 子 者 之 言 何 如
[Fu [san zi zhe]] zhi yan he ru
DEM 3 gentlman ZHE GEN word what like
lsquoHow what those three gentlemen saidrsquo (Analects Xianjin)
3
b 夫 二 人 者 魯國 社稷 之 臣 也
[Fu [er ren zhe]] Luguo sheji zhi chen ye
DEM two person ZHE Lu nation GEN minister COP
lsquoThese two men (will become) ministers of Lu at the national levelrsquo
(Zuozhuan Cheng 16)
(3) clearly shows that ZHE occupies a position below D since the constituent
projected by ZHE can be possessed
(3) 庾公之斯 衛子 之 善 射 者 也 (Mencius 8)
Yugongzhisi [Weizi zhi [shan she zhe]] ye
Yugongzhisi Weizi GEN well shoot ZHE COP
lsquoYugongzhisi is a skilled archer of Weirsquo
Aldridge proposes that ZHE is n a determiner on the DP spine but structurally lower
than the D position itself She follows Williamson (1987) Kayne (1994) and others in
assuming that a determiner can take a relative clause2 as its complement She further
assumes with Basilico (1996) that the determiner can serve as the binder for the gap to
form the relative clause The head position in the relative clause is limited to VP-external
2 Aldridge (2009241 fn 2) assumes that the relative clause is a TP and not a CP This is primarily because
Archaic Chinese lacked finite embedded clauses these all show evidence of nominalization On this
analysis then the phase head for the embedded clause is D rather than C
4
positions by the Phase Impenetrability Condition of Chomsky (2004) since the VP is no
longer accessible to the computational system by the time ZHE is merged
(4) DP D nP
TP ZHEi proi hellip
In this paper I propose that SUO forms relative clauses on VP-internal positions by
acting as the binder for an operator in the edge of vP This restricts SUO relativization to
VP-internal positions since the subject is licensed outside of vP and is not c-commanded
by SUO This leads to the question of why the operator resides in the edge of vP rather
than CP I propose that this is because there is no CP layer in these relative clauses The
TP headed by SUO and dominating vP is nominalized The subject moves to the specifier
of this nominalized T However since this T is not finite and therefore unable to license
nominative case the subject must value genitive case with D SUO relatives are thus
reduced relative clauses with genitive subjects in the sense of Krause (2001) Miyagawa
(2008 2011) Aygen (2011) and others
5
(5) DP D[Gen] TP
DP[Gen] Trsquo TNom vP SUOi OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
In this paper I argue in turn for each component of the analysis in (5) I begin by
showing that SUO occupies a clause-medial position and is not associated with the CP
layer I next offer indirect evidence for movement of the operator to [Spec vP] and
confirm that SUO is not this operator but rather must be analyzed as occupying a position
external to vP Finally I argue for the reduced nature of the clause by showing that the
subject receives genitive case However this case is valued under Agree with D the
subject remains in the specifier of the nominalized T and does not move to [Spec DP]
2 Position of SUO
This section argues that SUO occupies a clause-medial position between the genitive
subject and vP Crucially SUO is not associated with the CP layer or other TP-external
position
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
3
b 夫 二 人 者 魯國 社稷 之 臣 也
[Fu [er ren zhe]] Luguo sheji zhi chen ye
DEM two person ZHE Lu nation GEN minister COP
lsquoThese two men (will become) ministers of Lu at the national levelrsquo
(Zuozhuan Cheng 16)
(3) clearly shows that ZHE occupies a position below D since the constituent
projected by ZHE can be possessed
(3) 庾公之斯 衛子 之 善 射 者 也 (Mencius 8)
Yugongzhisi [Weizi zhi [shan she zhe]] ye
Yugongzhisi Weizi GEN well shoot ZHE COP
lsquoYugongzhisi is a skilled archer of Weirsquo
Aldridge proposes that ZHE is n a determiner on the DP spine but structurally lower
than the D position itself She follows Williamson (1987) Kayne (1994) and others in
assuming that a determiner can take a relative clause2 as its complement She further
assumes with Basilico (1996) that the determiner can serve as the binder for the gap to
form the relative clause The head position in the relative clause is limited to VP-external
2 Aldridge (2009241 fn 2) assumes that the relative clause is a TP and not a CP This is primarily because
Archaic Chinese lacked finite embedded clauses these all show evidence of nominalization On this
analysis then the phase head for the embedded clause is D rather than C
4
positions by the Phase Impenetrability Condition of Chomsky (2004) since the VP is no
longer accessible to the computational system by the time ZHE is merged
(4) DP D nP
TP ZHEi proi hellip
In this paper I propose that SUO forms relative clauses on VP-internal positions by
acting as the binder for an operator in the edge of vP This restricts SUO relativization to
VP-internal positions since the subject is licensed outside of vP and is not c-commanded
by SUO This leads to the question of why the operator resides in the edge of vP rather
than CP I propose that this is because there is no CP layer in these relative clauses The
TP headed by SUO and dominating vP is nominalized The subject moves to the specifier
of this nominalized T However since this T is not finite and therefore unable to license
nominative case the subject must value genitive case with D SUO relatives are thus
reduced relative clauses with genitive subjects in the sense of Krause (2001) Miyagawa
(2008 2011) Aygen (2011) and others
5
(5) DP D[Gen] TP
DP[Gen] Trsquo TNom vP SUOi OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
In this paper I argue in turn for each component of the analysis in (5) I begin by
showing that SUO occupies a clause-medial position and is not associated with the CP
layer I next offer indirect evidence for movement of the operator to [Spec vP] and
confirm that SUO is not this operator but rather must be analyzed as occupying a position
external to vP Finally I argue for the reduced nature of the clause by showing that the
subject receives genitive case However this case is valued under Agree with D the
subject remains in the specifier of the nominalized T and does not move to [Spec DP]
2 Position of SUO
This section argues that SUO occupies a clause-medial position between the genitive
subject and vP Crucially SUO is not associated with the CP layer or other TP-external
position
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
4
positions by the Phase Impenetrability Condition of Chomsky (2004) since the VP is no
longer accessible to the computational system by the time ZHE is merged
(4) DP D nP
TP ZHEi proi hellip
In this paper I propose that SUO forms relative clauses on VP-internal positions by
acting as the binder for an operator in the edge of vP This restricts SUO relativization to
VP-internal positions since the subject is licensed outside of vP and is not c-commanded
by SUO This leads to the question of why the operator resides in the edge of vP rather
than CP I propose that this is because there is no CP layer in these relative clauses The
TP headed by SUO and dominating vP is nominalized The subject moves to the specifier
of this nominalized T However since this T is not finite and therefore unable to license
nominative case the subject must value genitive case with D SUO relatives are thus
reduced relative clauses with genitive subjects in the sense of Krause (2001) Miyagawa
(2008 2011) Aygen (2011) and others
5
(5) DP D[Gen] TP
DP[Gen] Trsquo TNom vP SUOi OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
In this paper I argue in turn for each component of the analysis in (5) I begin by
showing that SUO occupies a clause-medial position and is not associated with the CP
layer I next offer indirect evidence for movement of the operator to [Spec vP] and
confirm that SUO is not this operator but rather must be analyzed as occupying a position
external to vP Finally I argue for the reduced nature of the clause by showing that the
subject receives genitive case However this case is valued under Agree with D the
subject remains in the specifier of the nominalized T and does not move to [Spec DP]
2 Position of SUO
This section argues that SUO occupies a clause-medial position between the genitive
subject and vP Crucially SUO is not associated with the CP layer or other TP-external
position
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
5
(5) DP D[Gen] TP
DP[Gen] Trsquo TNom vP SUOi OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
In this paper I argue in turn for each component of the analysis in (5) I begin by
showing that SUO occupies a clause-medial position and is not associated with the CP
layer I next offer indirect evidence for movement of the operator to [Spec vP] and
confirm that SUO is not this operator but rather must be analyzed as occupying a position
external to vP Finally I argue for the reduced nature of the clause by showing that the
subject receives genitive case However this case is valued under Agree with D the
subject remains in the specifier of the nominalized T and does not move to [Spec DP]
2 Position of SUO
This section argues that SUO occupies a clause-medial position between the genitive
subject and vP Crucially SUO is not associated with the CP layer or other TP-external
position
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
6
21 Below subject
To begin SUO is able to relativize on any VP-internal position The gap is the direct
object in (6a) a locative constituent in (6b) and the object of a preposition in (6c)
(6) a 人 之 所 畏 (Laozi 20)
ren zhi suo wei
person GEN SUO fear
lsquowhat people fearrsquo
b 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
Wen Wang zhi suo [VP bi feng yu e ]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
c 亂 之 所 自 起 (Mozi 14)
luan zhi suo [VP [PP zi e ] [VP qi ]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
However SUO never relativizes on subject position Even if the VP following SUO is
unaccusative the gap in the relative clause will never be the subject In (7) the
relativized position is a locative adjunct
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
7
(7) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 風 之 所 漂 (Guanzi 64)
[Feng zhi suo [VP piao e ]]
wind GEN SUO blow
lsquowhere the wind blowsrsquo
Aldridge (2010) shows that the EPP was active on T in late Archaic Chinese and
subjects were required to move out of vP to [Spec TP] Given that subjects always
precede SUO in linear order (including in unaccusatives) I assume that the EPP was
active for nominalized T as well We have seen above that SUO could not bind a gap in
subject position We can thus conclude that [Spec TP] is not in the c-command domain
of SUO Therefore SUO cannot occupy a position external to TP eg in the CP layer If
it did then it would c-command [Spec TP] and be able to relativize on subject position
conter to fact
One final piece of evidence for the TP-internal position of SUO comes from that fact
that SUO follows temporal adverbs On the assumption that temporal adverbs are
adjoined to TP or Trsquo then SUO clearly occurs within the TP projection Note that (8a)
contains the determiner ZHE in addition to SUO ZHE does not function as a relativizer
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
8
when it co-occurs with SUO but rather contributes definiteness or specificity I will have
more to say about ZHE in this function in section 42
(8) a 今 之 所 爭 者 (Zhuangzi 36)
[Jin zhi suo [VP zheng e ] zhe]
now GEN SUO fight ZHE
lsquowhat you are fighting over nowrsquo
b 始時 所 是 (Zhuangzi 35)
[Shishi suo [VP shi e ]]
beginning SUO agree
lsquothat with which (he) agreed in the beginningrsquo
22 Above vP
The preceding subsection showed that SUO is located below the position for the subject
This subsection argues that SUO is located outside of vP First SUO precedes everything
in the relative clause except the subject and temporallocative adverbials (9) shows that
SUO precedes both the clausal negator bu and the aspectual negator wei
(9) a 無 為 其 所 不 為 (Mencius 13)
Wu wei [qi suo bu wei]
NEGIMP do 3GEN SUO not do
lsquoDo not do what he does not dorsquo
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
9
b 皆 君 之 所 未 嘗 聞 (Hanfeizi 38)
Jie [jun zhi suo wei chang wen]
all lord GEN SUO notyet previously hear
lsquoAll of these are things which his lordship has never heard ofrsquo
I assume with Ernst (1995) and Hsieh (2001) that the clausal negator bu lsquonotrsquo is an
adjunct adjoined to VP However in contrast to bu Hsieh has argued that aspectual
negators in Chinese are functional heads projecting a NegP Supporting evidence for
positing wei lsquonot yetrsquo as a functional head can be found in examples of VP-ellipsis VP-
ellipsis in English is licensed only by negation when it is functional head not an adjunct
(Baltin 1993 Lobeck 1995 Potsdam 1997) Potsdam (1997) uses evidence like (10) to
argue that lsquonotrsquo in English is a functional head and not an adjunct
(10) Q Should I attend the meetings
A I suggest that you not ___ (Potsdam 1997538)
VP ellipsis can be licensed by wei lsquonot yetrsquo in Archaic Chinese
(11) Q 有 諸
You zhu
exist 3OBJ+Q
lsquoDid such a thing happenrsquo
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
10
A 未 也 (Mencius 4)
Wei __ ye
notyet COP
lsquoNot yetrsquo
In contrast the constituent negator bu does not license VP-ellipsis A different
negator fou lsquonot bersquo was required rather than bu Fou was a pronominal form replacing
a VP
(12) Q 自 織 之 與
Zi zhi zhi yu
self weave 3OBJ Q
lsquoDid weave it yourselfrsquo
A 否 以 粟 易 之 (Mencius 5)
Fou __ yi su yi zhi
notbe APPL grain trade it
lsquoNo I traded grain for itrsquo
From the above discussion it is clear that SUO precedes at least one type of
functional category the aspectual negator wei and is therefore located above the lexical
VP projection In section 5 I return to the status of wei as a functional cateory and show
how this fact provides evidence against an alternative analysis of SUO relative clauses
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
11
(13) provides additional evidence that SUO was located outside VP since it precedes
other functional categories which are likewise located outside the lexical verbal
projection SUO precedes the perfective auxiliary in (13a) and a modal in (13b)
(13) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi [suo yi cang] hai suo jiang shou
not use SUO PERF store harm SUO will receive
lsquoto not use [what you already have] to harm what you will receiversquo
b 非 人 之 所 能 為 也 (Mencius 9)
Fei [ren zhi suo neng [VP wei e ]] ye
notbe human GEN SUO can do COP
lsquo(That) is not something which a human being can dorsquo
Clinching evidence for the positioning of SUO outside the extended verbal projection
vP comes from the fact that it precedes the subject-oriented quantifier jie
(14) 此 天下 百姓 之 所 皆 難 也 (Mozi 15)
Ci [tianxia baixing zhi suo jie nan ] ye
this world commoner GEN SUO all suffer COP
lsquoThis is something which commoners the world over all agonize overrsquo
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
12
Like modern Mandarin dou jie quantifies only to its left (Luuml 1991 Li and Thompson
1981 Lee 1986 Chiu 1993 Cheng 1991 1995 and others)
(15) a Ta-men dou hen xihuan wo
3-PL all very like 1SG
lsquoThey all like mersquo
b Ta dou hen xihuan wo-men
3SG all very like 1-PL
lsquoHe likes all of usrsquo (Cheng 1995198)
Unlike dou jie quantifies only over subjects However these are not limited to
semantic subjects in [Spec vP] but include derived subjects as in the passive example in
(16) The exact analysis of how quantification takes place is not at issue in this paper The
reader is referred to Lee (1986) Cheng (1991 1995) Chiu (1993) Tsai (1994) among
many others for analyses of modern Mandarin dou What is at issue here is the relative
positions of the Archaic Chinese quantifier jie and SUO The fact that jie can quantify
over a derived subject in a passive suggests that jie is not a stranded quantifier in [Spec
vP] It also cannot be located inside VP since it never occurs in postverbal position and
never quantifies over a VP-internal constituent It is most natural then to assume that jie
is located outside vP Note further that the quantifier precedes the passive auxiliary ke
Assuming that a passive auxiliary is located no lower than Voice or v then the quanitifier
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
13
is clearly located outside of this projection Crucially given that SUO also precedes the
quantifier we can safely conclude that SUO is located outside of vP
(16) 皆 可 謂 能 禮 士 矣
Jie ke wei neng li shi yi
all PASS say can respect gentleman PERF
lsquo(They) all can be said to be able to respect a man of classrsquo
(Luumlshi Chunqiu 135)
3 Evidence for movement
Having shown that the position occupied by SUO is located between the subject and the
vP I turn now to evidence for null operator movement to the edge of vP I first offer
indirect evidence for the possibility of such operator movement from wh-fronting I then
provide supporting evidence for operator movement on the basis of island sensitivity in
SUO relatives
31 Parallel with wh-movement
Late Archaic Chinese VP-internal constituents generally underwent fronting to preverbal
position when they were wh-phrases Interestingly this operation is parallel to SUO
relatives in targeting any and only VP-internal constituents (17a) shows movement of a
direct object Note in the second clause of (17a) that basic word order in the language is
VO with the object following the verb when it was not a wh-constituent (17b) shows
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
14
movement of a locative or dative element And (17c) shows movement of the object of a
preposition
(17) a 吾 誰 欺 欺 天 乎 (Analects 9)
Wu shei [VP qi tshei ] Qi tian hu
1 who deceive deceive Heaven Q
lsquoWho do I deceive Do I deceive Heavenrsquo
b 其 子 焉 往 (Mencius 7)
qi zi yan [VP wang tyan ]
3GEN son where go
lsquoWhere will their sons gorsquo
c 吾 又 誰 與 爭 (Zuozhuan Zhao 4)
Wu you shei [VP [PP yu tshei ] zheng]
I then who with compete
lsquoThen who would we compete withrsquo
Raised wh-phrases precede the modal neng as well as negation which is also parallel
to SUO
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
15
(18) a 是 何 能 害 (Luumlshi Chunqiu 161)
Shi he neng hai the
this what can harm
lsquoWhat can this harmrsquo
b 然則 我 何 為 乎 何 不 為 乎
Ranze wo he wei hu He bu wei hu
then I what do Q what not do Q
lsquoThen what should I do What should I not dorsquo (Zhuangzi 210)
It is somewhat challenging to demonstrate that the landing site for wh-movement is
below the CP layer since subjects could also be questioned and appeared in preverbal
position Simply showing that the object wh-word follows a referential subject in surface
order does not suffice because subjects could easily be topicalized in Archaic Chinese
This leads to two possible analyses of Archaic Chinese wh-movement Either the wh-
word moves to [Spec CP] as in (19a) or it targets a position in the lower phase edge as
in (16b) On the former analysis the referential subject is analyzed as a topic also located
in the CP layer
(19) a [TopP Wo jiang [FocP he [TP two qiu the ]]]] (Wh-position in CP edge)
I will what askfor
b [TP Wo jiang [vP he [vrsquo two [vrsquo qiu the ]]]]] (Wh-position in vP edge)
I will what askfor
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
16
The choice between the two alternatives depends on indirect evidence specifically
the relative positions of subject and object wh-words with respect to other elements in the
clause (19a) predicts that both subjects and objects move to the same position since both
subjects and objects would be in the c-command domain of a probe on C triggering wh-
movement (20) shows that this prediction is not borne out As first observed by Wei
(1999) subject wh-words precede the modal jiang as in (20a) while object wh-words
follow jiang as in (20b) This clearly shows that object wh-movement targets a position
lower than subject wh-words Furthermore this position cannot be in the CP layer but
rather in a position which does not c-command the subject in [Spec TP]
(20) a 誰 將 治 之 (Yanzi Chunqui Nei 113)
Shei jiang zhi zhi
who will govern them
lsquoWho will govern themrsquo
b 我 將 何 求 (Zuozhuan Xi 28)
Wo jiang he qiu
I will what askfor
lsquoWhat will I ask forrsquo
Adopting proposals for the existence of Arsquo positions for topic and focus in the vP
layer in Italian (Belletti 2004) and modern Mandarin (Paul 2005) Aldridge (2010)
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
17
proposes that object wh-movement targets a focus position in the vP phase edge Subject
wh-words are assumed to remain in situ in [Spec TP] The interrogative interpretation is
obtained via unselective binding by an operator merged in [Spec CP] as per Tsairsquos (1994)
analysis of modern Mandarin wh-in-situ
(21) CP OP Crsquo C TP
DPSubj Trsquo T vP XP[Foc Wh] vrsquo
tSubj vrsquo v[Foc] VP hellip tXP hellip
Support for the unselective binding analysis comes from evidence that late Archaic
Chinese wh-words are indefinites and not quantificational operators As shown in (22)
wh-words could be used in this period with non-interrogative interpretations
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
18
(22) a 不 知 我 者 謂 我 何 求 (Shijing Shuli)
Bu zhi wo zhe wei [wo he [qiu the ]]
not know 1 DET say 1 what seek
lsquoThose who do not know me say that I am looking for somethingrsquo
b 誰 之 不 如 可以 求 之 (Guoyu Jin 6)
[Shei zhi bu [ru tshei ]] keyi qiu zhi
who GEN not compare can follow 3OBJ
lsquoIf you donrsquot measure up to someone you can follow himrsquo
What the above proposal suggests for SUO relative clauses is that there was a
position in the edge of vP to host operator movement The following example further
shows that SUO itself is not this operator since SUO clearly precedes the modal jiang
Recall from (20b) that object wh-words follow this modal
(23) a 不 以 所 已 藏 害 所 將 受 (Xunzi 21)
bu yi suo yi cang hai [suo jiang shou]
not use SUO PERF store harm SUO will receive
lsquoto not use what you already have to harm what you will receiversquo
b 其 所 將 為 (Zhuangzi 33)
[qi suo jiang wei]
3GEN SUO will do
lsquowhat he will dorsquo
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
19
32 Movement in SUO relatives
Having seen in the preceding subsection that late Archaic Chinese had wh-movement to
the edge of vP this subsection presents evidence that SUO relatives are likewise derived
through movement First there is no evidence that operator movement in a SUO relative
can cross an island barrier Specifically I have found no examples of SUO relativization
out of an island This indirectly suggests that operator movement in SUO relatives was
sensitive to locality constraints on movement More concrete evidence is provided by the
modern Mandarin reflex of SUO relative clauses Both subject and object relatives are
typically formed with the linking element DE in modern Mandarin as in (24a b)
However it is also possible to add SUO in object relatives as in (24c)
(24) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
20
Interestingly Chiu (1995) shows that while gaps in DE relatives are permitted in
some types of islands gaps are not permitted to occur inside islands in relatives involving
SUO In (25) the gaps are contained within sentential subjects within the relative clauses
(25a) without SUO is grammatical while the relative with SUO in (25b) is
ungrammatical I discuss approaches to modern Mandarin SUO in section 5
(25) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
At this point a word is in order regarding SUO relativization on the object of a
preposition as seen in (6c) and repeated below Assuming that adjunct PPs are islands to
extraction then movement of the operator from the gap position to the edge of vP would
be a counterexample to the generalization made above that SUO relativization is sensitive
to island barriers
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
21
(26) 亂 之 所 自 起 (Mozi 14)
luan zhi suo [vP [PP OP [Prsquo zi tOP ]] [vrsquo tSubj [VP qi ]]]
unrest GEN SUO from arise
lsquofrom whence unrest arisesrsquo
There is an alternative analysis which does not entail movement from the PP Aissen
(1996) proposes that wh-movement of the object of a preposition in Tzotzil involves
fronting of the wh-phrase to the left edge of the PP and then pied-piping of the entire PP
to the scope position (27a) shows that PPs are head-initial in Tzotzil However when a
PP contains a wh-word the wh-word must precede the preposition
(27) a I-kom [PP ta s-na] (Tzotzil Aissen 1996468)
CP-remain P A3-house
lsquoHe remained in his housersquo
b [CP [PP Buchrsquou [Prsquo ta [DP s-na twh ]]] [Crsquo C[+WH] [IP ch-a-bat tPP ]]]
who P A3-house ICP-B2-go
lsquoTo whose house did you gorsquo (Tzotzil Aissen 1996470-1)
There is indirect evidence for the pied-piping analysis in the case of Archaic Chinese
Certain types of PP could appear postverbally if they did not contain a wh-word or the
gap in a SUO relative clause as in (28) Note that the preposition is the same as the one
seen in (26)
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
22
(28) 秋 公 至 自 晉 (Zuozhuan Cheng 4)
Qiu gong zhi zi Qi
Autumn lord arrive from Qi
lsquoIn the Autumn the lord arrived from Qirsquo
However I know of no examples of wh-questions or SUO relative clauses in which a
preposition is stranded in postverbal position In other words the a preposition hosting a
gap must appear in preverbal position This suggests at least indirectly that the PP has
been moved to preverbal position I assume that this position is the operator position in
the edge of vP as shown by the bracketing in (28)
Interestingly wh-movement in Archaic Chinese mirrors SUO relativization A wh-
word which is the object of a preposition appears to the left of the preposition and the PP
surfaces in preverbal position Again this can analyzed as pied-piping of the PP
containing the wh-word to the [Spec vP] focus position
(29) 吾 又 誰 與 爭
Wu you [vP [PP shei [Prsquo yu tshei ]] [vrsquo tSubj [VP zheng]]]
I then who with compete
lsquoThen who would we compete withrsquo (Zuozhuan Zhao 4)
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
23
This section has presented evidence that SUO relativization is derived through
operator movement that is sensitive to locality constraints The next section addresses the
final component of the analysis proposed in (5) by showing that the structure above vP is
a nominalization and consequently that SUO relative clauses are a type of mixed
projection
4 SUO clauses as nominalizations
To recap the discussion so far SUO heads a functional projection above vP where it
binds an operator which has moved to the edge of vP to form a relative clause The
question which must be asked at this point is why the relative operator remains in [Spec
vP] and does not move further to [Spec CP] I propose that the reason is because there is
no CP layer in this type of relative clause What I show in this section is that a SUO
relative clause is a nominalized TP It is therefore nonfinite and is not dominated by a CP
layer Rather the nominalized TP is selected by D which supplies genitive case to the
subject
41 Genitive subjects
First note that SUO relatives very frequently have genitive subjects3 More examples are
shown below
3 Genitive case does not always appear overtly on the subject of a SUO relative The constraints governing
the morphological realization of genitive case in this configuration are complex and include prosodic as
well as pragmatic considerations For the purposes of this paper I assume that these subjects uniformly
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
24
(30) a 以 丘 之 所 聞 羊 也 (Guoyu Lu 2)
Yi [Qiu zhi suo wen] yang ye
by Qiu GEN SUO hear sheep COP
lsquoAccording to what I (Qiu) have heard (it) is a sheeprsquo
b 聖人 之 所 不 能 勝 也 (Zhuangzi 14)
[shengren zhi suo bu neng sheng] ye
saint GEN SUO not able overcome COP
lsquo(These) are things which (even) saints cannot overcomersquo
The zhi genitive case marker also appears with possessors in a DP
(31) a 先 王 之 道 (Analects 1)
Xian wang zhi dao
former king GEN way
lsquoways of the former kingsrsquo
value genitive case in the syntax The choice of phonetic form is determined post-syntactically in the
Morphological Component
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
25
b 寡人 之 身 (Mencius 1)
guaren zhi shen
1 GEN body
lsquomy bodyrsquo
Subjects of nominalized clauses also take this case marker (32a) is an example of a
sentential subject (32b) shows a complement clause Both types of embedded clauses
were obligatorily nominalized in late Archaic Chinese
(32) a 臣 之 事 君 義 也 (Zhuangzi 14)
[chen zhi shi jun] yi ye
minister GEN serve jun duty COP
lsquoA minister serving his lord is dutyrsquo
b 臣 固 知 王 之 不 忍 也 (Mencius 1)
Chen gu zhi [wang zhi bu ren] ye
I already know king GEN not bare COP
lsquoI already knew you would not be able to bare itrsquo
Given that complements of perception verbs were nominalized as in (32b) it is not
surprising that a SUO relative clause can appear in complement position In fact SUO
relativization was the only way to express an embedded object wh-question
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
26
(33) a 有司 未 知 所 之 (Mencius 2)
Yousi wei zhi [OP suo zhi]
servant not know SUO go
lsquoI do not know [where you are going]rsquo
b 不 知 亂 之 所 自 起
Bu zhi [luan zhi suo [VP [PP zi e ] [VP qi ]]]
not know unrest GEN SUO from arise
則 不 能 治 (Mozi 14)
ze bu neng zhi
CONJ not can govern
lsquoIf (one) does not know [from whence unrest arises] then (one) cannot
governrsquo
42 Low subject position
The final part of the analysis to be demonstrated is the proposal that the subject in SUO
relatives remains within the nominalized TP headed by SUO and values genitive case
under Agree with D rather than moving to [Spec DP] The first indication that this is the
case comes from the fact that genitive subjects follow demonstratives like bi in (34)
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
27
(34) a 彼 其 所 與 至 者 必 其 民 也
[Bi [qi suo yu zhi] zhe] bi qi min ye
DEM 3GEN SUO P arrive DET will 3GEN people COP
lsquoThat with which he will arrive will certainly be his peopersquo (Xunzi 15)
b 彼 其 所 殉 仁 義 也
[Bi [qi suo xun]] ren yi ye
DEM 3GEN SUO selfsacrifice benevolence righeousness COND
則 俗 謂 之 君子 (Zhuangzi 21)
ze su wei zhi junzi
then common call 3OBJ gentleman
lsquoIf that for which he sacrifices himself are the ideals of benevolence and
righteousness then commoners will call him a gentlemanrsquo
Even more convincing is the fact that a temporal adverb can precede the genitive
subject Miyagawa (2008) uses similar evidence to argue that genitive subjects in
Japanese relative clauses likewise do not raise to [Spec DP]
(35) a 此 昔 吾 先 王 之 所 以 霸
Ci [xi [wu xian wang] zhi suo yi ba]
this past 1GEN former king GEN SUO APPL conquer
lsquoThis is why our former king was victorious in the pastrsquo
(Luumlshi Chunqiu 145)
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
28
b 凡 古 今 天下 之 所 謂 善 者
[fan [gu jin [tianxia zhi suo wei shan]] zhe]
generally old now world GEN SUO call good DET
lsquogenerally what the world refers to as ldquogoodrdquo both now and in the pastrsquo
(Xunzi 23)
As can be seen in many of the SUO relatives given above SUO can also occur with
the determiner ZHE In these cases ZHE is not functioning as a relativizer since the
operator in vP is bound by SUO Rather ZHE adds definiteness or specificity to the DP
as a whole This is clearly shown in (36a) where the context indicating that speakers
have things to say is established in the first clause The referent in (36b) is indefinite
since it is unknown However it is clearly specific since the wife assumes that her
husband has been eating and drinking with someone
(36) a 言 者 有 言
Yan zhe you yan
speak ZHE have speech
其 所 言 者 特 未 定 (Zhuangzi 12)
[qi suo yan zhe] te wei ding
3GEN SUO say ZHE but not uniform
lsquoOnes who speak have things to say but what they have to say is not
uniformrsquo
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
29
b 其 良人 出 則 必 饜 酒 肉 而 返
Qi liangren chu ze bi yan jiu rou er fan
3GEN husband leave CONJ always fill liquor meat CONJ return
其 妻 問 所 與 飲 食 者
Qi qi wen [suo yu yin shi zhe]
3GEN wife ask SUO with drink eat ZHE
則 盡 富 貴 也 (Mencius 8)
ze jin fu gui ye
CONJ all rich powerful COP
lsquoWhenever the husband went out he would come back well fed and liquored
His wife asked who he ate and drank with and (the answer was) all rich and
powerful peoplersquo
As discussed in section 1 Aldridge (2009) argues that ZHE occupies a position
structurally lower that D With this in mind now observe the following examples in
which two SUO relatives with genitive subjects are coordinated under ZHE
(37) a 言 之 所 不 能 論
[[Yan zhi suo bu neng lun]
word GEN SUO not can debate
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
30
意 之 所 不 能 致 者
[yi zhi suo bu neng zhi] zhe]
intention GEN SUO not can bring DET
不 期 精 粗 焉 (Zhuangzi 210)
bu qi jing cu yan
not limit fine coarse there
lsquoThat which words cannot debate and intentions cannot summon is not limited
to what is coarse or what is finersquo
b 此 商 君 之 所 以 車裂 于 秦 而
Ci Shang Jun zhi suo yi che-lie yu Qin er
this Shang lord GEN SUO APPL cart-rip in Qin CONJ
吳 起 之 所 以 枝解 于 楚 者 也
Wu Qi zhi suo yi zhi-jie yu Chu zhe ye
Wu Qi GEN SUO APPL limb-remove in Chu DET COP
lsquoThis is the reason why Lord Shang was executed (by being torn apart by
moving carts) in Qin and Wu Qi was executed in Chu (by having his limbs
removed)rsquo (Hanfeizi 14)
Each of the conjuncts contains a genitive subject indicating that these subjects must
be located in a structurally lower position than ZHE Given that ZHE occupies a position
structurally lower than D it is clear that the genitive subjects in (37) do not occupy [Spec
DP]
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
31
(38) DP D[Gen] nP SuoP ZHE DP[Gen] Suorsquo
Suoi vP
OPi vrsquo
tSubj vrsquo v VP hellip tOP hellip
5 Loss of SUO Relativization
The analysis presented in the previous section of SUO relatives as nominalizations also
helps understand changes which took place in this construction in Middle Chinese
Beginning in early Middle Chinese in the 1st century CE SUO was no longer obligatory
for forming object relative clauses In this section I argue that the loss of SUO
relativization was due to the loss of the nominal layer dominating the vP The
morphological trigger for this loss was the loss of the Archaic Chinese morphological
case distinctions including the genitive case which appeared on subjects of embedded
clauses Without this morphological evidence for the nominalization I propose that
language learners acquired the default (in the sense of Roberts 1997 and Roberts and
Roussou 2003) structure for relative clauses involving a finite CP The addition of the CP
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
32
layer above the TP was acquired due to the necessity of transfer of the features from C to
finite T according to the C-T transfer hypothesis of Chomsky (2005 2008) and others
First observe that SUO is no longer obligatory in forming object relative clauses in
early Middle Chinese Both of the relative clauses in (39) are formed on the Archaic
Chinese subject relativizer ZHE
(39) a 寡人 好 者 音 也 (1st C CE Lunheng 64)
[Guaren hao zhe] yin ye
1 like DET music NMLZ
lsquoWhat I like is musicrsquo
b 我 請 君 塞 兩 耳
Wo qing jun sai liang er
1 ask lord close two ear
無 聽 談 者 (1st C BCE Zhanguoce Zhao 1)
wu ting [tan zhe]
donot listen discuss DET
lsquoI asked my lord to close his ears and not listen to what was being discussedrsquo
What has happened is that the subject relativization strategy has been extended to
object relatives As I pointed out in section 32 a single strategy is employed to form
both subject and object relative clauses in modern Mandarin The linking element DE
appears following the clause in both cases Interestingly DE traces its history as a
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
33
relativizer to Archaic Chinese ZHE Specifically DE (的) ndash DI (底) in late Middle
Chinese ndash was a lexical replacement for ZHE (者) (Luuml 1943 Ohta 1958 Cao 1986 Feng
1991 and others) Therefore it is clear that the subject relativization strategy was
extended to object relatives
(40) a [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
b [mai shu de] ren
buy book DE person
lsquoperson who bought athe bookrsquo
The questions to be addressed at this point are why SUO relatives were lost (ie
became non-obligatory) and why the subject relativization strategy was generalized to
object relativize clauses I propose here that what took place was the loss of the nominal
layer dominating the SUO TP and the reanalysis of SUO clauses as finite CPs With the
addition of a CP layer operator movement could access this phase edge which
effectively eliminated the formal distinction between subject and object relative clauses4
4 In this paper I have assumed with Aldridge (2009) that non-control embedded clauses in Late Archaic
Chinese were all nominalized Consequently the complement of the determiner ZHE in subject relative
clauses was a (nominalized) TP and not a CP I therefore further assume that ZHE relatives underwent a
reanalysis similar to SUO relatives in which the nominalized TP was reanalyzed as a finite CP
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
34
The subject relativization strategy which involved embedding of a full clausal structure
could then be employed in formation of both subject and object relative clauses
Evidence for the denominalization of SUO relative clauses in Middle Chinese comes
from two sources First is the loss of morphological genitive case on the subject I further
suggest that this was the trigger which led to the reanalysis Without this mophological
trigger learners of the language acquired relative clauses as embedded CPs rather than
DPs I discuss this phenomenon first Then I provide supporting evidence for the addition
of the CP layer in SUO relatives by showing that these could involve long distance
dependencies
First it is clear that genitive case has been lost by the 5th century CE from texts like
the following In the Baiyujing genitive case on subjects of SUO relatives is entirely
dictated by prosodic factors In Middle Chinese translations of Buddhist texts there was a
tendency to create four-syllable phrases Genitive case was frequently added or omitted
in order to ensure this pattern No genitive case appears on the subject of a SUO relative
if the clause or phrase already consists of four syllables as in (41a) But genitive case can
be added if this would result in a phrase four syllables long as in (41b)
(41) a 是 我 所 作 (5th C CE Baiyujing 41)
[Shi wo suo zuo]
this I SUO make
lsquoThis is something I madersquo
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
35
b 我 之 所 記 終 無 違失
[Wo zhi suo ji] zhong wu weishi
I Gen SUO record finally notexist exception
lsquoIn the end there were no exceptions to what I recordedrsquo
(5th C CE Baiyujing 41)
It can also be demonstrated that genitive case in general was lost by early Middle
Chinese A number of morphological case distinctions can be observed in late Archaic
Chinese Zhi was used for third person accusative pronouns Note that the same graph is
used to write the accusative pronoun and genitive case on subject and possessor DPs
This does not mean however that the two cases were morphologically nondistinct I
show below that genitive pronouns took a different form (42) shows that ZHI was used
for the theme in a dative construction the goal in a double object construction ZHI also
was used for an exceptionally case-marked embedded subject
(42) a 臣 之 子 亦 不 能 受 之 於 臣
Chen zhi zi yi bu neng [shou zhi yu chen]
1 GEN son also not can receive 3OBJ P 1
lsquoMy son also cannot receive it from mersquo (Zhuangzi 13)
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
36
b 卒 授 之 國 (Zhuangzi 5)
Zu [shou zhi guo]
finally bestow 3OBJ nation
lsquoIn the end I gave him (responsibility for) the nationrsquo
c 上 賢 使 之 為 三公 (Xunzi 12)
Shang xian shi [zhi wei sangong]
most able make 3OBJ be sangong
lsquoThe most capable make them into sangong (the highest official rank)rsquo
On the other hand subjects of nominalized embedded clauses were marked with
genitive case When this subject was a third person pronoun the form was qi Thus ZHI
is also clearly distinguished in this period from the genitive pronoun 其 qi lsquo3GENrsquo
(43) a 周 公 知 其 將 畔 而 使 之 與
Zhou gong zhi [qi jiang pan] er shi zhi yu
Zhou duke know 3GEN will rebel CONJ send 3OBJ Q
lsquoDid the duke of Zhou send him knowing [that he would rebel]rsquo
(Mencius 4)
b 莫 知 求 其 所 已 知 者 (Zhuangzi 23)
Mo zhi qiu [qi suo yi zhi zhe]
noone think seek 3GEN SUO PERF know DET
lsquoNoone thinks to seek [what he already knows]rsquo
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
37
From early Middle Chinese however QI came to be used for ECM embedded
subjects
(44) a 全 趙 令 其 止
Quan Zhao ling [TP qi zhi ]
protect Zhao make 3GEN stop
lsquo(You) protect Zhao and make them stoprsquo (1st C BCE Zhanguoce Xi Zhou)
b 欲 使 其 生 於 階庭 耳
Yu shi [qi sheng yu jieting] er
want make 3GEN grow in courtyard PART
lsquo(You) want to make them grow in a courtyardrsquo
(5th C CE Shishuo Xinyu Yanyu)
Later in Middle Chinese QI could be found as the goal in a double object
construction
(45) a 唯 願 世尊 與 其 長壽
Wei yuan Shizun yu qi changshou
only wish Buddha give 3GEN longevity
lsquo(I) only ask the Buddha to give him long lifersquo
(5th C CE Zabao Zangjing 47)
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
38
b 吾 與 其 眼 (5th C CE Xianyu Jing 6)
Wu yu qi yan
1 give 3GEN eye
lsquoI give him eyesrsquo
In Middle Chinese then we can conclude that the morphological form of the third
person pronoun no longer directly reflects the case licensed in the syntax One possible
explanation for the appearance of genitive pronouns in the environments just given may
be that the genitive form was used for accusative pronouns appearing in specifier
positions The fact that an accusative pronoun in complement position is never replaced
with a genitive offers indirect evidence for this suggestion
(46) a 王 乃 召 相 令 之 曰 (Zhanguoce Qin 5)
Wang nai zhao xiang ling zhi yue
king then call minister order 3OBJ say
lsquoThe king then summoned his prime minister and issued him an order
sayingrsquo
b 大 王 可 試 使 之 (Zhanguoce Zhao 4)
Da wang ke shi shi zhi
great king can try send 3OBJ
lsquoYour majesty could try to send himrsquo
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
39
Further support for the proposal that genitive case was lost in early Middle Chinese
comes from the fact that embedded subjects in general were no longer required to appear
with the genitive marking (47a) shows an Archaic period sentential subject with genitive
marking (47b) shows a similar sentence in a Han period historical chronical The Han
example does not use genitive case for the embedded subject
(47) a 天下 之 無 道 也 久 矣
[Tianxia zhi wu dao ye] jiu yi
world GEN nothave way NMLZ long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
(5th C BCE Analects 3)
b 天下 無 道 久 矣 (1st C CE Shiji 47)
[Tianxia wu dao ] jiu yi
world nothave way long PERF
lsquoIt is a long time since the world has been without the proper wayrsquo
I take examples like (47b) to be direct evidence for the emergence of embedded
(finite) CPs in the language The way that this came about was that the loss of genitive
case removed the morphological trigger for learners to acquire the more marked
nominalization structure for embedded clauses The learners then acquired the default
clausal structure involving a TP with a CP layer
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
40
I next turn to evidence for a finite CP structure in SUO relative clauses (48) shows
that a SUO relative could be formed on a long distance dependency The existence of a
clause boundary intervening between SUO and the gap suggests that there must be a CP
layer present to allow the operator to escape from the most deeply embedded clause and
thereby avoid violating the Phase Impenetrability Condition of Chomsky (2000 2001
2004 2005 2008) (48) is an example from a very late Archaic Chinese text I have
found no examples of this sort in early texts suggesting that this possibility emerged very
late in Archaic Chinese at the very cusp of the Middle Chinese period
(48) 是 所 使 夫 百吏 官人 為 也
Shi [suo shi [[fu baili guanren] [vP wei e ]]] ye
DEM SUO make DEM clerk official do NMLZ
lsquoThis is something which one makes those clerks and officials dorsquo
(Xunzi 11)
If we accept that SUO clauses have been reanalyzed as CPs in Middle Chinese it is
also not surprising that SUO comes to be employed in the formation of long passives in
early Middle Chinese This type of passive involves embedding a SUO clause under the
copula wei The matrix subject precedes the copula and is semantically construed as the
gap in the SUO clause
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
41
(49) a 微 趙君 幾 為 丞相 所 賣
Wei Zhao Jun ji wei [chengxiang suo mai]
notbe Zhao Jun almost COP primeminister SUO betray
lsquoWere it not for you Zhao Jun (I) would have been betrayed by the prime
ministerrsquo (1st C BCE Shiji Lisi)
b 高祖 擊 黥布 為 流 矢 所 中
Gaozu ji Qing Bu wei [liu shi suo zhong]
Gaozu attack Qing Bu COP stray arrow SUO hit
lsquoGaozu attacked Qing Bu was hit by a stray arrow and became very illrsquo
(1st C CE Lunheng Minglu)
I assume the following analysis (based on analysis of Mandarin long passives by Chiu
1993 1995 Huang 1999 Huang et al 2009) in which SUO heads TP in an embedded
CP Operator movement takes place to the edge of this CP where the operator can be
coindexed with the matrix subject The CP layer must be assumed in this case given that
the operator needs to be able to access a phase edge where it will be accessible to the
matrix subject If the SUO clause were a DP there would be no [Spec CP] for the
operator to move to and the operator would be left in the edge of the embedded vP This
would separate the operator and the matrix subject by the DP phase layer and prevent
construal of the two without violating the Phase Impenetrability Condition
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
42
(50) TP proi T vP ltprogt wei CP OPi TP chengxiang SUO vP ltOPgt ltchengxianggt mai ltOPgt
It is not controversial that SUO passives are historically related to modern Mandarin
long passives The copula wei was replaced by the modern Mandarin passive marker bei
in Middle Chinese in approximately the 6th century (Wang 1958 Ohta 1958 Bennett
1981 Tang 1987 Wei 1994 Sun 1996) At a later date SUO was lost (51) shows a
modern Mandarin long passive
(51) Zhangsan bei Lisi da-le
Zhangsan BEI Lisi hit-ASP
lsquoZhangsan was hit by Lisirsquo
What is important for the discussion at hand is that modern Mandarin long passives
can be built on long distance dependencies providing some of the evidence for the null
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
43
operator movement analysis of Chui (1993 1995) Ting (1998) and Huang et al (2009)
Thus we have further indirect evidence for the CP analysis of SUO clauses in Middle
Chinese
(52) Zhangsani bei [CP OPi [Lisi pai jingcha [CP ltOPgt [zhuazuo-le ltOPgt]]]]
Zhangsan Pass Lisi send police arrest-Perf
lsquoZhangsan suffered Lisi sending police to arrest himrsquo (Huang et al 2009125)
In this section I proposed that SUO relativization came into disuse as an obligatory
strategy for object relative clause formation in Middle Chinese Specifically the
nominalized TP projected by SUO was reanalyzed as a finite CP The trigger for this
reanalysis was the loss of genitive case Without this morphological cue learners of the
language acquired the less marked strategy for clausal constituents consequently
acquiring embedded clauses including relative clauses as finite CPs rather than
nominalizations
6 Comparing with Other Analyses
Before concluding this paper I consider proposals which have been made for the modern
Mandarin reflex of SUO and show that these approaches cannot be applied to SUO of the
late Archaic Chinese period As mentioned in section 32 SUO can optionally appear in a
relative clause formed on a VP-internal position in modern Mandarin (53a b) show that
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
44
object relatives can be formed with SUO and the linker DE or just with DE in modern
Mandarin (53c) shows that subject relatives must be formed with just the linker
(53) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b [Lisi mai de] shu
Lisi buy DE book
lsquobook which Lisi boughtrsquo
c [(suo) mai shu de] ren
SUO buy book DE person
lsquoperson who bought athe bookrsquo
In order to account for the fact that SUO occurs only in relatives formed on VP-
internal positions Chiu (1993 1995) proposes that SUO heads the accusative case-
checking projection The argument moving to the specifier of this projection is assigned
accusative case so SUO relativization is restricted to accusative objects The overt
appearance of SUO is triggered by operator movement to the specifier of SuoP
accounting for the restriction of SUO to relative clauses SuoP is located above TP due
to the fact that SUO precedes TP-internal elements like aspectual markers negation and
subject-oriented quantifiers the last of which Chiu locates in T
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
45
(54) a [Lisi suo mai de] shu
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
b CP OP C NomP Lisi Nom SuoP ltOPgt suo TP T VP ltLisigt mai ltOPgt
This analysis is able to account for the lack of subject relativization with SUO and
also for the fact that SUO relativization is sensitive to island barriers as discussed in
section 3
(55) a [[Lisi kan ei ] zui heshi de] shui
Lisi read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
46
b [[Lisi suo kan ei ] zui heshi de] shui
Lisi SUO read most appropriate DE book
lsquothe book that it is most appropriate for Lisi to readrsquo
Nevertheless this analysis cannot be adopted for late Archaic Chinese Most
damaging is the fact that late Archaic Chinese SUO does not always relativize on
accusative object position (56) repeats an example (6b) involving a locative gap The
locative here cannot be analyzed as the accusative marked argument because there is a
direct object inside the SUO relative which must value accusative case
(56) 文 王 之 所 避 風 雨 (Zuozhuan Xi 32)
[Wen Wang zhi suo [VP bi feng yu e ]]
Wen king GEN SUO escape wind rain
lsquowhere the (Zhou) king Wen escaped the stormrsquo
Furthermore SUO can relativize on unaccusative (including passive) VPs as seen in
section 21 The gaps in (57) are locative adjuncts The internal argument subject in (57a)
surfaces to the left of SUO and values genitive case The subject is topicalized in (57b)
and is null within the relative clause itself What is important for the argument at hand is
that given Burziorsquos (1986) generalization that accusative case is not available in
unaccusatives examples like the following cannot be accounted for on an approach
which requires SUO to head an accusative case projection
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
47
(57) a 榖食 之 所 生 (Zhuangzi 210)
[gushi zhi suo [VP sheng e ]]
grain GEN SUO grow
lsquowhere grain growsrsquo
b 是 不材 之 木 也
Shi bucai zhi mu ye
DEM worthless GEN tree TOP
無 所 可 用 (Zhuangzi 14)
wu [pro suo ke yong]
notexist SUO PASS use
lsquoThere is no place for this worthless tree to be usedrsquo
Ting (2003 2010) shows that even in modern Mandarin SUO relatives are not
limited to accusative object gaps (58) shows a SUO relative formed on a locative
constituent Consequently Chiursquos (1993 1995) proposal suffers from inadequacies even
in accounting for modern Mandarin
(58) Lisi suo gongzuo de difang (Ting 2003125)
Lisi SUO work DE place
lsquothe place where Lisi worksrsquo
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
48
Another problem for the accusative case approach is that SUO in modern Mandarin
can be separated from the gap by a clause boundary as in (59b) Assuming that structural
case on the operator forming the relative clause is valued locally by embedded v (as per
Chomsky 2004) the function of SUO in the higher clause cannot be case-checking or
valuing Even if accusative case is assumed to originate from a higher functional head
accusative case valuing on the operator cannot be attributed to SUO in (59b) since the
accusative case in the matrix clause is required by the matrix object
(59) a [wo rang Zhangsan [PRO suo goumai e ]] de shu
1SG make Zhangsan SUO buy DE book
lsquothe book that I made Zhangsan buyrsquo
b [wo suo rang Zhangsan [PRO goumai e ]] de shu
1SG SUO make Zhangsan buy DE book
lsquothe book that I made Zhangsan buyrsquo (Ting 2003127)
Ting (2003 2010) proposes an alternative analysis of modern Mandarin SUO as an
Arsquo-bound pronominal clitic SUO heads the NP which occurs in the gap position in the
relative clause and undergoes head movement out of NP and VP and adjoins to Infl It is
coindexed with and bound by a null operator in [Spec CP] of the relative clause
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
49
(60) [NP [CP OPi [IP Lisi [Irsquo suoi+I [VP mai [NP [Nrsquo tsuo ]]]]] de ] shui ]
Lisi SUO buy DE book
lsquobook which Lisi boughtrsquo
This analysis accounts for the position of SUO between the subject and VP It also
accounts for the fact that only VP-internal positions can be relativized with SUO since
the SUO projection is located below the position for the subject in [Spec IP] It also
avoids the problem of associating SUO with a case position Head-movement will be
allowed from any NP in the complement position of the verb
However there are several other problems inherent in this proposal Given that verbs
in modern Chinese move no higher than v or the head of the highest VP shell in the
clause (Tang 1990 Huang 1991 1993 1994 Gu 1995 Tang 2001 and others) then the
movement shown in (60) violates Travisrsquo (1984) Head Movement Constraint since SUO
must move over the verb to reach Infl
To address this challenge Ting (2003 2010) appeals to the possibility of long head
movement (in the sense of Roberts 1994) via excorporation However it bears
mentioning that Roberts (1994) restricts excorporation to L-related heads specifically
verbal and inflectional categories but not other heads like Neg or C An immediate
problem concerns how to treat cases like (59b) in which SUO is separated from its base
position by a clause boundary Ting (2010) argues that the embedded clause is a TP and
not a CP thereby circumventing the problem of excorporation from a non-L-related C
head
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
50
A more damging concern is that SUO precedes negators in the relative clause Note in
particular the negator wei lsquonot yetrsquo I argued in section 22 that wei lsquonot yetrsquo in late
Archaic Chinese was a functional head projecting a NegP along the lines of Hsiehrsquos
(2001) proposal for aspectual negators in modern Mandarin Given that wei projects
NegP and is therefore not an L-related head for the purposes of Robertsrsquo (1994)
excorporation then head movement of SUO should not be able to traverse through wei on
on its way to Infl
(61) a 君 之 所 未 嘗 食 唯 人 肉 耳
[Jun zhi suo wei chang shi] wei ren rou er
lord GEN SUO notyet ever eat only human meat FOC
lsquoWhat the lord had not yet tasted was only human fleshrsquo (Hanfeizi 10)
b 其 過 三 皆 君 之 所 未 嘗 聞
Qi guo san jie [jun zhi suo wei chang wen]
3GEN error 3 all lord GEN SUO notyet ever hear
lsquoHe made three mistakes All are things which his lordship has never heard
ofrsquo (Hanfeizi 38)
Another problem is the fact that SUO must be stipulated to act as a resumptive
pronoun only in relative clauses SUO does not surface in other extraction contexts like
topicalization A resumptive pronoun did appear in Archaic Chinese when an object was
topicalized But SUO was not employed in this way
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
51
(62) a 諸侯 之 禮 吾 未 之 學 也
[Zhuhou zhi li] wu wei zhi xue ye
feudallord GEN rite 1 notyet 3OBJ study COP
lsquoThe rites of the feudal lords I have not yet studiedrsquo (Mencius 6)
b 子路 人 告 之 以 有 過
Zilu ren gao zhi yi you guo
Zilu person tell 3OBJ that have error
lsquoZilu someone told him he made a mistakersquo (Mencius 3)
This is surprising given that other languages eg Hebrew (Sharvit 1999) and Greek
(Alexopoulou 2006) which allow resumption in relative clauses do not employ a special
set of pronouns specifically for resumption in relative clauses Rather one set of
pronouns appears in resumptive and non-resumptive contexts alike
Finally Tingrsquos approach has no account of the fact that Archaic Chinese SUO
relatives are nominalizations This is particularly problematic for her (2005) analysis of
late Archaic Chinese She claims that the crucial difference between modern and
Classical Chinese is that SUO was obligatory for object relativization in the latter and not
in the former In order to account for the obligatory nature of SUO relativization in late
Archaic Chinese she claims that Archaic SUO is itself an operator and as such must
undergo further head movement to C at LF By assuming a CP layer in the SUO relative
she leaves open the question of how genitive case is licensed on the subject
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
52
6 Conclusion
This paper has proposed that the late Archaic Chinese morpheme SUO forms relativize
clauses on VP-internal positions by binding an operator in edge of vP The operator in
SUO relatives does not move to [Spec CP] because the vP is dominated by a
nominalized TP and DP layer No CP is projected SUO relatives are mixed projections
specifically a type of what Kause (2001) terms reduced relatives with genitive subjects
References
Abney Steven Paul 1987 The English noun phrase in its sentential aspect Doctoral
dissertation Massachusetts Institute of Technology
Aissen Judith 1996 Pied-piping abstract agreement and functional projections in
Tzotzil Natural Language and Linguistic Theory 143447-491
Aldridge Edith 2009 Old Chinese determiner Zhe In Historical Syntax and Linguistic
Theory ed Paola Crisma and Giuseppe Longobardi 233-248 Oxford Oxford
University Press
Aldridge Edith 2010 Clause-internal wh-movement in Archaic Chinese Journal of East
Asian Linguistics 19(1)1-36
Alexopoulou Theodora 2006 Resumption in relative clauses Natural Language and
Linguistic Theory 2457-111
Aygen N Gulsat 2011 Reduced relatives and the location of agreement California
Linguistic Notes 36(1)1-30
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
53
Baker Mark C 2005 On gerunds and the theory of categories Ms Rutgers University
Baker Mark C 2011 Degrees of nominalization Clause-like constituents in Sakha
Lingua 1211164-1193
Baltin Mark 1993 Negation and clause structure Ms NYU
Basilico David 1996 Head position and enternally headed relative clauses Language
73(3) 498-532
Belletti Adriana 2004 Aspects of the low IP area In The Structure of CP and IP ed
Luigi Rizzi 16-51 Oxford Oxford University Press
Bennett Paul 1981 The evolution of passive and disposal sentences Journal of Chinese
Linguistics 961-90
Borsley Robert and Jaklin Kornfilt 2000 Mixed extended projections In Syntax and
semantics vol 32 The nature and function of syntactic categories ed Robert
Borsley 101-131 New York Academic Press
Burzio Luigi 1986 Italian syntax A Government-Binding approach Dordrecht Reidel
Cao Guang Shun 1986 Zutangji zhong de lsquodirsquo(lsquodirsquo) lsquoquersquo (lsquoliaorsquo) lsquozhursquo [ lsquodirsquo lsquoquersquo
lsquozhursquo in the Zutangji Zhongguo Yuwen 3
Cheng Lisa Lai-Shen 1991 On the typology of wh-questions Doctoral dissertation
Massachusetts Institute of Technology
Cheng Lisa Lai-Shen 1995 On dou-quantification Journal of East Asian Linguistics
4197ndash234
Chiu Bonnie 1993 The inflectional structure of Mandarin Chinese Doctoral dissertation
University of California Los Angeles
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
54
Chiu Bonnie 1995 An object projection in Mandarin Chinese Journal of East Asian
Linguistics 477-117
Chomsky N 2000 Minimalist inquiries In Martin R Michaels D Uriagereka J
(Eds) Step by Step Essays in Minimalist Syntax in Honor of Howard Lasnik MIT
Press Cambridge MA pp 89-155
Chomsky N 2001 Derivation by phase In Kenstowicz M (Ed) Ken Hale A Life in
Language MIT Press Cambridge MA pp 1-52
Chomsky N 2004 Beyond explanatory adequacy In Belletti A (Ed) Structures and
Beyond The Cartography of Syntactic Structures Volume 3 Oxford University Press
pp 104-191
Chomsky Noam 2005 Three factors in language design Linguistic Inquiry 36 1-22
Chomsky Noam 2008 On Phases In Foundational Issues in Linguistic Theory Essays
in Honor of J-R Vergnaud Cambridge MA MIT Press
Ernst Thomas 1995 Negation in Mandarin Chinese Natural Language and Linguistic
Theory 13665-707
Feng Chun Tian 1991 Jindai Hanyu Yufa Wenti Yanjiu [Research on Early Modern
Chinese Grammar] Shandong Jiaoyu [Shandong Education] Publishers
Grohmann Kleanthes and E Phoevos Panagiotidis 2009 Mixed projections Categorial
switches and prolific domains Linguistic Analysis 35141-161
Gu Yang 1995 Aspect licensing verb movement and feature checking Cahiers de
Linguistique Asie Orientale 24(1)49-83
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
55
Hsieh Miao-Ling 2001 Form and meaning Negation and question in Chinese Doctoral
dissertation University of Southern California
Huang C-T James 1991 Remarks on the status of the null object In Principles and
parameters in comparative grammar ed Robert Freidin 56-76 Cambridge MA
MIT Press
Huang C-T James 1993 Reconstruction and the structure of VP Some theoretical
consequences Linguistic Inquiry 24103-138
Huang C-T James 1994 Verb movement and some syntax-semantics mismatches in
Chinese Chinese Languages and Linguistics 2587-613
Huang C-T James 1999 Chinese Passives in Comparative Perspective Tsing Hua
Journal of Chinese Studies 29423-509
Huang C-T James Y-H Audrey Li Yafei Li 2009 The Syntax of Chinese Cambridge
University Press
Kayne Richard 1994 The Antisymmetry of Syntax Cambridge MA MIT Press
Krause Cornelia 2001 On Reduced Relatives with Genitive Subjects Doctoral
dissertation Massachusetts Institute of Technology
Lee Thomas H 1986 Studies on quantification in Chinese Doctoral dissertation
University of California Los Angeles
Li Charles and Sandra Thompson 1981 Mandarin Chinese A functional reference
grammar Taipei Wenhe
Lobeck Anne 1995 Ellipsis Functional heads licensing and identification Oxford
Oxford University Press
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
56
Luuml Shuxiang 1943 Lun di3 di4 zhi bian ji di3-zi de youlai [On the differences between
di3 and di4 and the origin of d3] In S-X Lu Hanyu Yufa Lunwenji 51-58 Beijing
Kexue Chubanshe
Luuml Shuxiang 1991 Xiandai Hanyu babai ci [Eight hundred words in Modern Chinese]
Beijing Commercial Press [1980]
Miyagawa Shigeru 2008 Genitive subjects in Altaic In Proceedings of the 4th
Workshop on Formal Altaic Linguistics (WAFL-4) ed Cedric Boeckx and Suleyman
Ulutas 181-198 Cambridge MA MIT Working Papers in Linguistics
Miyagawa Shigeru 2011 Genitive subjects in Altaic and specification of phase Lingua
1211265-1282
Ohta Tatsuo 1958 Chugokogo Rekishi Bunpo [Chinese Historical Grammar] translated
by S-Y Jiang amp C-H Xu Beijing Univesity Press 1987
Paul Waltraud 2005 Low IP and left periphery in Mandarin Chinese Recherches
Linguistiques de Vincennes 33111-134
Postsdam Eric 1997 NegP and subjunctive complements in English Linguistic Inquiry
28(3) 533-541
Roberts Ian 1994 Two types of head movement in Romance In Verb Movement ed
David Lightfood and Norbert Hornstein 207-242 Cambridge Cambridge University
Press
Roberts Ian 1997 Directionality and word order change in the history of English In Ans
van Kemenade amp Nigel Vincent (eds) Parameters of morphosyntactic change 397-
426 Cambridge Cambridge University Press
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
57
Roberts Ian amp Anna Roussou 2003 Syntactic change A Minimalist approach to
grammaticalization Cambridge Cambridge University Press
Sharvit Yael 1999 Resumptive pronouns in relative clauses Natural Language and
Linguistic Theory 17587-612
Sun Chaofen 1996 Word-order change and grammaticalization in the history of Chinese
Stanford CA Stanford University Press
Tang C-C Jane 1990 Chinese Phrase Structure and the Extended Xrsquo-Theory Doctoral
dissertation Cornell University
Tang Sze-Wing 2001 The (non-)existence of gapping in Chinese and its implications
for the theory of gapping Journal of East Asian Linguistics 10201ndash224
Tang Yuming 1987 Han Wei Liuchao beidongshi luumlelun [Brief discussion on passive
constructions in the Han Wei and Six dynasties periods] Zhongguo Yuwen
198216-222
Ting Jen 1998 Deriving the bei Construction in Mandarin Chinese Journal of East
Asian Linguistics 7 319-354
Ting Jen 2003 The nature of the particle SUO in Mandarin Chinese Journal of East
Asian Linguistics 12121-139
Ting Jen 2005 On the syntax of the suo construction in Classical Chinese Journal of
Chinese Linguistics 33233-267
Ting Jen 2010 On the climbing of the particle suo in Mandarin Chinese and its
implications for the theory of clitic placement The Linguistic Review 27(4)449-483
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
58
Travis Lisa 1984 Parameters and effects of word order variation Doctoral dissertation
Massachusetts Institute of Technology
Tsai W-T Dylan 1994 On economizing the theory of A-bar dependencies Doctoral
dissertation Massachusetts Institute of Technology
Wang Li 2004 Hanyu shigao [Lectures on Chinese] Beijing Zhonghua Shuju [1958]
Wei Pei-chuan 1994 Guhanyu beidongshi de fazhan yu yanbian jizhi [On the
development and mechanism of change of the passive construction in Classical
Chinese] Chinese Languages and Linguistics 2293-319
Wei Peichuan 1999 Lun xian-Qin Hanyu yunfu de weizhi [On the position of operators
in Pre-Qin Chinese] In Linguistic essays in honor of Mei Tsu-lin Studies in Chinese
historical syntax ed Alain Peyaube and Chaofen Sun 259-297 Paris Center de
Recherches Linguistiques sur l Asie Orientale
Williamson Janis 1987 An indefiniteness restriction for relative clauses in Lakhota In
The Representation of (In)definites ed Eric Reuland and AGB ter Meulen 198-190
Cambridge MA MIT Press
Zhou Fagao 1963 Stages in the development of the Chinese Language Zhongguo
Yuwen Luncong [Collected papers on Chinese Language] 432-438 Taipei
Zhengzhong Shuju
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan
59
Primary Sources
Hanji Dianzi Wenxian [Electronic Corpus of Chinese Texts]
httphanjisinicaedutwindexhtml
Institute of Linguistics Academia Sinica Taipei Taiwan