UNIVERSITY OF CALIFORNIA San Di ego Interaction of Morphology and Syntax American Sign Language A dissertation submitted in partial satisfaction of the requirements for the degree of Doctor of Philosophy in Linguistics by Carol A. Padden Committee in charge: Professor David M. Perlmutter, Chair Professor Ursula Bellugi Professor Sandra l. Chung Professor Michael Cole Professor Edward S. Klima Professor Alan Timberlake 1983
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
UNIVERSITY OF CALIFORNIA
San Di ego
Interaction of Morphology and Syntax
American Sign Language
A dissertation submitted in partial satisfaction of the
requirements for the degree of Doctor of Philosophy
in Linguistics
by
Carol A. Padden
Committee in charge: Professor David M. Perlmutter, Chair Professor Ursula Bellugi Professor Sandra l. Chung Professor Michael Cole Professor Edward S. Klima Professor Alan Timberlake
1983
UNIVERSITY OF CALIFORNIA
San Otego
Interaction of Morphology and Syntax
1n
American Sign language
A dlssertatlon submitted tn partial satisfaction of the
requirements for the degree of Doctor of Philosophy
in linguistics
by
Carol A. Padden
Committee 1n charge: Professor Oavid M. Perlmutter. Chair Professor Ursula Bel1ugi Professor Sandra l. Chung Professor Michael Cole Professor Edward S. Klima Professor Alan Timberlake
19B3
Copyri ght by Carol A. Padden
1983
The dissertation of Carol A. Padden is approved. and it is acc~ptable in quality and form for pUblication on microfilm:
University of California. San Diego
1983 I
iii
To my parents who gave me the language and mucn more.
1v
1.1 1.2 1.3 1.4
2
2.1 2.1.1 2.1.2 2.1.2.1 2.1.2.2 2.l.2.3 2.2 2 " 2.3.1 2.3.1.1 Z. 3. 'l 2.3.2.1 2.3.2.2 2.4
3
3.1 3.1.1 3.1.2 3.1.3 3.1. 4
4
Table of Contents
Notational Conve.ntions ••••••••••••••••••••••••••••••• list of Figures •••••••••••••••••••••••••••••••••••••• Ac~nowledgements ••••••••••••••••••••••••••••••••••••• V1ta. Publications and Fields of Study ••••••••••••••• Abstract of the Dissertation •••••••••••••••••••••••••
Distinctions between grammatical categories •••••••••• Adjectives ••••••••••••••••••••••••••••••••••••••••••• liouns •••••••••••••••••••••••••••••••••••••••••••••••• Verbs ............................................... . ~?redicate~ as a gral'lmat leal re 1 at ion •••••••••••••••• Argurnents for predicat!"!hood ........................ .. Sp.ntential complements .............................. . Short anslOcrs •••••••••••••••••••••••••••••.•••••••••• Distinction between predicate and attributive adjectives ••••••••••••••••••••••••••••••••••••••••••• Resultative inflection ••••••••••••••••••••••••••••••• Adverbs •••••••••••••••••••••••••••••••••••••••••••••• Facial adverbs ••••••••••••••••••••••••••••••••••••••• Footnotes to Ch~Pter 4 •••••••••••••••••••••••••••••••
Subjecthood •••• I ••••••••••••••••••••••••••••••••••••••
Verb agreement ••••••••••••••••••••••••••••••••••••••• Two analyses of verb agreement ...................... . Agreement marker omission •••••••••••••••••••••••••••• FORCE-type verbs ••••••••••••••••••••••••••••••••••••• An alternative analysis •••••••••••••••••••••••••••••• Hoda Is ••••••••••••••••••••••••••••••••••••••••••••••• Arguments for stating the rule In terms of "subjectn. ArgL!r,lents for modals as predicates .................. . SELF pronouns •••••••••••••••••••••••••••••••••••••••• Test c~ses: Sentence-initial nominals ••••••••••••••• Modals ••••••••••••••••••••••••••••••••••••••••••••••• SELF pronouns •••••••••••••••••••••••••••••••••••••••• Subject agreement •••••••••••••••••••••••••••••••••••• Ambiguous sentences •••••••••••••••••••••••••••••••••• Possible analyses •••••••••••••••••••••••••••••••••••• Footnotes to Chapter 5 •••••••••••••••••••••••••••••••
6.1 Arguments against the source.goal analysis of verb agreement ••••••••••••••••••••••••••••••••••••••••••••
6.1.1 Backwards verbs ..................................... . 6.1.2 The class of Inflecting ... erbs •••••••••••••••••••••••• 6.2 Stating the verb agreement rule ..................... . G.2.1 Arguments for the advancement analysis ••••••••••••••• G.2.1.1 An argument for final 2-hood: Sign Order •••••••••••• G.2.1.2 An argument for initial 2-hood: locus shifting ...... 6.2.1.2.1 An argument against a thematic analysts of locus
shifting ••••••••••••••••••••••••••••••••••••••••••••• 6.2.1.3 The indirect object relation •••••••••••••••••••••••••
Footnotes to Chapter G •••••••••••••••••••••••••••••••
82 8; 86 8~ 90 9[ 95 n
100 101 le3 106 109
113
114 115 111 120 123 117 [28 131 m 137 139 gO 1'[ !42 [43 [46
149
[51 152 154 155 153 158 16[
167 111 113
ClaSsifiers and Indices .............................. 174
vi
7.1.1 7.1.2 7.1.2.1 7.1.2.2 7.1.2.3 7.1.3
7.2 7.2.1 7.3 7.3.1 7.3.1.1
7.3.1.2
7.3. I. 3 7.3.2 7.4 7.4.1
7.4.1.1 7.4.1.2 7.4.2 1.4.3
Arguments for clausehood of Noun + Class:ifier sequences •••••••••••••••••••••••••••••••••••••• ,...... 177 Sentential complements. ••••••••••••• •••• ••••••••••••• 178 Arguillents for predicatehuod of classifiers ........... 179 Short answers............................ .•.•.•••.••• 179 linear position of verbs •.•••••.•••••••••.•••••.••••• 181 Facial adverbs ••••••••••••••••••••••••••••••••••••••• 182 An argument for l~hoOd of the sentence-initial nominal: Modals ••••••••••••••••••••••••••••••••••••• 183 Sign order.... .•••.•••••••••.•••••••••••••••.•••••••• 184 Discourse ordering constraint........................ 186 Other complex classiffer sequences ••••••••••••••••••• la9 Other "locat ive object-subject~verb" sequences....... 190 An argument against 2-hood of sentence-initial nominal: Locus shifting.............................. 190 An argument against sentence_initial nominal as oblique: Topicalization ............................. 191 A proposed account .......... : •••••••••••••••••••••••• 194 "S-O-V" :iequences.................................... 197 Indices.............................................. 199 Arguments for the clausehood of some Noun + Index sequen(~s.. •••••••••••••••••••••••••••••••••••••••••• 201 Sentential complements ••••••••••••••••••••••••••••••• 201 Short answers..... •••••••••••••• ••••••• •••••••••••••• 202 An argument against predicatehood of some indices.... 204 519n order........................................... 207 footnotes to Chapter 7............................... 210
Since there is not yet a standJrd transcription system for ASL. the following notational conventions are used:
1) Signs are represented with English glosses in capitalized letters, e.g. CAT ('cat'), WOMAN ('woman'). When more than one English word 15 needed to represent a single sign, hyphens appear betneen them, e.g., DON'T-KNOW ('don't know').
2) fingerspelled words are represented with hyphens between the fingerspelled letters, e.g., C-A-R-O-l ('Carol').
3) Non-manual features "l'ihich appear simultaneously with the manual segment are represented in the overltnes. e.g. ---t--- (topic mar~ing) in the following example!
-t-{a} CAR. WOMAN BUY.
'A car. the woman bought.'
Other non-manual ~ymbols are:
(b) ---q--- yes-no question marking
(e) --whq-- WH-question marking
(d) ---n-- negative marking
(e) --when-- temporal clause marki ng
(f) --if -- conditional clause marking
(,) --hn-_ head nod
(h) --rc-- relative clause marking
4) A comma is used to represent a break in phrase timing, characterized by a ~ause and/or a hold on the preceding Sign.
5) For signs which involve articulation in a given locus position, the pOSition of the sign is indicated with a subscript whiCh prec~des the sign. These include the personal and locative pronouns (INDEX)' possessive (POSS) and reflexive (SElf) pronouns as well as certain nouns which can be marked in a particular locus pOSition. Iperson marking is indicated with the 1 subscript. the 2person marking. the 1
viii
subscript. and either locative points or 3person marking with letter ~ h ~ .... etc. For example:
(a) IINOEX ('I')
(b) 2POSS ('your')
(c) ISElF ('himself')
(d) IINOEX ('there')
(e) ICHAIR ('this chair there')
6} Agreement markers on verbs are rep,resented by subscrlpts~ For example. the lperson subject agreement marker on a verb like HATE takes the form of a subscript before the gloss, and the 2person object agreement marker. a subscript after the gloss:
ta) IHATE2 ('I~hate-you')
7) locative affixes on a class of verbs are indicated by subscripts. For example, the form of the verb GO moves between two locative points. The first locative pOint Is indicated by the subscript before the gloss and the second point, the subscript after the gloss:
(a) jGO j ('here-go-there')
8) Some verbs have only a single agreement marker or a single locative affix. In these cases, only a single subscript indicating the agreement m~rker or the locative affix appears before the gloss:
(a) tWANT ('she-wants! want(s)_1t')
9} Other symbols representIng lexical items or affixes include:
(a) CL: classifier
(b) D: derivational form
Example: Cl:3 ('classifier for vehicles')
Example: D:SJT[+noun] (derived noun)
(c) mult ('multiple plural Inflection')
(d) exhaus ('exhaustive plural inflection')
eel du (dual inflection)
(f) r (reciprocal inflection)
(g) pI (plural)
(h) fAg (Agentl,e suffix)
1x
10) Some symbols are combined, e.g.:
(a) rl.ZG1VE ('you-and-I~glve-to-each-other')
(11) Certain sequences in ASl Involve simultaneous articulation of one sign with one hand and a different sign with the other hand. Sb;/ns
"articulated with the left hand dominant appear on th'e line marked: "lhand", and those articulated with the right hand dominant appear on the line marked "R-hand":·' The representati on al so scherndti ca 11y i nd; cates the temporal sequencing of the two hands relative to the otherj for example, In (a) below, the two hands articulate PUT simultaneously. (An illustration of (1) appears in Figure 15.) In (b), the sign, i CL :4 appears following articulation of FENCE and Is held (d~picted by the
.dotted Hne) for the duration of the sign sequence: CAT jCl:V-CfWUCH.
lowe a great deal to several individuals and must in some small
way express my gratitude for their support through some of the ~ost
harrowing times of my lHe. This final culmination of five years of Iny
lHe "efl,ects In more ways than they realize. the impact of their
contributions.
First, 1 wish to thank two ind1viduals who were p~obably most
responsible_ for my coming to San Diego and for sl'stdining the ty[)e of
environment crucial for serious in-depth work In ASl research: Ursula
Hellug! and Edward Klima. Their constant demand for highly defined and
rigorous work In ASl has produced a generation of linguists of which I
am proud to be a part.
When I first arrived at UCSD, a nervous graduate student who
used a strange language. an excellent faculty tn tile Oepartmer,t of
Linguistics rose to the challenge and gave generously of their time and
wisdom. 1 am grateful for an uniformly competent and demanding faculty.
It is difficult to llAagine what my yelrs of gradUate study "ould
have been like without David Perlmutter. H1s influence on my thinUng
about the nature of human language and 1anguage structure Is clear
throughout these pages. His many unexpected and delightful ruminaf:ions
about ASl have led me into fruitful areas of research, and enriched my
understanding and appreciation for the tight complexity of human
languages. Above all. I must thank him 'for his several qualities ...-hlch
xlv
made all the difference for me: his honesty, unrelenting insistence on
clearer and better work, and unflagging support.
The following comrades, along with good cheer and excellent
ccnversation, turned consultant sessions into a refreShing break: Ben
Bahan, Carlene Canady-Pedersen, Venita Driscoll, Sue Hays, leslie
Jamison Hanaumi, Ella lentz. Pat Richey, Dennis Scherllenauer. Sam Supalla
and Jamie Tucker. For friendly commisseration and helpful advice on
aspects of my work. I thank Rick lacy, Ruth Loew, Richard Meier and
laura PettHo. In addition to members of my committee. Scott Liddell and
Robert Johnson reviewed earlier drafts of the dissertation and gave
valuable critical conunents.
Illustrations are by Frank. Paul, and display his usual clarity
and consistency.
I yratefully acknowledge financial support from the Salk
Institute. fellowships granted from the University of California, San
Diego, the National Science Foundation and the ford Foundation through
Hichael Cole and the laboratory for Comparative Human Cognition.
And finally, for reaSSuring me that eventually it would make
sense and that
therapist, and
I would survive, I thank my husband., best friend. . '
favorite entertainer, Tom Humphr1es.
XV
VITA
April 3, 1955 - Born - Washington. D.C.
1978 - B.A •• Georgetown University 1980 - M.A •• University of California. San Diego 1981 - C.Phil •• University of California, San Diego
PUBLICATIONS
Focusing on the non-manual components of American Sign language. (authored with Charlotte Baker). Understanding lan,uage ~hrou9~ ~ 19~Buase Research. edt by Patricia Slple. New lor: Aca em)c ress. •
A Basic Course in .American SiSn language (authored with Tom Hu~,phries and - lerren~o'U'rke). Sliver prlng. MD: TJ Publishers, Inc •• 1980.
Some arguments for syntactic patterning in Mierican Sign languillje. SiQn language Studies, 32, 239-259. Silver Spring, MO: linstok Press. ~.
Major field: linguistics Syntax:
FIELDS OF, STUOY
Studies in Professors langacker.
David Perlmutter, Sandra Chung, Edward Klima and Ronald
Studies 1n Phonology: Professors Matthew Chen. Jeffrey Elman and Sanford Schane.
Studies in American Sign Language linguistics: Professor Edward Klima
xvi
ABSTRACT OF TIlE OISSERTATIOH
Interaction of Morphology and Syntax 1n American Sign language
by
Carol A. Padden
Doctor of Philosophy in Linguistics
University of Cal1fornia, San Otego, 1983
Professor David H. Perlmutter, Chair
This study explores basic clause structure 1n American Sign
language {ASL}, a visual-gestural language used by Deaf people 1n North
Amerlca. Three major classes of verbs: Plain, Inflecting and Spatial,
are identified on the basis of differences 1n morpholog1cal form and
their interaction with various sYntactic phenomena including verb
agreement and s1gn order.
Previous analyses of ASL verbs have noted that verbs differ
markedly with respect to which morphology can be added to the verb stem.
However. no clear criteria have been offered for distinguishing between
Inflecting and certain forms of Spatial verbs whiCh have highly similar
/I'.orphology. The present analysis distinguishes between person and number
agreement morphology. which can be added only to Inflecting verbs. and
other morphological forms such as manner and location. This study
proposes that for Inflecting verbs (those that inflect for person and
xvii
number}. the verb agreement rule is best stated 1n terms of grammatical
relations.
The present investigation uses the framework of relational
grammar as developed by Perlmutter & Postal (1914. 1911. 1983) in which
grammatical relations such as ·subject·. Nd1rect object". etc. are ta~en
as primitives of lingu1stlc theory. Additionally, grammatical relations
are represented in this framework at more than one level of syntactic
representation. The multfpl~_ l~vels crucially interact with certain
grammatical phenomena. As well as accounting for facts about verb
agreement. the proposed structures provide, means for capturing other
syntactic and morphological generalizations in the grammar of ASl.
In relational grammar, upredlcate~ 15 a relational rather than
categorial notion. USing various diagnostics of predicatehood, cerl~in
adjectives and nomina1s as well as verbs are shown to bear the predicate
relation. In addition, these diagnostics show that some index1c forilis
and classifier elements are also predicates. Stating the slyn order rule
In terms of ·predicate" instead of "verb" c<!ptures a generalization
about the position of elements which bear the ?redicate relation. As a
consequence, s1gn order in ASL is demonstrated to be less fleKible than
previously thought.
xviii
r Introduction
Recent investigations into verb morphology In American Sign
languagE! (ASl). a visual-gestural language used by members of the Deaf
community in North America. have shown that verbs are among the most
complex and varied forms in the language (Fischer & Gough 1978, Bellugl
& Kl ima 1979, Supalla 1982). What is interesting is not only the
variety of verb affixes contained In different morphological categories:
person, number, aspect. locative. noun class. manner. among others, but
also that verbs differ with respect to which of the categories of
affhes can be added. Friednlan (1975). and others have proposed that
verbs can be grouped into classes on the basis of which affixes can be
6dded to them. For example. verbs like GIVE, HATE inflect for person and
number, but other verbs llke KNOW, LIKE do not. Supalla (1982) has
detailed the morphological structure of a class of verbs which he terms
·v'erbs of motion and location", these include verbs like CL:3-GO
('vehicle- move') and CL:I-WAlK ('person-walk') which mark for noun
class and manner among other categories.
Previous analyses of clause structure 1n ASl have characterized
clause structure and syntactic phenomena on the basis of verb
mor~holo9Y. In clauses containing verbs with person and number agreement
milrkf!rs, e.g. HATE. gram:natlcal relations such as "subject" and "direct
object" are determined by the fOrm of the markers that appear on the
vef'b (Friedman 1976). Kegl (1976) suggested that in clauses containin9
verbs ~hich inflect for person and number, sign order is more flexible
1
than in clauses containing verbs which do not. In the case of clauses
with verbs which lack agreement markers. e.g. KNOW, grammatical
relations are determined by sign order (Fischer 1975, Kegl 1976). Thus
according to these analyses, determining what is "subject" and ~direct
object M In a clause depends on verb agreement with respect to one class
of verbs and sign order in another class. And in cases of verbs of
motion and location, e.g. CL:3-GO ('vehicle- move') using either of
these criteria to determine grammatical relations is problematic since
these verbs also lack agreement markers and it appears that sign order
In clauses containing these verbs Is unl1ke that tn 'claus(!s containing
other verbs (Liddell 1917, McIntire 1980, Supalla 1982). As a solution,
some have proposed that thematic relations are mor~ appropriate as an
account of verb morphology and clause structure (friedman 1975).
The present study examines various cases of interaction of
morphology and syntax in ASl and proposes an account which differs trom
previous analys~s 1n that clause structure and sy~tactic phenomena 3re
not dertned in terms of verb agreement. or sigo order. ,but in terms of
grammatical relations_. Using the. fra1l1ew9rk of relational grammar
develop.ed by Perlmutter & Postal (lgH, 1977. 1983) in which grammatical
relations Such as "subject", "direct object". etc. are taken as
primitives of linguistiC theory, facts about syntactic phenomena,
including verb agreement .and s1gn order are accounted for in. a general
way.
Th1s approach has several cons_equences. First, by referencing
2
-----------~-------------
grarrmatlcal relations such as "subject" and "direct object" rather than
thematic notions, facts about the sYntactic behavior of nominals can be
generalized across clauses containing verbs of different classes.
Second. the notion of "predicate" as a grammatical relation allows for I
general statement of rules with regard to the similar behavior and
distrf'bution of predicate norninals. adjectives and verbs. These
arguments also point to a distinction between adjectives which bear the
prpdlcate relation and those that do not, a distinction which has not
been made In previous investigations. Third. the present investigation
argues that In order to state generalizations about verb agreement and
other phenomena In ASl, representing grammatical relations in more than
one level of syntactiC structure is necessary. An analysis which refers
to a Single-level notion of "direct object~ prov~s less insightful than
one which distinguishes between "lnitiaP and "final" direct object. And
finally. on the basis of arguments for the notions of "subject M•
"predicate u• and "initial direct object". 1 examine clauses which have
flexible or dtfterent sign order and show that facts about Sign order In
these clauses as well as other syntactic phenomena can be accounted for
by the same set of rules needed for other clauses.
This study IS organized In the following way. Chapter 1 presents
an overview of the framework central to this investigation, relational
9ra~nar. Chapter 2 contains certain preliminary Information about the
distinction between embedded and coordinate structures needed for
arguments in later chapters which crucially depend on the analysis of
certain multi-clausal structures as embedded. nQ.t coordinate. Chapter 3
3
reviews the case· for grouping verbs into different classes on the basis
of morphological features. introduce several criteria for
distinguishing between agreement morphology and other types of
formationally similar morphology. Since in Chapter 5 agreement
morphology is used in arguments for the notion of Msubject~. these
criteria are essential in identifying the correct cases for the
"adjective" and "verb". ~nd presents 'arguments for the notion of
"predicateM as a relational. not categorial notion. 'Chapters 5 and 6
detail the verb agreement data in ASl and present arguments for a verb
agreement rule wh1ch refers to "final" grall\'!latical relcltions. Chapter
analyzes clauses containing Inflecting and Spatial verbs.
r Chapter 1
Relational Grammar
The present study adopts the framework of relational grammar as
diSCUSSed by Perlmutter (1978, 1979), Perlmutter & Postal (1974,1977),
and a number of other works which are c1ted throughout the study. Basic
elements of the theoretical framework are outl ined below; only those
portions of the framework which bear on the grammatical description in
this study are discussed.
1.1 Basic concepts
The fundamental goals of linguistic theory are to characterize
the \.,alS languages are alike and the ways they differ. A major claim of
relational graflvnar is that generalizations about grammatical
constructions within a particular language and across the class of human
lanSuages are best stated 1n terms of grammatical relations.
Gral1~nat1cal relations are conceived of as primitives; they are not
defined in terms of other features such as word order. case marking or
phrase structure conflgurations. but ratper. determine these featUres.
1.2 Clause structure
In this theory, a clause involves a set of I1ngulst1c elements,
the grJmmatlcal relations these elements bear to the clause and the
level(s) at which the elements bear grammatical relat10ns to other
elements.
5
The inventory of grammatical relations include: the Predicate
relation, represC!nted with the "P" R-s1gn. "subject" ("1"). "direct
object" ("Z"). "indirect object" ("3"). The latter three relations are
called ~ relations. The 1 and 2 relations together comprise the
~ relations. Nonterm relations include the oblique relations
such as "Goal", "locative", "Benefactive", "Instrumental", etc.} and the
"Chomeur" relation. Finally, there Is the set of overlay relations, one
of which is discussed In Chapter 7: "Topic".
In order to highlight these features of clause structure, the
following elements are used:
(a) a set of nodes which represent primitive linguistic elements
(b) III set of "R~signs" representing the names of grammatical relations
(c) a set of coordinates which indicate the levels at which elements
bear grammatical relations to other elements
An "arc" consists of an ordered pair of nodes, one R-sign. and a
non~nul1 sequence of coordinates. ' It 15 a formal representation of the
grammatical relation ~ a given el:ement-! bears to another element
b at a certain level. S.:. Element' ,!.'is called the "head- of the
arc, and 2,. the tal1:
(1 )
GRx
a
As a concrete example, consider the follOWing representation: in
(2). the element ~ which bears the subject relation to clause ~.
·heads· a subject arc. with the R-sign. MIM. 1n that clause. Further.
WOMAN bears the 1 relation to the clause at level ~
(2)
A clause is the set at arcs which have a common clause node as
tail. In the following clause, (3). WAlT heads a Predicate arc (.pM),
and ~OMAN. a 1 arc. In the following representation, the elements have
the same coordinate. ~ or 1n other words, the structure contains a
single level or ·stratum·.
(3) WOMAN WAIT.
'The woman is waiting.'
A WA.f'f WOH-AN
Some structures contain more than one stratum. as in for example.
an "advancement" structure discussed in a later chapter. (4) is termed
a "3·2 Advancement" structure and involves elements which bear relations
to the clause 1n two different strata: c1 and c2. GIVE bears the
predicate relation to the clause 1n both the c1 and c2 strata; WOMAN
likewise heads a 1 arc 1n both strata, but BOOK and MAN head two arcs,
each in different strata. BOOK bears the 2 relation in the c11 or
~ stratum. and the Chomeur relation 1n the next, c2 stratum,
or the ~~. MAN heads a 3-arc in the initial stratum, and a
7
2-arc in the ttnal stratum. The representation 1n (4) 15 called a
"relational network".
(4) WOMAN jGIVE j HAN BOOK.
'The woman gave the· man a book.'
(5) below, termed a "stratal diagram" provides an equivalent
representation of (4) 1n an abbreviated form. Instead of noting the
coordinates at which an element bears a grammatical relation, the
representation of levels is shown with a visual analog. Other ele.;lents
in the clause such as adverbs, aspectual markers, etc. are not l~cluded
in the diagrams in order to highlight the relevant elements. For the
most part, diagrams such as (5) will be used throughout this study:
(5) WOMAN jGIVEj MAil BOOK.
'The woman gave the man a book.'
HhN
1.3 Constructions, rules and laws
Previous investigations within the relational grammar framework
have proposed a set of "well-formed" relational networks. or the set of
possible grammatical constructions 1n human languages. These include
B
Passive constructions (Perlmutter & Postal 1977). 3-2 Advancement
(Aissen 1979. Crain 1979, Chung 1976), and Inversion (Harris 1981.
Davies 1981) among others. The grammars of specific languages state
which subset of the range of possible constructions are found in the
language. For example, since there 1s no ev1dence for the Passive
construction in ASl, but evidence for 3-2 Advancement. only the latter
construction ;s included in the subset of well-formed constructions in
the gl"ammar of ASl.
Rules in a grammar are well-fonmedness conditions on relational
networks. For example, the condition on antecedents of SELF pronouns in
ASl, snown below in (6), slates that only relational networks which meet
the conditiOn are wel1-fonmed.
(6) ~ pronouns
Only nomlnals which head a I-arc may be antecedents of SELF pronouns.
In addition to language-specific rules, ~ are fonmulated as
wel1-formedness conditions on the class of possible relational networks.
The la,,;s which bear on the present 9ra~atlcal description of ASL are
discussed b~low. These are:
(.) (bl (c
Stratal Uniqueness law Mutivated Chomage law Fina 1 1 law
The Stratal Uniqueness law (Perlmutter & Postal 1977. 1983) 05
stated below infonmally. rules out constructions such as (8) 1n which
two nominals head subject arcs in the same stratum:
9 10
(7) ~ Uniqueness ~
At most one nominal can bear a given term R-s1gn in a given stratum.
'0
(8)
c.
Returning to the earlier example of a multi-stratal structure,
(4), repeated below. we See that the nominal bearing the 3-relation at
level 'I advances to bear the 2-relation at the next level, c2-
Since this would violate the Stratal Uniqueness Law, the nominal which
bears the direct object relation at level c1 cannot bear this relation
at level '2' but instead must be placed "en chomage"; it bears the
~ relation. The following law restricts the domain of the
chomeur relation to certain contexts.
(4 )
The Motivated Chomage law (Perlmutter & Postal 1977, 1983), stoClted
informally, reads as follows:
(9) Motivated Chomage law
If a clause contains a nominal a which heads an arc with the tenn R-sign n in stratum c1 and-a ChoRleur arc in stratum c1+I, then tnere must be a nominal ~ which heads an arc with R-s go ~ in c1+1 o where all arcs have the same tan.
The effect of (9) Is to rule out cases such as (10) In which the
nominal BOOK heads a direct object arc In stratum c1 and a Chomeur arc
In c1+1" but there 1s no nominal which heads the direct object arc In
that stratum. Thus. the law restricts the occurrence of the chomeur
relation to only those contexts where there would otherwise be a
violation of the Stratal Uniqueness law.
(10 )
According to the Final 1 Law (Perlmutter & Postal. 1983) as
stated below. (12) is not a well~fonned relational network:
(11) £.!!!.!l 1 law
Every basic clause"must contain I l-arc In the final stratum.
(\2 )
1.4 Why relat10nal grammar?
A primary goal of the theoretical approach of relational grammar
Is the formulation of universal properties of human languages. A central
concept of relational grammar is that generalizations about syntactic
structure within individual languages and across languages are best
captured in terms of grammatical relations. Much of the work
~ ..
11
contributing to this goal Is based on examination of a variety of
typologically diverse languages. From thts data, specific proposals have
been made within the framework concerning those properties which
languages share, and the ways languages differ.
Relational grammar provides a means for examining ASl syntax in
at least two ways: first. the characterization of clause structure 1n
terms of gralMlatical rel'ations facilitates comparison bet~leen ASl Clnd
other languages with different language-particular featur~s. including
different modality (i.e., spoken or signed). Second, ..... hile the theory
brings out new problems 1n the analysis of ASl syntax. the theory also
leads to poss1ble solutions for syntactic phenomenon in ASl. As an
example. it has been noted that some ASl sentences lack verbs, but other
constituents in the sentence behave in certain respects Mlike verbs". In
"elat10nal grammar. the not10n of "Predicate" as a grammatical relation
has provided a means of capturing significant generalizations about the
similar behavior of verbs and other constituents 1n certaln sentences.
In this way, the theory points to a solution for syntactic phenom~non in
ASl as well as to depict the ways in which phenomenon in ASl resemble
those in other languages.
fORC(, l(AS(, l(ACH, SE.ND. AS'!:.. iE.ll. "
" A c.onta 1fIS verbs like: GIVE. thiS cl asS are
fl£FE~!". BLAME .. BORROW. COPY. and
HlVllE. ("Ppend u "
crl.!
\2
Chapter 2
Verb Classes
Previous analyses of ASl verbs have noted that verbs differ
markedly with respect to which morphology can be added to the verb stem
19HZ}. Inflections for person and number are added only to certain verb
stems, markers for location and manner can only be affixed to another
set of verb stems, and verbs of yet another set are more restricted in
the range of possible affixation. The present study re~examines ASl verb
morphology and identifies three major classes of verbs on the basis of
gener;;l morphological characteristics.
Chapter 2 discusses the morphology of the three verb classes:
~!nflecting·. ·Spatial· and uPlain-. Inflecting verbs, unlike the other
two classes, mark for person and number. Spatial verbs mark for
location and position, and a sub~class marks for path and manner of
movement. In contrast, Plain verbs do no~ mark for these categories.
Chapters 4·7 discuss in further detai 1 the syntactic characteristics of
verbs in these classes, inclUding verb agreement on inflecting verbs and
shifting ot nominal locus pas1tion in inflecting and spatial verbs.
2.1 Inflecting verbs
Verbs of this class inflect for perSon and number. Included in
13
this class are verbs 11ke: GIVE. SEND, ASK, TELL. FORCE. TEASE, TEACH,
SElL, DEFEAT. BLAME, BORROW. COPY, and INVITE. (Appendix A contains
additional examples.)
As can be seen from the illustrations of verb forms in sentenc~s
(1-2), the position of the beginning point of the sign varies depending
on whether the person of the subject of the clause is Iperson (I) or
2person (2) and in sentences (2-4) the end point of the sign also varies
in pos1tion depending on the person of the object. In {5-7}. it can be
seen that- as the number of the subject and object varies. tne beginning
and end points likewise change in form. (Illustrations of verb forms
appear in Figure 1.)1
(I) IINDEX IG1VEj BOOK.
'I gave him the book.,'
(2) 21NOEX pVEj BOOK.
'You gave, her the book.'
(3) IINDEX IGIVE2 BOOK.
'1 gave you the book.'
(4) jlNDEX iGIVEj BOOK.
'She gave him the book.'
I'
k .
'I gave you both the book.'
(6) IINOEX 1G1VE"exhaus BOOK.
'I gAve each of them the book.'
(7) l,dulNOEX l,duGIVEI BOOK.
'The two of them gave me the book.'
---Figure 1---
These verb forms cons1st of a linear movement (the stem). with
agreement markers at either end; the beginn1ng point of the s1gn 1s the
subject agreement marker, and at the end point. the object agreement
marker.
Not all inflecting verbs contain a linear movement between two
distinct points of locat1on. Certain inflecting verbs. for example.
WANT, FLUNK, ARREST inflect for either the subject (WANT) or the direct
object (FLUNK. ARREST. WANT). [n these cases, tne form of the verb lacks
a linear movement, and the sign 1s articulated in a single location
(Fischer & Gough 1978. Heier 1982). Despite these differences.
morphologically these verbs behave as inflecting verbs except that
agreement 1s expressed with only one nominal. 2
Inflections for person and number. as we will see in the
following discussion, consist of discrete and specific morphological
15
forms which are added to the verb stem.
2.1.1 P.erson agreement
The forms of. the person agreement markers are as follows:
Iperson: near signer's body 2person: in direction of addressee as determined bY3eye contact
with addressee. either real or discourse-marked 3person: the agreement fIlarker will have the Sil(l1e 10cus ;>oint t in
neutral space as the assigned 3person nominal locus point I·
While the form of the Iperson marker 1s fixed (a position near
the body of the signer). 2person and 3person markers pote.'\lial1y ha'/e an
Infinite number of possible locations. hi addressee can be positioned
in any location relative to the signer. accordingly a 2person dgreement
marker 15 likewise variable. As for 3person. the particular agreement
marker form used is dependent on a number of conditions, only a few of
which will be discussed here. For example. in (8-9). the ~ub';Qct is
followed by an INDEX to a particular locus point (il. as such. is
"assigned" that locus point. Subsequent agreement wah the subject win
have the same locus point. A noun such as DOG in (10) can be articulated
1n a particular location in place of an INDEX, and the object a'Jree~lerlt
marker in FEED has the same locus point as DOG. (Note: In the glosses.
the subscripts, .!.z. h.h ... are used to specify any distinctive locus
point in sign space. Henceforth, elements with the same subscript have
the same locus point.)
(B) WOMAN ,INDEX, ISEII01 FLOWER.
16
'That woman there, she sent me flowers.
(9 ) 'DOG jlNOEX i81TEj CAT.
'That dog there bit the cat. I
~t--
(10) 100G, iBlTEj CAT.
'That dog there bit the cat.'
The selection of a particular locus point for a 3person UmEX 1s
determlnpd by several factors. among them: the real, or supposed
location of the 3person referent, e.g. the woman being referred to in
(9) normally occupies an office 1n the general vicinity of locus point
(i), or an arbitrary locus point is chosen. Selection of a locus point
for 3person agreement is constrained as follows: coreferential nominals
are assigned the same locus point. Consequently. non-coreferential
nominals are usually assigned. different locus points.
The possibility of more than a frw 3person agreement markers 1s
not unusual for human languages. Southern Tlwa, for example, has
nUlTn:!rous 3person forms (Allen & Frantz. 1978). But what is perhaps
unique in the case of ASl 3person agreement is the potentially enormous
number of possible agreement forms. i.e •• any point in the signing space
around the signer's body 1s available for 3person agreement. However. no
more than possibly four or five dlfferent 3person locus points are used
at one point in the COnversation and new 3person locus points are
17
selected when the topic of the conversation changes, or the time
reference of the event changes. Additional discussion of some other
interacting conditions which determine selection of locus points for
3person agreement can be found 1n Chapter 6.
Although both 2person and 3person agreement forms have available
any locus point within s1gning space, distinctions between 1. 2 and
3person agreement forms are clear. as outlined above. Sentences in which
an inflecting verb's marking does not agree with the subject or object
are ungrammatical as in (11) where the SUbject is Iperson, but the
agreement marker is 3person and (12) where the direct object 15 2person
but the agreement marker is Iperson.
-----hn----(II) *IINDEX JASKj WHERE, WILL IINOEX.
'I'll ask where 1t 1s.'
-----n----(12) ·CAREFUL. MAYBE iASKl 21NOEX, NOT I1NDEX.
'Be carefull He might ask ~. not me.'
2.1.2 Number agreement
Number agreement may be either unmarked. for singular or
collective plural as 1n (13-14). or marked for dual. trial (three) or
more than three. In number marked forms. distinct morphological units
for number which take the form of characterist1c movements are added to
1 18
the stems as shown in the following sections 4(Kl1ma & Bellug1 1919}.
(13) II"DEX IASK, SISTER ZI"DEX GO.
'I asked ~ sister if she was going.'
. (14) 'pi INDEX ,HATEj jpIINDEX.
'They hated them.'
2.1.2.1 Dual
The dual number mark1ng has either one of the following two
forms:
(i) The verb stem is executed tw1ce. with the inflected end poInt displaced the second time.
or: (il) The verb stem 1s doubled to a two-handed form and executed either: (a) simultaneously or (b) twice 1n sequence.
-------t-------(15) SISTER. BROTHER, lASK i •du '
'I asked my sister and brother.'
---Figure 4---
While either form 1s poss1ble for dual objects, only the second
form (il) is possible for dual subjects.
(16) BOTH CIIILOREN '.duASKI SIMULTANEOUSLY.
'Both of the children asked me at the same time.'
19
---Figure 5---
Sentences like (17-lS) 1n which the number of the subject 1s
greater than two but the agreement marker 1s dual. are ruled out •
(17) *5 CHllDRErC 1•duASK1 SIHU,lTANEOUSlY.
'The children asked 'me all at'"the same time.'
----t----(18) 'CROWO DOG, IFEEO"du'
II fed scads of dogs.'
Another form of the dual 1s the reciprocal; the basic
morphological form ts similar to the dual with the following
distinction:
Reciprocal
The doubled dual form in which the end po1nts of each one-handed form either (a) are adjacent. or (b) have the same agreement Darker as the other's beginning point.
II waved (a paper with) the telephone number in the air.1
These examples demonstrate that while the two verbs appear
similar 1n form. they belong to different morphological classes: GIVE is
an Inflecting verb. but CARRY·BY.HANO 1s Spatial. Consequently.
CARRY-BY-HAND does not inflect for person or number. but instead marks
for location In a manner exclUsive to Spatial verbs. In addition to a
morphological distinction. the two forms are different forrnatlonal'y tn
one respect. Although both forms have identical handshapes and identical
linear movern~nts, the orientation of the hand in the two forms 1n
(58-59): tGIVEj and iCARRY-8Y-HANOj is different. In
iG1VEj' the palm is oriented from the direction of the 3person
locus. 1. whereas with the CARRY-BY-HANO form. the palm 1s or1ented
from the direction of the Iperson locus. In both these examples, the
palm 1s oriented from the direction of the subject of the sentence. From
these and similar examples. it can be seen that palm orientation plays a
role 1n verb morphology. an observation which has been made by oth~rs
with different examples (Fischer & Gough 1978. Meier 1982). Since this
Is a potentially interesting area of verb morphology which I do not
treat here. I mention 1t as an area for further investigation.
Another pair of formational1y similar verbs: DRIVE and DRIVE-TO
39
.~'
carl also be demonstrated to be dissimilar morphologically. DRIVE-TO is
Spatial but DRIVE exhibits none of the characteristics of Spatial verbs.
1nstead 1s morphologically Plain. DRIVE (see Fig·ure 21) cannot add
person or number inflections. nor can it indicate spatial location.
---1f--(61) MY-TURN, llNOEX ORIVE_
'If you'll let me. I'll drive~'
(62) llNOEX HURRY 1DRIVE-TOj' jARRIVE ON-TIME_
'I rushed to drive over there and arrived just 1n time.'
---Figure 21---
These pairs of formatlonal1y similar but morphologically distinct
verbs point to a tripartite system of morphological verb classes: Plain.
Inflecting and Spatial. Although two verbs may be quite similar in
form. their morphological class limits the range of possible affixes.
The above characterization of differences between Plain.
lnqecting and Spatial verbs point to an 'Important feature
d1stinguishing the three classes: the different ways tn which verbs of
each class exploit the spatial dimension in sign. Plain verbs occupy an
articulatory space. the sign is articulated in a fixed location. e.g.
CELEBRATE: neutral space. THINK: on the forehead. LIVE: on the chest.
In contrast. Inflecting and spatial verbs vary their location in space,
but fn significantly different ways. Inflecting verbs vary their
10cat10n 1n space depending on the categor1es of person and number of
40
'-.:...:,;..-
the subject and object, categories to which Spatial verbs are oblivious.
Instead, Spatial 'Ierbs mark for location, exploiting a more detailed
spatial dimension in which all spatial points around the signer's body
are avallable.
~l .-J.
Footnotes
1 The terms "subject" and "(direct/indirect) object" are useD here only as labels to facilitate description of the 'Ierb forms. A5 mentioned 1n the ~ntroduction to the chapter, a syntactic account of verb morphology. including grammatical relations and 'Ierb agreement 1s d1scussed 1n Chapter 4-7.
2 I will not attempt here to characterize beyond an informal description the morphophonological structure of the different Inflecting 'Ierbs; Meier (l980) and Liddell (1982) discuss several cmalyses of these verbs.
3 In direct speech. e.g. "So t said to him: 'What do you want?''', the signer can address an imaginary 2person referent.-other than the "real II or "present" 2person addressee. In these cases, the 2person form 1s eye contact directed toward an imaginary addres~ee.
4 Se'leral rnarkings for plurill are not discussed in the sections below, including: apportionative, seriated and allocative (Klima & Bellugi 1979}: as with the number forms discussed below. each have distincti'le forms added to the verh stem. for the present purposes of examining number inflectional marking on ASL verbs, it will be sufficient to limit our discussion to the following number forms: dual, reciprocal, exhausti'le, and multiple.
5 When 'Ierbs are_ inflected for plural object, subject agreement markers are typically omitted (indicated by the 0 subscript). further discussion of omitted subject agreement markers can be found in Chapter 5.
6 It is possible to accompany a Plain verb with a simultaneous index, and shift the body 1n the direction of each index, e.g.: e.g. :
(1 ) L-hand: t JUDEX
R-hand: FORGET
j INDEX k ~tl~~X ,
FORGET FORGET.
'He forgot and she forgot and he forgot (too).'
These forms appear not to be "inflected" but rather, a type of seriated verb construction in which the indices are used Simultaneously with the verb. Tests for number inflection discussed in this chapter involve fixed body pOSition and no simultaneous index.
7 Liddell (1977:15D-152) has made a similar observation regarding the distinction between verbs which 1 term here as "Inflecting M and MSpatiaP.
IG1VEj
(I) IINDEX IGIVEj BOOK.
'I gave him a book.'
ZGIVE j
(Z) ZINDEX ZGIVE j BOOK.
'You gave her the book.'
Figure 1
43 44
IG1VEZ
(3) lINDEX IGIVEZ BOOK.
II gave you the book.'
i GIVE j (4) tINOEX tGIVEj BOOK.
'She gave him the book.'
Figure 1 (continued)
45 46
IGIVE 2du (5) lINOEX IGIVE 2du BOOK.
II gave you both the book.'
i,duG1VEl
(7) l,dulNDEX i,duGIVEj BOOK. 'The two off them gave me a book.'
IGIVE; .exhaus
(6) jINOEX jGIVEi, •• haus BOOK.
'I gave each of them the book.1
Figure 1 (continued) Figure 1 (continued)
DOG tiNDEX
(9) DOG tiNDEX tBITEj CAT.
'That dog there bit the cat,'
--t--i DOG
--t--(10) tDOG. tBrTEj CAT.
'That dog there bit the cat,'
Figure 2
47
IINDEX
(11) 'lINDEX tASKj WHERE. WILL liNDEX.
'I'll ask where it is.'
·~-n------(12) 'CAREFUL. MAYBE tASKl 2INOEX. NOT liNOEX.
'Be careful! He might ask l£!!.. not me,'
figure _3
49
50
-------t------- i ,duASKl (15) SISTER, BROTHER, lASKi,du' (16) BOTH CHILDREN i,duASKl SIMULTANEOUSLY.
'I asked my sister and brother.' 'Both of the children asked me at the same time.'
'Money. one company gave him a trophy and another gave him.'
Topical ization provides yet another argument for the distinction
between err,bedded and non-embedded structures in ASL: a topical ized
subject or direct object of an embedded. but not a non-embedded clause
may appear in initial position.
3.1.4 Conjunctions/Discourse markers
Certain lexical items such as BUT, MlO, FINISH ('then'), \rIElL
('so'). PLUS ('and') and other non-manual markers such as ---head nod
--- ('and then') cannot appear between a matrix and embedded clause.
While many of these Signs translate as conjunctions in English,
additional evidence distinguishing between coordinate and poly
sentential structures will determine whether these are mare
appropriately termed conjunctions or discourse markers.
(41) *lPERSUAOE 1 BUT CHANGE MIND.
'I persuaded her but to change her mind.'
(42) 'PERMIT j FINISH WASH CAR.
'I let her then wash the car.'
However, there are sentences similar to (41-42) which are judged
gralTlAlatical, shown below as (43). In cases of sentences I ike (43), the
sentences have dHferent non-manual markers. A pau~e typically ajJpear~
between the clauses and the second clause may be preceded by a sharp
head shake.
(43) lERSUADE p BUT CHANGE MIND.
'I persuaded her to do it but then I/she/he changed my mind.'
(44) IINDEX PERMIT i • FINISH WASH CAR.
'I permitted her to do. it then I/she/he went to wash the car.'
1
J
t ............................. .
It can be shown with any of the above arguments that (43-44) are
not embedded structures; (45-46) shows that when clause! is negated.
a negative marking cannot appear over both clauses for the same
meaning. As discussed in Section 2.1.2. the fact that a negative
marking cannot appear over both clauses in (45-46) demonstrates that
they are not embedded structures.
---n---(45) PERMIT1 fINISH WASH CAR.
'I didn't let her do it so then she went to wash the car.' ______ n __ _
(46) IPERSUAUE l BUT CHANGE MIND.
'1 didn't talk her into anything but she changed her mind
(anyway) •
Additionally. a subject pronoun copy of the first clause cannot
appear following the second clause in these sentences.
(47) 'I'ERMIT1 fINISH lINDEX WASH CAR IINDEX.
'I let her so then she went to wash the car. I did.'
(4B) 'ZPERSUADE 1 BUT iINDEX CHANGE MItlD ZUIDEX.
'You talked her into it but she changed her mind. you did.'
Further, 1n sentences like (49~50). the subject or d1rect object
of the second clause cannot be topicalized and appear 1n lnitial
position. an additional argument against the analysis of these
\ ,--,.
79 80
structures as embedded.
_~t-
(49) 'CAR, PERMIT1 FINISH 1 INDEX WASH.
'The car. I let her so then she washed it.'
--t--(50) 'HOUSE, 2'ERSUADE 1 BUT IItlDEX BUY.
'The house, you persuaded me but I bought it anyway.'
With four dlffere~t arguments. we have thus far demonstrated the
existence of embedded structure's In ASl. Apart from the information
this provides about interna) structure of ASl sentences wllleh lilek
su.rface markers. the tests are essential in determining clause structure
of bi-clausal sequences in ASl.
.1
i\
.1
I I 'I
Footnotes
1 The discussion of topicalization 1n this section ;s limited to the issues of the argument; further discussion can be found 1n Chapters 6 and 7.
2 Although the examples used for the argument include only those 1n which the topicalized constituent appears in sentence-initial posit1on of an embedded structure. some signers will allow topicaltzed nomlnals in clause-initial position at any level of embedding. e.g.:
'I hope my sister manages to persuade my mother. the exercise class. to take it.'
3 A nur.lber of non-manual features appear to mark boundaries between claus~s. such as a pause and movement of the head. It is unclear to me whether specific features can be identified as preceding only embedded as opposed to conjoined clauses. Further description of non-manual components 1s necessary before these markers can be used as arguments d1stir.~uishing one structure froln another.
Chapter 4
Predicates
The preceding chapter discussed morphological properties of one
grammatical category of s1gns 1n ASL: verbs. In this chapter.
properties of -nouns·, ~verbs·1 and N adjectlves· are examined. and
several characteristics emerge which determine category merr,bership of
s1gns. These are discussed in Section 4.1. Sections 4.2 and 4.3
examine data wh1ch show that category membership alone cannot account
for cases of signs which are members of different categories, but
display similar syntactic characteristics. 1 argue that in order to
account for this data, the not10n of "predicate~ as a grammatical
relation is needed. Several arguments are brought to bear on the
distinction between grammatical category and-grammatical reldticn,
showing: (I) some rules refer to category only. (2) others refer to
grammatical relations and (3) yet others refer to both category and
grammatical relations. This chapter addresses the question of
morphological and syntactic properties of ASL as well as interaction of
both.
4.1 Distinctions ~etween grammatical categories
Determining whether a lexical item is. for example, a ·noun K as
opposed some other category such as ·verb- or -adjective- involves
identifying properties 1nherent to nouns which are not shared by members
82
of other classes. These properties of category membership are
determined to be invariant; lexical items which are "nouns" exhibit the
same properties of "nouns" regardless of which clause they appear in.
With respect'to ASl. distinctions have been made between
semantically related pairs of nouns and verbs. e.g. CHAIR - SIT-DOWN,
AIRPLANE _ FLY. etc. (Supalla &- Newport. 1978) on the basis of
differences in surface movement forms: while verbs can have either
single or repeated movement. the noun counterpart of the pair 1s always
repeated and restrained. Klima & Bellug1 (1979) have also identified
movement differences between nounS and derived adject1ves. However.
determining category membership on the basis of form alone is
inconclusive in contexts where differences 1n form between certain
verbs, nouns and adjectives become neutral lIed. This problem 1s similar
to relyln9 exclusively on the appearance of the '-1ng' inflection to
distinguish between verbs and adjectives in English. In addition.
derivational nlorphology 15 not found on many s1gns. thus a contrast
between different categories is absent.
Consequently some other determinant of category membership is
needed which 1s not dependent on form alone. In the following section,
several determinants of category membership are examined which support a
distinction between the grammatical categories of noun, verb and
adjective in ASL. Although the present study Is limited to distinctions
between these three categories. this is not to say that other categories
such as adverbs and prepositions do not exist In ASL. but In the
~- .. ~.~---"-----------
83
interests of focus. Identifying properties of these categories 1s
postponed to a later study.
4.1.1 Adjectives
Only adjectives may inflect for the -Intensive- aspect ('very').
The form of the inflection Is the addition of a characteristic movement
s~quence to the stem: a 10n9 tense hold at the beginning of the sign and
~ very rapid release to a final hold. For example. the uninflected SICK
has a short lax movement toward the head. but when inflected for
intensive. the form changes to a larger movement as described above.
(I) LAST-WEEK IINOEX SICK.
'I was sick last week.'
(2) LAST-WEEK IINOEX SICK[+intenslve).
II was very sick last week.'
All the underlined signs in the following sentences may b~
1nflected for 1ntensive aspect, thus they are argued to be ·adjectives~.
(3) UP-TO-NOW WOMAN SICK/SICK[+intenslve].
'The woman has been s1ck/very s1ck for some time.'
(4) BOOK HAVE BLUE/BLUE[+;nt.n,!,.) ON-BACK-SURFACE.
'The book has a blue/bright blue cover.'
I
J
(5) GIRL THIN/TIIlN[+intenslve].
'The girl was thin/very thin.'
(6) LAND ilNOEX FLAT/FlAT[+intensiveJ.
'The land there 1s flat/very flat.'
The configuration of the intensive inflection 1s not to be
confused with other modulations on base forms which also involve
addition of a (+tense] feature, such as emphatic stress. The emphatic
form can be distinguished from the intensive in the following respect:
emphasis is added to the surface form. but not to the stem, as in the
case of the intensive. Compare the surface uninflected, non-emphatic
form of REO in (7) with (8) and (9) below; (8) involves the
characteristic movement contour of the intensive form: a single sharp
release from a tense beginning to a final hold, in (9) REO has
tenseness, but the repetition of the surface form remains and tenseness
is distributed evenly across the execution of the movement, tn contrast
to the sequence described for intensive ~spect.
(7) IINDEX BUY TOP-GARMENT RED.
'1 bought a red top.'
--when-(B) SUN-SET. SKY RED[+lnt.nslv.]_
'Wh1le the sun was setting. the sky was bright red.'
In an analysis such as (65) where there is no relational
distinct10n between adjectives which appear after the noun and those
102
that precede the noun. f.e •• both appear 1n nominals. an ad~hoc
statement about the resultatlve inflection would be needed to account
for (70-11). If. however. there were grounds for positing a relational
distinction between the two adjectives. the facts about (70-71) would
follow from this distinction.
Having shown that adjectives which precede and follow the noun do
not beha~e alike with respect to affixation of the resullatlve
inflection, we tUrn now to further arguments which support the structure
presented in (66) as an account for (71).
4.2.2.2 Adverbs
The adverbs, including! FAST ('quickly') EASY, ALWAYS precede
verbs 1n a verb phrase, but not nouns:
(72) MOTHER ALWAyS HOPE 1,jduiNOEX MARRY.
'Their mother is always hoping they'll marry.'
------t~ __________ _ (73) KrmW lPOSS 8ASEMENT, WATER FAST SEEP-THROUGH-WALL.
'You know, 1n my basement, the water very quick.ly gets
through the walls.'
--------t-----------(74) "K"OW IPOSS 8ASEMENT, FAST WATER SEEP-THROUGH-WALL,
(75) HORSE FAST EAT-Orr-SURFACE GRASS.
103
'The horse quickly ate all the grass.'
(76) -HORSE EAT-ON-SURFACE FAST GRASS_
In the following pair. of sentences, EASY can precede the
adjective in (77), (79) but not in (78), (80):
(71) ilNOEX MAN ALWAYS MIND-WARPED.
'The man 15 always crazy.'
(7S) • t INDEX ALWAYS MIND-WARPED MAN.
'He's a continually crazy man.'
(79) i,plINOEX PEOPLE ALWAYS HUNGRY.
'Those people are always hungry.'
{SO} *1,PlINOEX ALWAYS HUNGRY PEOPLE.
'They're continually hungry people.'
If (7S) and (77) are analyzed as having the following structure,
1n which the adject1ve bears a predicate relation:
IN
J
(81) HORSE FAST EAT-ON-SORFACE GRASS.
'The horse quickly ate all the grass,'
(82) ;INDEX HAN ALWAYS MINO-WARPED.
'The man is always crazy,'
1\ HINO- • n.JOE"1-. HI'-.N WMtPE"O L
and order of const1tuents in the ASl sentence is ordered: Subject _
Predicate, then an account of the grammatical and ungrammat1cal strings
in (72-80) as well as (70-71) can be accounted for by the following
statement:
(83) Adverbs
Adverbs precede any constituent but nomlnals.
The statement effectively exclUdes nominals, including those containing
attributive adjectives such as (78) and (80). but under an analysis of
adjectives in (77) and (79) as predicates, the fact that the statement
does not exclude adverbs from preceding predicate adjectives 1s
correctly predicted by the structures. (81-82) and the s1gn order'
105
constra1nt.
4.2.2.3 Facial adverbs
A certain set of facial expressions, illustrated below along with
their notation symbols. appear concurrently with a specified range of
manual segments. As can be seen from the Hlustrations, each facial
expression can be .identified by particular configurations of the mouth
area, eyebrows. and head position. L'iddell (l978) has demonstrated that
the appearance of these facial expressions 1s not a "comment- by the
Signer about her feelings or emotions concerning the content of
discourse, but instead are grammatical modifiers which he calls -faClal
adverbs".
---th--- 'carelessly. without attention'
---mm--- 'leisurely. without effort'
---em--- 'c~reful1y. with attention'
The range of manual segments with which facial adverbs co-occur
is 11mited 111 the f.ollowing ways. fi,rst. while they can co-octur with
verbs. they cannot with nouns:
--mm-(84) liNDEX DRIVE.
'J was driving along easily.'
106
107
--~--t----- ----th----(as) NEW CLOTHES. DAUGHTER DEPOSIT1.exhaus'
'MY daughter throws around her new clothes.'
--th-(86) *1INOEX BUY CHAIR BREAk.
'I bought a badly broken chair.'
And they co-occur only with adject1ves which follow the noun. not
those that precede the noun:
----th-----(87)TENOENCY jPOSS. BOY MISCHIEVOUS.
'The boy Is known for getting In trouble.'
----th-----(a8) ~lINOEX MISCHIEVOUS BOY.
'He's a mischievous boy.'
An analys1s which posits a predicate relation for the adjective
In (a7), but not for (a8) needs no additional statement to account for
the different distribution of the facial adverbs In (87-88). These
sentences would be accurately predicted by the earlier statement about
adverbs. repeated below:
(89)~
Adverbs co-occur with any constituent but nom1nals.
The favorable outcome of this analYSis contrasts with the
alternative analysis 1n which adjectives. regardless of position. are
contained 1n nominals. Such an analysis would incorrectly predict (87)
to be ungralMlattcal and the additional statement needed to account for
the grammatical sentences would obscure a larger generalization about
the behavior of predicates.
The preceding arguments show there are grounds for distinguishing
between attributive and predicate adjectives in ASl despite the fact
th~t In many sentences. their forms are virtually indiSt1nguishable.
Predicate adject1ves. unlike attributive adjectives may affix the
resultatlve inflection. and they can be preceded by adverbs or co-occur
with facial adverbs. In these respects. predicate adjectives behave
similarly to verbs. a generalization captured 1n an analysis which
pOSits a pred1cate relation for both pre~icate adjectives and verbs.
108
Footnotes
In some cases, the distinction between emphatic and the intensive fnflection 1s not clear, as in verbs like: WANT, LIKE. There is a possible variation on WANT or LIKE that resemble the intenSlve inflection form, and it is not clear whether these fonus are actually emphatic. However, it can be shown with a different test that they are not adjectives. Adjectives can be preceded by VERY, hut not verbs. This test was not included in 4.1 for the reason that many signers feel sentences containing VERY rather than the adjective inflected for intensive is a borrowing from English.
(i) 11NO£' VERY HAPPY.
'I'm very happy. I
(i I) '11NO£' VERY WANT BOOK.
(i1l} "'IHiOEX VeRY LIKE MOVIES.
2 In sentences lik.e (30-31), a head nod can appear concurrently with the predicate nominal. as in (i) below:
---hn(i) i INDEX OOCTOR W~<AN.
'The doctor is a woman.'
The head nod in structures such as (1) has heen analyzed by others (Liddell 1977, Coulter 1978) as a predicate. If this analysis is correct, and 1 have no reason at this point to suggest otherwise, sentences like (i) do not create problems for my analYSis of the nominal .,IS predicate. Structures lik.:! (i) are tentatively analyzed in this framc'"ork. as bi-clausal structures represented as (11) below, i.e •• the head nod b~ars the predicate relation in the higher clause, and pr~dicate nominal, the predicate relation in the lower clause. {iil helow is an exampl~ of such a SubjectwRais1ng structure; the subject of the lower clause raises to bear the subject relation 1n the higher clause:
--hn-(i1) jll<O£X WOOAN OOCTOR.
'The woman is a doctor.'
109
?\ ~INOe:i- WOHAN 'Poc-ToR.
3 A statement such as (ii) below would be acceptable as a response to (i) but cannot be used outside of a particular question-response context. Note similar pairs in English, (tii-tv):
----wh--_____ ----wh- --wh--(1) WHICH GO-AWAY BROTHER SIST£R
'Which one left? The hrother or the sister?'
(ii) llNOEX THINK SISTER.
'I think the sister (was the one).
(ill) How does President Reagan eat?
(Iv) I think wfth a fork.
4 The starred responses would be acceptdble as additional infor,1\at10n volunteered by the respondee. e.g. in response to Mis that woman a doctor1~, one could respond, MYes, and a woman (at thatll~, but these are not considered simple confirmational ManswersM to the question.
110
'~.
RED
(7) lINDEX BUY TOP-GARMENT RED.
'I bought a red top.'
RED[-+intensive]
-when--(8) SUN-SET, SKY RED [+intensive].
'While the sun was setting. the sky was
bright red. I
Figure 22
III
RED{+emphat1c}
-----0-----------(9) 2I~DEX UNDERSTAND, IINOEX WANT RED [+emphatic],
2G1VEI BLACK.
'You didn't understand. I want the red one.
but you've given me the ~ one.'
Figure 22 (continued)
112
Chapter 5
Subjecthood
In this chapter 1 examine the status of the sentence-initial
nominal (underl ined) in sentences 1 ike (1-9) below and argue that they
are "subjects", J will show that with respect to the statement of
syntactic rules, a theory which relies on semantic roles such as "source
and goal". "agent", "experiencer", etc. will fail to account for the
behavior of these nominals across different syntactic rules and that the
grammatical relation of "subject"l is needed in order to capture
generalizations obout their behavior.
(I) HOUSE BLOW-UP.
'The house blew up.'
(2) WOMAN .GIVE. BOY BOOK. -_1 J
'The woman gave the boy a book.'
(3) ~ j TAKE j LICENSE.
'The poltce took his driver's license.'
(4) ~ SlCK[+resultative].
'My sister became 111. . 113
'----...-'-
(5) ~ jSELF SEAL-TOP.
'The machine seals the top (of the jar) by itself.'
(6) TEACH,Aq ANGRY.
'The teacher 15 angry. I
(7) CUP jCL:C-FALL-OFF-FLAT_SURFACE.
'The cup fell off the edge of the table.'
(8) ~ JBITEJ T~A-I-L.
'The monkey btt the tail/ bit Its tall.'
(9) IPOSS SISTER jlNVlTEj kPOSS BROTHER.
'My sister invited his brother.'
USing .facts ab9ut verb agreement. agreement marker deletion, a
coreferent1altty constraint on certain complement structures. ffiodal
structures and the SElF pronoun 1n ASl, arguments are presented sho .... ing
that nom1nals which bear th~ 1 relation behave Similarly ocross
different syntactic rules. Additionally. these rules provide tests for
distinguishing between subjects and nonsubjects. Chapter 6 continues
with further discussion of the need for other gralMlatical relations such
as "direct object" in the statement of other syntactic rules in ASl.
5.1 Verb agreement
11'
The- fonm of Inflect1ng verbs was described 1n Chapter 2 as
consisting of a linear movement (the stem) and agreement markers at both
ends. Sentence flO) contains an example of inflecting verbs like GIVE.
SEND, ASK, SHOW, BAWL-OUT, FINGERSPElL_TO. etc. 1n which the agreement
marker at the init1al point marks for person and number of the 1
The initial point of Inflecting verbs mark for person and nu~ber of the final 1 and the end point, the final 2 of the clause.
(14) Back ..... ards ~
One sub-class of Inflecting verbs is morphologically backwards. Accordingly. with respect to verb agreement, the initial paint marks for final 2. and the end point, the final 1 of the clause.
In the following sections. the two analyses using either semantic
roles or gran~atical relations are compared with respect to other
syntactic rules in ASL.
5.1.2 Agreement marker omission
One of two agreement markers on Inflectlng verbs may optionally
be omitted. When the agreement marker at the initial point of an
Inflecting verb like GIVE. SErlO. ASK, BAWL-OUT. CATCH is omitted, the
resulting form has d reduced linear movement and agreement 1s no longer
mdrked. Additionally. omission of the agreement marker (indicated by
the MOM subscript) is possible without corresponding omission of the
117
nominal which it marks (as shown In (15). 3
(15) WIJI1AN OGIVE I NEWSPAPER.
'The woman gave me a newspaper.'
(16) OGIVE I NEWSPAPER.
'Someone gave me a newspaperl 1 was given a newspaper.'
(17) jINDEX JSENO j BOX fINISIl. LATER OSENOj CANDY.
'She sent him a box then later she sent him some candy. I
But omission of the agreement marker at the end paint is
d1sallo~led. In (18). the end paint agre{!ment marker has be!!n deleted and
the resulting sentence form is unacceptable. Only (19) is possible.
Since 3person and omitted agreement markers are similar 1n form. overt
Iperson object nominals are used in the following senter,ces to
Iperson nominals are absent except when emphatic. If omission were
possible in the presence of a Iperson nominal. as ,1s with omitte.d
in1tial point agreement markers, then the sentences should be
acceptable. but as ShOwn, they are not.
(18) 'WIJI1AN iG1VEO )INOEX BOOK.
'The woman gave ~ a book.'
(19) WIJI1AN iG1VEI IINDEX BOOK.
118 I
'The. woman gave!!!!. a book.'
An analys1s stated in terms of semantic roles would state the
facts prese'l~ed in (l5-19) in the following way:
(20) ~ Analysis: Agreement Marker On1ssion
The agreement marker for the source may optionally be omitted.
If a general statement such as (20) were correct. then the
analysls would predict (21) to be bad and (22), good. As can be seen
below. both these predictions Clre incorrect. Backwards verbs permit
omission of the end point agreement marker. and disallow omiss1on of the
initial point agreement marker. a pattern exactly opposite that of
regular Inflecting verbs.
(21) liND" JTAY£-OUT O FRIEND SISTER.
'I'm taking out my friend's sister.'
(22) 'IINDEA OTAKE-OUT I FRIEND SISTER.
'I'm taking out my friend's sister.'
Under an analysis in terms of semantic roles. these facts about
(lgreernent marker omission with both sub-classes of verbs must be
accounted for separately. With verbs like GIVE, etc. the source
agreement marker is omitted. but with verbs like INVITE. etc., it is not
the source, but the goal agreement marker which is omitted.
119
Consequently. an add1tional statement 1s needed in the grammar to
account for.backwards verbs:
(23) SR Analysis: Agreement Marker Gn1ss1on on Backwards Verbs
The agreement marker for the 2oa1 of backwards verbs may optionally delete.
However. under an analysis with grammatical relations, these
facts about agreement mar~er, om,ission ~n both sub-classes of Inflecting
verbs can be accounted for with a single statement. Since statements
(13-14) have sp~cified that the subject agreement marker appears at the
initial pOint of a regular Inflecting verb, but at the end point of a
backwards verb. a statement suct\ as (24) adequ(ltely Clccounts for
(15-22).
(24) GR Analysis: Agreement Marker fuisslon
The subject agreement marker may optionally be omitted.
As such. an analysis 1n terms of ·subject" provides a general
statement and captures similarities in syntactic behavior of both
sub-classes of Inflecting verbs. While an analysis in terms of semantic
roles may have appeal with respect to unifying the two sub-classes. it
does not offer any benefit in terms of accounting for agreement marker
omission.
5.1.3 FORCE-type verbs
120
r
The class of complement structures including matrix verbs like
FORCE. PERMIT. ADV1SE. URGE. COMf>'IANO, ASK behilve differently from other
complem~nt structures with matrix verbs 1 ike HOPE, INFORM. ASK.IF 1n
that a coreferentia11ty constraint applies with the group of
~fORC£-type" verbs. From (25-28). it can be seen that unless the
agreement markers at the end potnt of the matrix verb and at the
beYlnning point of the embedded verb are coreferential. the sentences
are ruled out. as (27-28) are.
(25) lWOEX IFORCE Z ZGIVE 1 MONEY.
'I'll force you to give me the money.'
(Z6) ZlNOE' 2URGEI lSENOj LETTER.
'You talked me into sending the letter/ You urged me to send the
letter. '
(27) '*lINOEX IFORCE Z 1G1'1E2 MONEY.
'I'll force you that he would give you the money.'
(28) 'ZINOE' 2URGEj lSEN02 LETTER.
'You urged him that I send you a letter.'
On the basis of sentences like (25-28) above. an analysis in
terms of semantic roles would posit the following:
(29) SR Analysis: Coreference Constraint on FORCE-type Verbs
121
The goal of the matrix verb and the source of the. embedded verb must be coreferential.
If (29) were sufficiently general, the following sentences in
which the source of the embedded backwards verb and the goal of the
matrix verb are coreferential should be good. but they are not:
(30) 'lURGEj jINVITEj SISTER.
'I urged him that she 1nvite him.
{31} '*CAN'T 1FORCEZ ZTAKE 1•
'He can't force you that I take you.'
If the embedded verb is one of the backwards verbs like INVITE,
the position of the coreferential agreement marker is at the end point,
not the beginning point of the embedded verb, This can be seen in (32)
below:
(32) lURGEj jlNVITEj SISTER.
'I urged him to invite his sister.'
A grammar in which syntactic rules are stated in terms of
semantic roles would need an additional statement such as (33) below to
correctly predict (30-31) to be ungrammat1ctll and (32). grarrmatlcal:
(33) SR Analysis: Coreference Constraint on FORCE-type ~ with 'tmbedded Backwards Verbs
122 1
The goal of the m~tr1x verb and the goal of the embedded backwards verb must be coreferent1al.
However. under the alternative analysis wh1ch references the
notion of ·subject M• only one statement 1s needed to state the
coreferentlality constraint for (25-28). (30_33):4
(3~) GR Analysts: Coreferenttal1tY constraint on FORCE-type verbs
The final 2 of the matrix clause and the 1 of the embedded clause must be coreferent1al.
As in the case of accounting for agreement marker omission. an
analysis of verb agreement which refers to the notion of "subject"
123
allows for the most general statement of the coreferential "tty constraint
for FORCE-type structures. An analysis in terms of the semant1c notions
of "sourcf!u" and "90a'" may be attractive as a llIeans of providing a
s~ng1e statement of agreement marker position on regular and backwards
Inflecting verbs, but fails to provide further advantage with respect to
an ac.count of other interacting syntactic rules in the language. e.g.
agreement marker omission and the FORCE-type coreferent1ality
constraint.
5.2 An. alternative analysis
In the previous se"c.t1ons, we have seen that if the verb agrp.ement
rule is stated in terms of "source" and "goal". the rules for agreement
marker omission and c.oreferent1al1ty constra1nt on FORCE-type complement
structures must be complicated in order to correctly predict the
grammatical and ungrammatical strings. On the basis of (35-36) below.
one might propose a revision 1n which the tnitlal potnt of regular
1nflecting verbs and the end potnt of backwards verbs is marked
(+Agent]. (37).
(35) WOMAN iGlVEl BOOK.
'The -woman gave me a bool(. , ..
(36) WOMAN j lNV1TE j FRIENO.
'The w~man invited a friend. I
(37) SR Analysis. Revised: Verb Agreement
The verb agrees in person <lrld number with the source and goal of the clause. The initial point of regular inflecting verbs is marked [+Agent). and the end point "of backwards verbs is marked (+Agent].
124
Additionally. if we revise the agreement marker omission rule and
the coreference constraint 1n FORCE-type structures as in (38-39) below:
(38) SR Analysis, Revised: Agreement Marker emission
The agent agreement marker may optionally be omitted.
(39) SR Malysis. Revised: Coreference constraint on FORCE-tY[le complem€"nt structures
The goal of the matrix clause and the agent of the embedded clause must be coreferential.
The revised rules. (37~39) would correctly predict sentences
(40-41) with agreement marker omission to be grammatical. and (42-43)
to be ungrammatical. And with respect to the coreference constraint, the
revised rules would likewise make correct predictions ahout (44-41):
(40) Wrn.AN OGIVE 1 BOOK.
'The woman gave me ~ book.'
(41) WOMAN jlNVITEo FRIEND.
'The woman invited a friend.'
(42:) ·C~O jGiFTO ]INDq MONEY.
'The company gave me some money.'
(43) *lINDEX OTAKEI 2ItlDEX.
'I'm taking you.'
(44) ;fORCE I 1 SEN0 2 LETTER.
'Ile (orced me to send you the letter.'
(45) ;fORCE 1 21NVITEI'
'Ile (orced me to invite you. I
(46) "jFORCE 1 2SEIIOI LETTER.
\
'He forced me that you send me the letter,'
(47) *jFORCE 1 IINVITE2'
'He forced me that you be invited by me.'
125
However. 1t cannot be correct that the initial pOint of all
regular inflecting verbs or the end point of all backwards verbs are
marked [+Agent). For example. the initial point of HATE, lOOK~AT.
GRASP.MENTAllY. ANALYZE would be labeled [+Experiencer]. as would be
the end point of backwards verbs: PERCEIVE. EAVESDROP, INTERNALIZE. But
with backwards verbs: RECEIVE, TAKE, the end point is neith~r agent or
experiencer. but [+Reclpient). Thus, in order to account for which
agreement markers are omitted and the coreference constraint. the
agreement rule liS well as the omission and coreference rules I~ould l1eed
to refer to a disjunction of several different semantic roles. for
example:
(48) SR Analysis: Agreement Marker Omission, Revised
Either the agent. exper1encer. or recipient of the clause may be omitted.
However. (48) would incorrectly allow omission of the end point
agreement marker on a regular inflecting verb like GIFT, 'to present or
to award' since the end point marks for the recipient of the clause. the
individual receiving an award. As was shown earlier. end point
agreement markers on regular inflecting verbs cannot delete.
(49) "C-O JGIFTO IINOEX TROPHY.
'The company awarded me a trophy,'
These various problems point to the difficulty of stating the
126
\ rules in terms of semantic roles. In addition to complicating the verb
agreement rule with two statements, agreement marker omission and the
coreferente constraint 1n FORCE·type complement structures must also
refer to a disjunction of several semantic roles. In an analysis which
references the not10n of "subject". cempl ications such as shown above
are avoided. '~m1nals which bear the 1 relation to the clause behave
alike with respect to verb agreement. a9ree~ent marker omission and
coreference with the matrix 2 1n FORCE-type complement structures; this
cCIl'rr.on behavior is captured by reference to "subject" In the statements
(13-14). (24) .nd (34) .bove.
The following sections discuss additional evidence which supports
the notton of "subject" in ASL clauses: modals and reflexive pronouns.
5.3 Modals
In pairs of sentences·llke the followIng. the modals: CAN. WILL.
SHOULO, MUST are preceded by nomtnals. 5
(50) IINOEX MUST STOP S<OKE['h.bltual).
'I must stop smoking.'
(51) I IIlDEX CAN MEET SUPERIOR.
'He can approach the head 9~. I
(5Z) SISTER SHOULD jCOMEj • FINISH TWO-US TOGETHER jORIVEk•
127
'My sister 1s supposed to come here then we will drive 'over there
together.'
(53) IINDEX WILL IGO-UP-TOj.
• I '11 go right up to him. face-to-face.'
I argue that the nominals which precede modals in the above
sentences bear the 1 relatton to the clause. I suggest. addittonally.
that the fact that Is precede modals follows from a general linear order
constraint: Subject ~ Predicate. As support for this hypothesis. I show
that the generalization is best stated in terms of Msu~jectN dnd not
semantic roles. and that there is evidence for analyzing modals as
predicates not auxl11aries. 6
5.3.1 Arguments for stating the rule tn terms of "subject"
Instead of stating a linear order constraint in terms of
"subject". a possible alternative would reference semantic roles and
characterize the structure as a linear sequence of constituents in the
following way:
(54) ~ Analysis: Sign Order (Partial)
Agent - Modal - Predicate.
128
But from sentences like the following, it can be seen that not only
agents precede modals (55), but also patients (56), experlencers (57),
If modals aie analyzed as pred1cates, sentences like (77-79) are
predicted to be acceptable by the analysis. If modal structures are
indeed biclausal. that 1s, 1n for example, (77). SELL heads a predicate
arc in the embedded. not matrix clause. then the facts about topical1zed
clauses in (77-79) are accounted for in a general way without need for a
special condition on modals 1n addition to (72).
As additional support, although not conclusive.S for analyzing
modals as predicates. the following show that modals may appear as short
answers to questions like (80) b~low. As discussed in Chapter 4, short
answers contain the 1 (which can be deleted) and Predicate of the
clause. The fact that modals appear alone 1n short answers, as 1n (81)
'follows from an analysls of modals as predicates, and nO add1t10nal
condition 1s needed.
------,---q----------(80) ZINOE' MUST lEAVE NOW.
'You must leave now1'
(81) YES, MUST.
'Yes, I do.'
In sum, there 1s evidence for analyzing modals 1n ASl as
133
predicates, and consequently, as multi-clausal. not mopoclausal
structures.
5.4 SELF pronouns
The SELF pronouns ('by oneself, on one's own'). like personal
pronouns, are· indexic: they mark for person and number. Either of two
forms of SELF pronouns are possible: a lax repeated form which is also
used for the emphatic ('me. myself', 'that one', etc.) pronoun "nd a
Single movement form. Since the first form 1s potentially ambiguous. the
second single movement form .111 be used for the foilowlng ""'pIes. 9
(82) SISTER ;SElF TELEPHONE CoO.
'My sister will call the company herself.'
(83) DOG lSElF ESCAPE.
'The dog ran off on its own,'
{84} DAUGHTER iSElF PERCEIVE[+random plural].
'My daughter picked it all up on her own.'
(85) ;OOOR lSElF OOOR-ClOSE.
'The door closed by itself.'
(86) iHOUSE iSElF BLOW-UP.
~The house blew up by itself.'
(87) WOMAN 1 SELF lSENOj LETTER.
'The woman sent the letter to him on her own.'
(38) BOY lSELF SEAL ENVELOPE.
'The boy sealed the envelope by himself.'
A comparison of two possible ways to state the condition on
antecedents ·of SELF pronouns shows that. as 1n the case of modal
structures. if the condition 1s stated 1n terms of semantic roles, at
least five roles are needed to account for the data.
(89) g Analysis: SELF Pronouns
The antecedent of a SElf pronoun 1s one of: Agent Experiencer Patient Instrument Recipient
As stated above, the condition would incorrectly predict the
sentence-inltial "pattents" 1n the following sentences to be possible
antecedents for SELF pronouns.
(91) *jBUTTER lSELF LEFT.
But if the condition 1s stated 1n terms of grammatical relat10ns
135
',,-,",,, :
and the initial nominals in (90~91) shown to be non~subjects, the
notion of "subject" would be sufficient to account for (82-88) and
(90-91).
(92) GR Analysts: SELF Pro~ouns
The antecedent of a SELF pronoun must be the 1 of the clause.
The condition would! likewise predict that, in clauses with
Inflecting verbs. nominals marked for subject agreement may also be
antecedents for SELF pronouns. The following sentences support this.
(93) WOMAN llNOEX jSELF lG1VEj TELEPHONE NUMBER.
'The woman w11l give you the phone number herself.'
(94) ISELF ISEN02 PICTURE.
'I'll send you the pictures myself.'
(95) 2SELF 2ASKl lBRINGj SISTER.
'Ask him yourself to bring his sister.'
(96) 2SELF lTAKE2 BEER.
'Help yourself to some beer.'
(97) BROTHER lSELF jlNVITEj WOMAN.
'My brother invited the woman h1mself.'
,--,,/
136
But nominal s marked for object agreement should not be
antecedents for SELF pronouns; this prediction is also supported.
(9B) *2S£Lf ISEND2 PICTURE.
As with the conditions on agreement marker deletion,
coreFerent1ality in FORCE·type complement stru,tures. stat1ng the
condition on antecedents of SELF pronouns in terms of "subject"
captures a generalization that nominals which bear a "subject" relation
behave similarly across different syntactic rules.
5.5 Test cases: Sentence~init1al nominals
Previous sections have argued that the sentence-initial nominals
In the following sentences are 15. In sentences with Inflecting verbs
(102-103). Is determine ,agreement. Is also may precede modals (104),
dnd be antecedents for SELF pronouns (105). Only subject agreement
markers may be deleted, as shown in (106).
(102) C-O jGIFTI TROPHY.
137
'The company gave me a trophy.'
(103) TRUE CAN 21f<DEX WRITE, 2SENDi LETTER.
'Actually, you can write up a letter and send it to him.'
(104) jINDEX SHOULD LEAVE BUTTER.
'He was supposed to leave the butter here.'
(105) CHURCH jSELF OG1VE imult fOOD.
'The church distributes the food itself.'
(106) BOY oSTEAL, BICYCLE.
'The boy stole a bicycle,'
On the basls of sentences like (102-106). it might be concluded
that all sentence-in 1t i al nominal s are Is. The fo l10wing sentence
counterparts to (102-106) present interesting test cases; in {107-1131,
the non-subject nomlnals in (102-106) now appear in sentence-initial
position and subject nominals are absent. IO
(107) TROPHY OGIfTI •
'They gave me a trophy.'
(108) LETTER OSEND j •
'The letter was sent to him.'
139
(109) BUTTER LEFT_
'The butter was left here.'
(lID) FOOD OGIVElmulto
'Food was distributed.'
(UI) BICYCLE OSTEALj _
'The bicycle was stolen.'
At this point. two questions arise: (1) do these sentence-inittal
nominals behave like Is 1n (l02-106)? and (2) if they do not. what 1s
the structure of sentences like (t07-111)? We can test whether these
nominals behave like Is with respect to the various arguments presented
in earlier sections for the notion of ·subject M• with the exception of
one: since all of these sentence-initial nominals are inanimate. they do
not meet the condition for the coreferentia11ty constraint in FORCE~type
co,nplement structures. thus this test cannot be used,
5.5. 1 Moda 15
We have previously established that Is precede modals. If the
sentence~1n1t1al nominals in (107.111) were Is, then they should be able
to precede modals. but as (112~116) show, they cannot. 11
(112) 'TROPHY WILL OGIFT1-
'The trophy can award to me.'
(U3) ·lETTER MUST OSEND1o
'The letter must send it to him.'
(ll4) *BUTTER SHOULD LEFT.
'The butter should leave it here.'
(115) ·FOOD.MUST OGIVElmul~' 'The food must distribute It.'
(116) 'BICYCLE CAN OSTEAL j _
'The b1cycle could steal it.'
5.5.2 SELF pronouns
Only Is can be antecedents of SElF pronouns. If the sentence
1n1t1al nom1nals in the following sentences were Is. then sentences in
which they appear as antecedents of SELF pronouns should be acceptable,
but as can be seen. the sentences are judged to be ungrammat1cal.
(U7) 'jTROPHY jSELF OGIFT I _
'The trophy awarded itself to me.'
(lIB) 'jLETTER jSELF OSENO j _
'The letter sent itself to him.'
1'0
(119) 'jBUTTER jSELf LEfT.
'The bulter left itself here,'
(120) *jFOOO iSELF OGIVEjmult.
'The food distributed itself.'
(121) 'iBICYCLE iSElf OSTEALj •
'The bicycle stole by itself.'
From these sentences. we see that the sentence-initial nom1nals
cannot be antecedents of SELF pronouns. thus 1n this respect, do not
behave llke Is.
5.5.3 Subject agreement
If the beginning point of the Inflecting verbs 1n the followtng
sentences were marked for agreement with the sentence-initial nom1nals.
the result1ng sentences would not be acceptable.
(122) 'iTROPHY IGlfTl"
'They gave it to me.'
(123) 'ilOOO LETTER iSENOj.
'The letter sent it to him.'
(124) \ INDEX FOOD tG1V£jmult.
141
'The food distributed SOme to them.'
(125) 'iBICYCLE jSTEALi •
'The bicycle stole it.'
Sentences (122-125) show that the sentence-initial nominals do
not determine subject agreement. thus in another way are demonstrated
not to behave like Is with respect to this rule.
5.5.4 Ambiguous sentences
Sentences like the following are ambiguous between two possible
readings:
(126) OOG SAVE.
'The dog was saved.'
'The dog saved someone.'
(127) MAN OPERATE.
'The man was operated on.'
'The man operated on someone.'
But if the sentence-initial nominal is antecedent of a SElf
pronoun or precedes a modal. only one reading is possible:
(128) OOG CAN SAVE.
142
'The dog can save anyone.'
·'The dog can be saved by someone.'
(129) tOOG t'ElF SAVE.
'The dog saved someone by itself.'
·'Someone.saved the dog by h1mself.'
(130) HAN HUST OPERATE.
'The man has to perform an operat1on.'
·'Someone else has .to operate on the man,'
(,131) jMAN t SELF. OPERATE.
'The man did the operating himself.'
·'Someone else did the operating on the man by himself.'
The fact that only one reading 1s possible for (128-131) suggests
a structural d1stinctton betw~en the two poss1ble readings for
(126-121).
The preceding arguments show that although the sentence-initial
nominals in (102-107) appear 1n the same position as Is, they do not
behave like Is with respect to three rules; verb agreement. modal
structures and possible antecedents for SELF pronouns. We turn now to
the question of what relation these sentence-initial nominals bear.
5.5.5 Possible analyses
143
Earlier. we established that the sentence-initial nominal in
sentences like (132) bears the 1 relation to the clause; the structure
o.f (132) is ~epresented as follows;
(132) WOMAN WAIT.
'The woman waIted.'
S1nce we have shown that the sentence-initial nomlnals in
(l02-107) do not behave like Is in sentences like (132). there are at
least two possible alternative analyses for these sentences, In (133).
the sentence-initial nominals are 2s at the 1nitial level and bear the
1 relation at the final level. in other words. these sentences are
paSSive structures. Or as in (134). the sentence-initial nominals are
direct objects which are frontedj the 1 of the sentence is unspecified
and does not appear on the surface.
(133) (134 )
The fact that these sentence-initial nominals may topicaltze does
144
not provide evidence for one structure over another since any nominal
heading a lor 2 arc may topicalize. In order to motivate the passive
structure represented in (133) for sentences I ike (102-107), there would
need to be evidence for distinguishing between ~ and final Is
which I have not yet been able to produce. That is, I have not yet found
rules which refer only to initial Is and those which refer only to final
15. 12 As a consequence. the rules discussed earlier which reference the
notion of subject do not refer to levels of syntactic representation.
In the absence of evidence for 1nltlal or final Is, an analysis such as
(134) seems the most probable at this point.
145
We have thus far in this chapter reviewed arguments supporting
the relational notion of Nsubject" in the statement of syntactic rules
in ASl. By employing the notion of Msubject U rather than semantic roles.
more general and simpler statements of conditions on syntact1c rules can
be made. Additionally.·condit10ns on rules which refer to only Is
provide a means of distinguishing subjects from nonsubjects, thus
contributing to a grammatical description of sentence-initial nominals
in certain clauses. Although there is ev,idence for the notion of
"subject" 1n ASl. there 1s. as of yet, no clear evidence d1stinguishing
between Minit'ial M and "final" Is. Consequently, rules discussed in this
chapter which reference the notion of subject do not refer to more than
one level of syntactic representation.
Footnotes
1 As will be discussed in the last section of this chapter, I am unable to find rules in ASL which refer to Is at more than one ie ... el. i.e •• initial or final Is. In the absence of such rules, I will assume all rules which refer to Is are stated in terms of final 1.
2 Chapter 6 contains arguments motivating the notion of Mflnal 2- in the statement of the verb agreement rule.
3 I do not fully understand the conditions for omission of the subject agreement marker. Examples in this chapter of omitted subject mar~ers include sentences in which an omitted subject marker appears with a fully specified subject. and those in which both the subject nominal and the agreement marker are omitted. further examination of a variety of complex structures may yield possible conditions on subject ilSjrt'f:'mr·(,t marker omission.
4 Although for the purposes of discussing the source-goal analysis of verb agreement. 1 have used examples of embedded sentences with Inflecting verbs, it is not r.he case that the corefercntlality constraint appl ies between agrepment markers. As stated. the constraint applies to matrix 2s and embedded Is. as the following sentences demonstrate. In (1l) below. the matrix 2 and embedded 1 are not coreferential, and the sentence is ruled out:
(1) 1INOEX 1FORCE I WOMAN 11NDEX VIS1T BROTHER.
'I forced the woman to visit her brother.'
(11) 'llNDEX 1FORCE I WDMAri j1NDEX VISlT BROTHER.
'I forced him that the woman visit her brother.'
5 Modals appear in other positions as well. initial as in (i) or final. (ii-iii).
(I) MUST 11NOEX STOP s-IOKE[+h.bltu.l).
'I have got to stop smoking.'
---------t--~--------(11) STOP SMOKE[+habitualJ MUST lINDH.
1 will not discuss here possible analyses for that th~se are biclausal structures as well. analysis of (Iii).
(1~11), except to suggest Section 5.3.2 discusses an
6 1f :1'.odals are analyzed as predicates, these modal structures In ASL are most likely subject~raising structures. i.e. (1) below:
(1) Subject RaiSing
The I of the embedded clause ascends to bear a 1 relation to the clause.
(il) SISTER SHOULO ICOMEj •
'My sister is supposed to come here.'
Aside_ from the evidence I produce whitt"! delnonstrate that modal structures are not monoclausal. but biclausal, there is not yet clear e'lidenCE that the nominal whic.h hears the 1 relation to the matrix clause also bears a 1 relation to the embedded clause.
7 These sentences are not to be confused \~ith grammatical strings such as (i) below in which the initial nominal is a topicalized 2:
--t-(1) BOOK. CAN WOMAN iGIVEj MAN.
'The book, the woman can give It to the man.'
8 Although the short answer examples are compatible with a blc.lausal analySiS, they do not argue against the auxiliary analYSis of modals. That is, the following formulation of the condition on short answers IS possible:
(i) Alternative: ~~
Short answers are composed of: Subject Aux Pred
147
--=------- .. ~-=--------
9 SELF pronouns can be ordered either after the subject as in the example sentences, or before the subject, as in (i) below:
(i) iSELF iHOUSE BLOW-UP.
'The house blew up by itself.'
Since position of the SELF pronoun is not relevant to the p,esent diSCUSSion, I shall consistently order SELF pronouns after the subject in the example sentences for purposes of consistency and ease of illustration.
10 The initial nominals in the follOWing sentences (l02~106) may appear with or without a topic marking of raised eyebrows which has been discussed elsewhere as an indicator of topicalized 2s (Liddell 1918. McIntire 1980). As I discuss here and in Chapter 7, topic marking may appear over Is and as well as 2s. thus is not relevant to the oresent discussion of distinguishing between Is and 2s. I have chosen then to use examples in which topic marking 1s absent from the initial n~ninal for ease of discussion and representation.
11 Sentences (112~116) are not to be confused with another set of similar sentences. shown below in which modals are preceded by a constituent break and the subject'of the sentencl'! 15 unspeCified. The (a) sentences are related to the (b) sente~,es except that the 2 is fronted and generally toptcallzed.
(ia) WILL OGIFT1 TROPHY.
'They will award me a trophy.'
--t-~
(ib) TROPHY. WILL OGIFTI•
'They will aw~rd me a trophy.'
(iia) MUST oSENO j LETTER.
'They've got to send him the letter.'
---t--(iib) LETTER. MUST OSE"Oj'
'They've got to send him the letter.'
12 A posslble argument for flnal I-hood would be word order, that is. nominals which appear in sentence-initial position are final Is. However, in the absence of other independent evidence for final I-huod. the argument cannot be supported.
1'8
Chapter 6
Verb Agreement
This chapter is concerned with the statempnt of the verb
agreement rule in clauses like (1~2) below in which agreement is
determined by Ule subject and notlonal ~ object, but 1n (3~4).
the subject and notional lndirect object.
(I) GIRL jHATE j BOY.
'The girl hates the boy.'
(2) OOG jBtTE j CAT.
'The dog bit the cat.'
--t-~
(3) 2 MAN. WOMAN oGIVEjdu BOOK.
'The .... oman gave the two men books.'
(4) SISTER j SHOW I PAPER.
'~IY sister showed me the paper.'
The structures proposed for sentences like (1) and (3) bear on
the statement of the verb agreement rule. Friedman (1975) proposed that
the form of agreement morphology could be predicted from the "sourceM
149
and "goal" of the clause, as in (5) beiow. Since the notions of source
~nd goal have been prt1poscd elsewhere (Fl11mor(! 1971, JtlcK(;ndoff 1972,
Gruber 1978) for "verbs of motion~, at first appearance, they seemed
partlcularly opt for the descript10n of a class of verbs in ASL. Under
Friedman's (1975) analYSiS, the d1rection of the linear movement in
verbs was described as originating at the locus pOSition of the source
nominal of the sentence and moving toward the locus position of the
goal.
(5) Source-Goal Analysis: Verb Agreement
l~O
The initial point of the d1rectional verb marks for per SOil and number of the source ond the end point~ the goal.
5.1 presented several arguments against statfng the a9re~mrnt
rule in terms of "source". While it appeared that the notion of
"source" captured certain generalitles about verb morphology, the same
benefit failed to appear elsewhere in the grammar. From the data on
agreement marker omission and the co-referential1ty constraint in
FORCE~type complement structures. it can be seen that stating these
rules in terms of "source" resulted 1r, statemer.ts which were less
general than if stated in terms of "subject."
Keg} (1978) suggested that sentences like (3~4) are -advancement
structures" 1n whiCh the M1nitlal- indirect object advances to .final"
direct object. Under this analysis. the verb agreement rule is state.;l in
terms of grammatical relations and levels of syntactic representation.
shown below 1n (6);
l
.~ _________ ...c:_-_::-
(6) Advancement analys1s! Verb Agreement
The inttial point of inflecting verbs marks for person and number of the final I, and the end point. the final 2 of the clause.
"In th1s chapter. I present arguments for the latter statement of
verb agreement. showing that the rule 1s best stated in terms of
gra:nmatical relations and levels of representation. In section 6.1.
argue against a verb agreement rule which references the semantic
notion of -goal-. In 6.2. I argue that sentences like (3~4) are best
represented by the multi-level advancement structure shown in (7). The
arguments in support of (7) draw from the facts about verb agreement and
other syntactic phenomenon in ASl.
--t--(7) 2 MAN, W()1AN oGIVEjdu BOOK.
'The woman gave the two men books. I
6.1 Arguments against the source.goal analysis of verb agreement
As defined by Gruber. ~goals- are "the ultimate destination of a
motion" (1978:66) such as (8-13) below (goals are underlined):
(8) John sent Bill a book. (9) John threw""""the ball to B111. (10) John received a book. frOiiilHll. (II) JOhfi inherited a million dollars. (12) J01iii ran ~ the deck.
151
(13) The ball rolled out of the house and 1nto the hole.
By referencing the notion of -goal- as defined above. the source
goal analysiS attempted to account in a general way for the form of verb
agreement 1n ASl. In sentences like (1~2). the -goaP 1s the notional
direct object. th,:!s the cnd,point of the verb agrees with the direct
object. But since in (3-4). the notional indirect object 1s the -goal-.
the verb instead agrees with the indirect object and not the direct
object. However, there are sets of verbs 1n ASl which exhibit agree:r.ent
morphology. but agreement 1s not expressed with the ~goal~ of the
clause. as defined by Gruber.
6.1.1- Backwards verbs
152
Backwards verbs were first discussed in Chapter 5: briefly, they
comprise the relatively smaller sub-class of Inflecting verbs for ~h1ch
agreement markers are Nbackwards·. In contrast to other Inflecting
verbs, the subject agreement marker or. Backwards verbs is not located at
the beginning point, but at the ~ point. and the agreement marker
for the notional direct object 1s instead located at the beginning
In order to account for (17-18). the source~goa·l analysis would
need to be mOdified. either by adding a special condition for sentences
like (18). or somehow refining the notion of goal to exc.1ude cases like
(17). Under either of these modifications, the position that the
notion of Mgoal K is useful as a means of determining agreement
morphology 1s weakened.
But if the agreement rule were stated 1n terms of grammatical
relations. as in (6), along with an additional statement about the form
of agreement markers on backwards verbs:
(19) Backwards ~
One sub-class of Inflecting verbs is morphologically backwards. Accordingly. with respect to verb agreement, the initial point marKS for final 2, and the end point. the 1 of the claose.
only (18) would be acceptable.
6.1.2 The class of Inflecting verbs
Additional objections to the source-goal analysis can be ralsed
on different grounds: many Inflecting verbs which contain a linear
movement do not entail "motion" except in an extended. looser sense,
e.g. HATE, LOOK-AT, PERCEIVE. REAO-EACH-OTHERS-HINOS. TEACH. INFORK. In
the case of these verbs. determining what constitutes a ~goal· is less
obv1ous.
IS.
As it is presently formulated. the source-goal analysis of verb
agreement in ASL requires certain complications in the grammar which are
not needed if verb agreement 'is stated tn terms of grammatical
relatloIJS. While problematic for Inflecting verbs. the notions of
source and goal. however. appear to be useful 1n characterizing the
direction of linear movement in directional Spatial verbs (Newport
IS8!. Supalla 1982). As a class, directional Spathl verbs ental1
motion and location. e.g.: 1Cl:V-WAlKj ('person walk from here to
there'). tCl:3.PASS-8.Yj (,vehicle drive from here to there'). and do
not mark for person and number agreement. I propose that syntactic
phenomenon, such as person and number agreement are best stated In terms
of grammattcal relations.
6.2 Stating the verb agreement rule
An adequate verb agreement rule must account for the form of
agreement morphology In sentences l1ke (20-21) below. In (20), the
notional direct obje,ct fs 3person dual, and the agreement marker 1s
likewise 3person dual. In sentences 11ke (21). agreement fs not
expressed with the notional direct object, but instead with the notfonal
indirect object. (22-23) show that if the verb marks for agreement with
the direct object instead of the Indirect object, the resulting form Is
either incorrect for the translation or unacceptable:
---t----(20) DOG, CAT, MOST IINOEX OFEE01,du'
155
'I'll have to feed the dog and the cat.'
(21) WOIIAN lGlVEj MAN 2 BOOK.
'The woman gave the man two books.'
(22) WOMAN lGlVEj,du MAN 2 BOOK.
*'The woman gave the man 2 books.'
(23) ·C-O OGIFT1 ~ TROPHY.
'The company gave ~ (of all peoplel) a trophy. I
Two different analyses for verb agreement are poSSible, each with
different consequences for the statement of the verb agreement rule.
(24) below 1s a unf-level structure; MAN bears the 3 relation to the
clause. The verb agreement rule under this analysis would need to be
stated disjunct1vely, shown below In (25):
(24) WOMAN lGlVEj MAN BOOK.
'The woman gave the man a book.'
(25) Unl-level Analysis: Verb Agreement
The verb agrees in person and number with 1 and 2. if no 3. otherwise 1 and 3.
A second analysis posits a 3-2 advancement structure for
I
156
sentences like (21). Similar structures have been proposed for a number
of languages. including: Baha Indones 1 an (Chung 1976). Chamorra (Crain
1979), s. Tiwa (Allen & Frantz 1983). and TzotlO (Alssen 1983). The
principal characteristic of 3-2 advancement clause 1s a nominal wh1ch
heads an 3-arc In one stratum and a 2-arc in the following stratum. As
represented below. the nominal ~ in (26) 1s referred to as an
advancement~. The inltial 2. BOOK bears the ,horneur relation
.~.
157
will show that (a) the nominal wh1ch heads the 3~arc ilt the in1tial
stratum heads the 2·arc 1n the final stratum. and (b) the nomH.al which
heads the 2·arc in the initial stratum does not head the 2~arc 1n the
final stratum.
6.2.1 Arguments for the advancement analysis
in the final stratum. 6.2.1.1 An argument for final 2-hood: Sign order
(26) WOMAN ,GIVEj HAN BOOK.
IThe woman gave the man a book.'
Under this analysis. the agreement rule refers to final
grammatical relations:
(27) Advancement Analysis: Verb Agreement
The v.?rb agrees in person' and number with the final 1 and 2.
The analysis accounts for the ungrammatical (22-23): the initial
2 bears the chomeur relation in the final stratum, consequently, It is
not marked for agreement.
In order to argue that clauses like (21) are advancement
structures, that 15. have the structure represented as (2o) above,
In sentences where agreement Is determined by the subject and
direct object. like (28-29) below. the direct object follows the
predicate:
(28) IINOEX FINISH oFEEO, ~OG.
'I've already fed the d09.'
(29) OOG ,BITEj CAT.
'The dog bit the cat.'
In sentences where agreement Is determined by the indirect
object, the indirect object follows the predicate. A sentence like (31)
below. where the direct ubject is ordered after the predicate and before
the indirect object is ungrammatical.
(30) WOMAN ,G1VEj HAN BOOK.
'The woman gave the man a book.'
158
I
I
(31) 'HOMAN lG1VEj BOOK MAN.
'The woman gave the man a book.'
If a structure such as (32) below were postted for sentences like
(30). the s1g" order rule, like the verb agreement rule, would also need
a disjunct1ve statement:
(32) WOMAN lG1VEj MAN BOOK.
'The woman gave the man a book.'
(33) Uni--level Analys15: 2.!!t!!. ~ Order of constituents In ASL 1s:
1 - P - 2 (if no 3) - Obliques otherwise: I-P - 3 - 2 -Obliques
An advancement analysis of {30} would avoid such a complication;
HAN bears the ftnal 2 relation, and BOOK, which bears the chomeur
relation, and is thus a non-term, 1s ordered after the final 2:
(34) Advancement analys 15: ~~ Ordt"r
Order of constituents in ASl Is: Final 1 - P - 2 - Nonterms
The advancement analysis captures a generalization about the
Similar behavior of fll1al 2s which is less obvious in the disjunctive
159
statement: they are marked for agreement on the verb and are ordered
after the predicate.
Alternatively, one might argue on the basis of the data on
.agreement and 519"0 order that MAN in (30) bears the 3.. not the 3
relation to the clause. 1.e •• (35), and BOOK bears some unspecified
grammatical relation!. perhaps an oblique. Under such an analYSis.
no diSjunctive statements are neededj the rules are cOn~jderably
simpler. Furthermore, they do not need to be stated 1n terms of levels
as 1n the advancement analys1s:
(35) WOMAN lG1VEj MAN BOOK.
'The woman gave the man a book.'
~fG.~2C HAN
Gwr; WOHAN ~DOr...
(36) Unl-1evel Analysis B: Verb Agreement
The verb agrees in person and number with the 1 and 2
(37) Unl-1evel Analysis!:.. Slqn ~
Order of constituents Is: 1 - P - 2 _ Obliques
The next section provides an additional argument for the
advancement analysis by showing that a distinction Is needed between
initial and f1nal 2s, one which the alternative uni-level analYSis fans
to make.
160
r
I ,
6.2.1.2 An argument for initial 2-hood: locus shifting
Since the form of locus shHting h perhaps unusual to ASl. the
phenomenon is discussed at length here as a preface to the presentation
of the argument itself.
Chapter 2 discussed verbs in which agreement and locative markers
take the form of some locus pOSItion. The form of agreement markers
involves pOSitions at points within sign space around the signer's body
as in, for example. the 3person agreemer.t markers: iGIVEj ('he gave
her'); the sign origlnate.s at locus position 1. and moves toward
position 1. locative markers on Spatial verbs likewise involve locus
positions, e.g. lCl:V-WALK-TOj and ,Cl:V:STANO. For those
Inflecting and spatial verbs which contain a linear movement between two
locus points,' I use a general tern: -directional verb-I 1
Otne. lexical itemS in ASl are indicated with locus pOSitions.
Including pronouns such as tlllOEX and specially marked nouns. e.g.
tDGG (as shown in 2.1.1).
In (38-39), it can be seen that the same locus point, 1. appears
for coreferential pronouns and agreement markers. In (38). the subject
pronoun and subject agreement markers are indexed In the same location.
In (39), the subject of the first sentence and the direct object of the
second are coreferentlal. and the location of the indexlc pronouns Is
the same. (Note: The subscripts, h h.h ... have previously been
161 162
used 1n the glosses to specify any distinctive locus position in sign
space. When these subscripts appear 'in the transliltions of the
sentences below, they are USed to denote coreference, i.e. nominals
which appear In the translation with same subscripts are coreferential.)
'Shel walked over there. stopped and thought a bit, then she1
decided to wait there.'
(43) MAN ,INDEX GET-UP. lCl:V-WAlK-TOj. jlNOEX SEEM DEPRESSEO.
'The man 1 got up and walked away. lief seemed to be depressed.'
In (44), the direct object In the first clause, BOOK, Is indexcd in
163
location i. but In the second clause, the locus position of BOOK has shift
to location 1. As in (42-43). although the two pronouns are indexed in
different locations, they are understood to be coreferential:
('4) 11NDf.X BOOK llNDEX jMOVE j • D:[+ld'omatlc]:WRDNG jlNOEX
FAll.
'I moved the book! from here to over there. But itt fell off.'
A number of conditions apply to constrain locus sh1ft1ng in ASL.
First. whtle subjects of Intransitive clauses shIft (42-43). subjects of
transitive clauses cannot. as (45) demonstrates:
(45) BOOK l1NOEX lG1VEj. jlNOEX ~ANT BOOK BACK.
-IShe t gave him the book. (How) shej wants It back.'
IShe, gave him the book. Hej wants it back.'
Second, if a clause does not contain a directional verb. i.e., it
has either a Plain verb or a Spatial verb without a linear movement,
then shifting does not take place. In (46) below. a Plain verb, WALK Is
substituted for the Spatial verb, Cl~V-WAlK~TO in (43), and ~ see th('lt
shifting does not take place: the pronouns indexed at location 1. and
1 are judged non~coreferentlal.
(46) MAN tINDEX GET-UP, WAlK[+continuatlve]. jINDEX SEEM OEPRESSED.
·'The man! got up and walked for a while. Het seemed to be
depressed.'
'The man t got up and walked for a while. Hej seemed to be
depressed. '
As can be seen from the examples. evidence of locus shifting
appears only when the nominal Is fe-indexed in a later sentence. I have
chosen as illustration of this phenomenon. examples which contain
pronouns or index1c markers 1n the ~shifted~ location. but in ceita1n
other discourse contexts. pronouns cannot appear 1n the shtfted
location. but must be indexed at the original point. There appear to be
,
164
I \,
a number of factors at the level of discourse ol'ganization ... hich
determine whether a pronoun appears in its original or shifted location,
such as Mbackgrounded- or -foregroundedM information, but for our
purposes here. we concern ourselves only with ~ nomillals may shift
and select as tests those contexts in which the pronoun or agreement
markers must appear in the shifted location. Thus the question of ~
pronouns or agreement markers appear in a shifted location 1s
independent of lihich nomina1s may shlft, the focus of the remainder of
this section. 2 Since no prevlous descriptions of the indexic system 1n
ASL (Friedman 1975. lacy 1974), have to my knowledge, discussed this
phenomenon of locus shifting. I discuss two alternative analyses.
Th~ nominals which undergo locus shifting appear to be visible
count~rparts to the semantic notion of Mtheme M in verbs of mot ton
proposed by Fillmore (1968, 1977), Gruber (1978). Jackendoff (1972).
Def1nitions of Mtheme" appear below:
(41) ~ Verbs of ~
·With verbs of motion. the Theme 1s defined as the NP understood as undergoing the motion." (Jackendoff 1972:29)
165
MFor motional verbs ••• the entity which ts 1n mot ton (is) the Theme of the sentence." (Gruber 1978:38).
(48) John w"lked to the store.
(49) l!!! ~ moved away.
(50) Harry gave lli ~ away.
(51) Will inherited ~~~.
(52) Charlie bought thE" 1 amp from Max.
From the data presented thus far. it ... ould appear that
determining which nominals may shift can be predicted from the ser;,antic
roles of nominals In the sentence. (53) provides a thematic analysis of
the phenomenon:
(53) Thematic relations gnalysls: locus shifting
In sentences with directional verbs, themes shift to the locus pOSition of the end point of the verb.
As an alternative analysts. I propose that locus shifting in ASl
is not stated 1n terms of the semantic notion of theme. but rather the
relational notion of Mabsolut1ve N• specifically: Minitial
absolutiveM• An absolutive arc shall be defined following Perlmutter
(to appear):
A stratum is transitive 1f and only 1f it contains a I-arc or a 2-arc. A stratum is intransitive if and only if it is not transitive. An arc a is an ergative arc in stratum S. if and only if a is a "t~arc and S1 is transitive. 1
M arc a is an absolutive arc in stratum $. if and only if a 1s a Wuclear term arc and a 1s not an ergative arc in "Sjo -
show that stating the rule in terms of the semantic n~ti~n of
theme makes certain inaccurate predictions in ASl about which nominals
m~ sh1ft. but if the condition 1s stated tn terms of "1nitial
absolut1ve". a more general statement can be prov1ded.
1£6
J
6.2.1.2.1 An argument against a thematic analysis of locus shift1ng
To repeat, the notion of "theme" is def1ned by Jackendoff (1972)
as "the NP understood as undergoing the motion", and similarly by Gruber
(1978:38) as "the ent1ty which is conceived as moving or undergoing
transitions," Examples appeared earlier of nom1nals which would be
·tnemes· 1n English (48-52).
The rollowing sentences in English, {57-60} provide interesting
test cases for the not ton of -theme-; to my knowledge. these sentences
have not been previously treated. In sentences such as these. it would
appear th~t the ~ntittes which are conceived as moving are both the
subject and direct object. Thus, these sentences contain ~ themes.
(54) The boat shipped the coals up the river.
(55) She brought cookies to the party.
(56) He carried the child to the sofa.
(57) She drove the car to New York.
If there were clauses comparable to (S4.S7) 1n ASl which contain
two themes, then a condition on locus shifting stated 1n tenms of
-theme K would predict that both the subject and direct object of these
trans1tive verbs shjft. Exact translat10ns of the verbs 1n (S4·S7) are
difficult, but the ASL verb. BRING/CARRY appears to be the closest
translat10n to the English 'bring' or 'carry'. Using a context 1n which
shifted pronouns can appear, we can test whether both the subject and
167
the direct object shift in ASL sentences similar to (54-57). t.s (58)
shows~ the subject does not sh1ft. but from (59), we see the d1rect
object does:
(58) WOMAN tlNOEX t8RINGj COOKIE. jlNOEX MAY£.
·'That womao 1 brough_t the cookies. She; made them.'
'That woman t brought the cookies. Hej made them.'
,----q--------(59) COOKIES tINOEX. WOMAN tlNOEX t8RINGj COOKIE. IINOn
4 j TAKE l'
'Remember those cookies? Well. the woman brought them here
and I took some.'
Consultants report that a closer translation of (59) must consist
of two clauses, one where WOMAN is subject of an intransitive clause
containin9 a directional verb such as WALK. and another claus.e with
'That woman 1 came here and brought cookies with her;,: ·Stef
made them. I
If -theme K is def1ned as g1ven 1n Gruber (I978)~ Jackendoff
(1972), the not10n is too general with respect to the phenomenon of
locus shifting in ASL. The condition on locus sh1ftfng as stJted 1n (53)
168
r
would pre~lct that in clauses with more than one theme. both the subject
and direct object shift, when 1n fact, only the direct object dues. A
·uniqueness cond1tlon u could be added In which only one theme appears In
a clause. The problem. however. 15 not removed: which nominal in (60) is
designated as -thcme w1'The following condition would be needed:
(61) Theme: Uniqueness Condition
In a clause with more than one theme. only the direct object shifts in locus.
Stating the condition on locus shifting in terms of -theme W does
not dispense with reference to grammatical relations. and requ1res an
additional condition.-
The ph~nomenon of locus shifting in ASL. as I show below. is
~tated more generally 11'1 terms of Winitial absolutive-. Rules which
reference the notion of absolutive have been proposed for a number af
languages. inclUding: Chamorro (Crain 1979). Udi (Harris. to appear).
"nd Southern Tiwa (Allen & Frantz I97a). Initial absolut1ve is defined
as a numinal which bears the absolut1ve re'lation 11'1 the 1n1tial
(62) Gil Analysis: locus Shifting
In clauses containing directional verbs. the lacus position of the initial absolutive of the clause shifts to the position of the end potnt of the verb.
169
Earlier. two different analyses for sentences like (63) were compared.
Since the indirect object in (63) behaves like a direct 'Object hi a
simple transitive clause with respect to verb agreement and sign order,
two possible analyses were entertained: a uni-level analysis such as
(64) In which MArl bears a direct object relation to the clause and SOaK,
some other grammatical relation ~. and (65), an advancement Structure
in which BOOK Is initial direct object and MAN, final direct object:
(63) W~AN jGIVEj MAN BOOK.
'The woman gave the boy a book..·
If we ~ssume a structure like (64) for sentences like (63). the
condit,ion on locus shifting would require a disjunctive statement. and
would be less general:
(68) Un1-1evel Analysis: locus shifting
In clauses containin9 directional verbs, intransitive subjects or nominals bearing grammatical relation!. shift to the locus position of the end point of the verb.
The advancement analysis. however. would avoid the complication
of a disjunctive statement. An analYSis of locus shifting in terms of
the relational notion M1n1tial absolut1ve- makes correct predlct10ns
1 110
about sentences (58-59). According to (62). only the inittal absolut1ve
shifts and not the ergative nominal. I conclude that the facts about
locus shifting in ASL are best stated 1n terms of the not10n "inittal
absolutive".
6.2.1.3 The indirect object relation
The data on locus shifting. verb agreement and sign order show
that the nominal which bears the final direct object relat10n 1n an
advancement structure does not bear the direct object relation 1n the
initial stratum. The question now remains as to what relation the
advancement nominal bears In the initial stratum. Although It seems
most plausible that the advancement nominal 15 an Initial Indirect
object. there are. unfortunately, no rules to my knowledge which refer
only to indirect objects. Whether or not the advancement nominal bears
an Indirect object relaLion in the initial stratum or some other
grammatical relation. ! does not alter the facts presented above
supporting the advancement analysis for sentences like (63).
In conclusion. I have argued In this chapter as well as the
previous chapter that the statement of the verb agreement rule in ASL
necessarily refers to grammatical relations. have shown In this
chapter, additionally, that more than one syntact1c level Is needed 1n
order to state the facts about verb agreement, sign order and locus
shifting In ASL. By posttlng an adVancement structure for sentences
like (63) the facts about verb agreement, sign order and locus shifting
in ASL are captured 1n a general way. Additionally. h~ve shown that
although Inflecting verbs 1n ASl appear to mim1c in a transparent way
semantic notions such as "theme~. ·source" and "goal", stating syntactic
rules 1n terms of these notlons does not allow for correct
generalizations across a range of sentences and verb SUb-classes.
172
r
L ..
Footnotes
1 Although the term Mdirectional verbM has been used elsewhere (I.'ood~ilrd 1974, Friedman 1975) as a name fOf' a verb class in ASL. 1t is used here to refer to verbs which contdin a 1 inear movement between two locus points. Verbs of this form can be m~nbers of different morphological classes. either Inflecting or Spatial.
2 M example of a conlext in which pronouns would not appear 1n the shifted location are sentences like (1) below. If the second sentence in the sequence is negated, pronouns must appear In the original, not snlfted location:
{l} ilNDEX NOT jWAlKjo llNOEX SCARED.
'He i didn't walk over there. Hei was scared.' ·'He l didn't walk over there. Sfiej was scared,'
The conditions under which pronouns appear in shifted or original locations are intriguing as a means of examining organization of back grounded or foregrounded information <IS well as temporal sequencing. Further discussion of this topic is beyond the scope of the present work and is reserved for future investigation.
3 The phenomenon of locus ~hif~ing refers to ~ iha,gehin ~ocation from i to samle' other locat 10n J at the end POlil 0 t e dlrectional verb.- In clauses containing dTrectional verbs. direct objects of simple transitive verbs determine the form of the end point agreement. including its locus position. Although these nominals are inHlal absolutiv~s. the present discussion of the condition on locus shifting does not concern them for the reason that they do not Kshift to" the locus position of 'the end point ± of the verb from some other position !. but are always positioned at •
4 TAKE 1s a backwards Inflecting verb; the direct object agreement marker is located at the beginning point.;
173
Chapter 7.
Classifiers and Indices
In this ,hapter. I examine the structure of sequentes which
contain indices and classifiers. While there has been som~ det~iled
morphological description of classifiers (Supalla 1982. N~\\port 1981)
and indices (Keg1 1978. Lacy 1974). comparatively little is understood
about their syntactic behavior. Both indices and classifiers involve
specifying a location in the sign area. Indices are "pointing" ~estures.
oriented toward a locus position in nEutral space. These includ~ the
personal pronouns. (e.g. lINOEX 'I') and the locative pronouns. (e.g.
jlNDEX 'there'). Like indices. classifiers specify locus positions.
e.g. lCl:3 ('vehlcle here').
In complex sequences involving classifiers such as (1-2). Liddell
(1979) has proposed that the initial nominals (underlined) are "locative
objectsK or obliques, and that constituents in such s~ntences al'e
ordered differently from other sentences containing v~rbs of different
classes. Specifically. in sentences w1th Plain or Inflecting v2rbs. he
argues that "basic" ("underlyingn) order of constituents is
"subject~verb~object". but in sentences like (1-2) with classifier verbs
(underlined). constituents must be ordered: "locative object~subject
verb". Under th1s analysis, two order rules ar~ needed, one for clauses 174
J
containing Plain and Inflecting verbs. (3) below. and another for
clauses containing a sub~c1ass of Spatial verbs, class1fier verbs, (4)
below. Additionally. the analysis assumes that (1-2) are monoclausal.
McIntire (1979) has proposed a similar analysis.
(I) l-haod: t Cl :B -------------
R-hand: TABLE MONKEY }l:V.
'The monkey stood on the table.'
(2) l-haod: t Cl : 4 ------------------
R~hand: fENCE CAT .;Cl: V-CROUCH.
'The cat sat next to the fence.'
In clauses cont41nlng Plain and Inflect1ng verbs. order of sign constituents is: Subject - Verb - Object.
(4)~~!
tn clauses containing Classifier verbs. order of s1gn constituents is: local ive object - Subject - Verb.
From sentences l1ke (5-1) below. Keg1 (1976) has claimed that In
clauses containing Inflecting verbs, sign order is less restricted than
with Plain verbs. She proposes that orders of "S·O-V" as 1n (5), or
MO_S_V" as 1n {6}, or even "V.S-O", as in (7) are possible In clauses
containing Inflecting verbs. As in the analysis of sentences with
classifiers. this analysis of sign order crucially rests on the analysis
of (5-7) as monoclausal; the ind1ces are either extra-linguistic
5JH!clflers or some type of dctC!rmlner. 1
175
(5) (BOY tINOEX) [GIRL jINOEX) tKICKj'
'The boy kicked the girl.'
(6) [GIRL jINOEX) [BOY tINOEX) tKICKj"
'The boy kicked the girl.'
(7) tKICKj [BOY tINOO) (GIRL jINOEX].
'The boy kicked the girl.'
These various analyses propose that sign order 15 determined ~y
the class of the verb In the clause. Using arguments presented in
earlier chapters for predlcatehood. 5ubjecthood. and direct objecthood
as tests. I re-examine clause structure of sentences liKe (1-2) and
(5-7). I show first that both classifiers In the sequence are
predicates. and the nouns preceding them are subjects. On the basis of
this data and other arguments which 1 present 1n the following sections,
1 argue that sentences like (1-2) and (5-7) are multi-clausal, not
monoclausal. Specifically. 1 show that there are grounds for analyzing
the clauses 1n (1) as 1n (a) below, and those in (5) as (9).(1 leave
open the question of whether the separate clauses I posit in each case
together constitute a s1ngle sentence. or separate sentences. 2)
176
r (B) l-h.nd: jCl:B ------------_
R~hand: TAULE MONKEY j~_
'The mon~ey stood on the table.'
(9) (BOY jINOEX] (GIRL jINCEX) jKICKj •
'The boy kicked the girl.'
finally, I demonstrate that this alternative analys1s has
consequences for the statement of s1gn order in ASl.
7.1 Arguments for clausehood of Noun + Classifer sequences
Consider the two analyses of sentences like (10-11) as outlined
above. Under the monoclausal analYSiS, the init1al sequence of Noun +
Classifier (underlined) is a ·locative object" phrase. 3 The classifier
serves to identify a locus position for the noun preced1ng it, as such,
is a type of specifier. The structure of (10.11) is represented as (12)
117 ",,",--,,-'
below. and an order rule, (4) applies to clauses with classifier verbs.
(10) l-h.nd: jCl:B ______________ _
R~hand: TABLE MONKEY jCl: V.
'The monkey stood on the table.'
(II) l-h.nd: jCl:4 - _______ _
R-hand: FENCE CAT jCL:V.
'The cat sat next to the fence.'
(IZ)
HOtVK€,(
In the following sections, I argue against the rnonoclausal
analysis and for an alternative analysis in which the initial setjuence
(underlined) of Noun + Classifier is not a noun phrase, but a claus~.
present arguments showing that both classifiers in the sequence are
predicates and the nouns preceding them are subjects. In Section 7.3,
discuss other related structures and propose an account for them under a
mu1t1clausal analysis.
7.1.1 Sentent1al complements
As discussed 1n Chapter 3. THINK takes only a clausal direct
object. Thus. sentences containing non-clausal direct objects. as in
(14) are ruled out:
178
(13) IINDEX THINK 2INDEX FINISH LEAVE.
'I thought you already left.'
The fact that sequences of Noun + Classif1er may appear as
complements of THINK provides the first argument that these sequences
are not noun phrases, but clauses.
(15) liNDE' THINK WCi<AN jCL,V.
'I think the woman stood over here.'
(16) liNDE' HOPE CAR jCL,3 STILL.
'I hope the car is still parked there.'
7.1.2 Arguments for predlcatehood of class1f1ers
7.1.2.1 Short answers
Short answers contain the subject and predicate of the clause, and
subjects in such answers may be deleted. Thus answers conta1ning only
the predicate are acceptable as in (18) but short answers containing
only the subject or any other nominal 15 not, as 1n (19-20):
•••••••••••••• q ••••••••••
(17) 2INDE' HAVE MONEY. CAN PAY.
179
'Do you have enough money to pay for it?'
----hn---(18) YES, HAVE.
·Yes. I do.'
------hn----(19) ·YES. I INDEX.
-----hn---- . (20) ·YES. MONEY.
If classif1ers 1~ (21) and (23) are predicates, they should be
able to appear alone 1n short answers, and (22) and (24) show this to be
true:
.•••. Q .........••. (21) QUESTION COW jCL,VV.
'Was the cow standing there?
(22) YES. jCL,VV.
'Yes. 1t was. I
------------q-----------(23) QUESTION 2POSS CAR jCL,3.
'Is your car parked over there nowl'
(24) YES. jCL,3.
'Yes. it 15.'
The preceding arguments show that sequences of Ibun + Classifier
are clausal and that the classifiers head a predicate arc. The
following arguments prov1de further support for postting a predicate arc
for these classif1ers by showing that they are verbs, and since all
verbs are predicates, cl~ss1f1ers must th~n be predicates.
180
7.1.2.2 linear position of verbs
As demonstrated in Section 4.1.3. verbs cannot be attributive.
that is appear before nouns in a noun phrase:
(25) *JINDEX SEE SIT-DOWN BOY.
'I saw the sitting boy.'
(26) *11NDEX SEE WAIT W~~AN.
'I saw the waiting woman.'
If classifier predicates are verbs, it should be the case that
they also cannot be attributive. (21-29) support this prediction:
(27) *lINDEX SEE jCL:3 CAR.
'I saw the parked car.'
(28) *ll"DEX SEE jCL:V WOMAN.
'I saw the standing woman.'
(29) "DAUGHTER WATCH jCL:VV COW.
'My daughter watched the standing cow.'
If the sequence of Noun + ClassHier ... ·ere analyzed as nomlnals,
that Is, the classifier is analyzed as some type of specifier or
181
determiner, then tn order to handle the ungrammatical ?equences of
(21-29) above. and allow the grammatical (30-32), a special rule wculd
be needed to order the classifier after the no~n head of the phrase.
Under the multi-clausal analysis. no additional rule would be needed.
The generalization about order of verbs observed in Section 4.1.3
handles the data here if these classifiers are analyzed as verbs.
(30) 11NDEX SEE CAR jCL:3.
'I saw that the car was parked there.'
(31) 11NDEX SEE WOMAN jCL:V.
'I saw that the woman was standing there.'
(32) DAUGHTER SEE COW jCL:VV.
'MY daughter saw that the cow was standing there.'
1.1.2.3 Facial adverbs
The facial adverbs. --th--, --Plll--. and --em-- cannot ·co-occur
with nominals. If the classifiers in the following sentences were
cont.ained in nominals. then sentences with facial adverbs co-occurring
with classifiers should be bad. But, as (33-35) show, they are
acceptable. These facts are consistent with the preceding arguments for
classifiers as verbs.
182 l
__ rtm_
(33) CAR jCL;3.
'The car is parked there w1thout any apparent problem.'
---tn-----(34) TREE jCL;l-LEAN.
'The tree 1s leaning over awkwardly.'
--Il'I!I--
{35) l_hand: iCl:4 ------------___ em_
R_hand: FENCE CAT kCl:V.
'The fence was just there, and there was th1s cat s1tt1ng right next to the fence.'
183
The preceding arguments have shown that 1n sequences of Noun +
Classifier. the classifier is a verb and bears the predicate relation to
its clause. The following argument showS that 1n such sequences. the
noun bears the 1 relation to the clause. thus providing additional
evidence along with the argumerlt with THINK complement structures that
such sequences ·are clausal.
7.1.3 An argument for the I_hood of the sentence~1nittal nomtnal:
Modals 4
Chapter S contains a discussion of modal structures in which 1
argue that only Is can precede modals in a modal structure. If it 1s
the case t~at ·'ocative object- sequences such as (36) below are
monoclausal, then modals should not appear within the oblique noun
phrase. but only following the subject norrdnal. but as (36)
demonstrates, modals can appear following either noun.
184
(36) L-hand; jCL;8 -------------- ---
R-hand: TABLE MUST
'The table has to be pos1tioned here so the monkey can jump
onto 1t.'
Under an analysts of (36) as monoclausal and the initial sequence
of Noun + Classifier as an ohl ique nominal. an additional compl kat Ion
wOuld be needed to account for sentences like (36) in which the nlOdal
appears ~ a noun phrase. Such a rule would not only be ad hoc,
but highly suspect.
But under a mult1~clausal analysis, no extra complication is
needed to account for (36). The sentence-lnitial nominal, TABLE hears
the 1 relation to the clause, thus can precede the modal in a modal
structure. likewise. since MONKEY heads the 1 relation tn the second
clause, it also may precede the modal.
On the basis of this argument. along with the preceding argL~ents
showing that these classifiers are predicates, t conclude that these
structures are not monoclausal. but multi-clausal sequences.
7.2 Sign Order
I have argued elsewhere that constituents tn ASL clauses are
ordc!red: Final 1 - Predicate - 2 - Nonterms. lhe fact that Is 1n the
follO\Jing clauses must be ordered before the classifier predicate is
consistent with the sign order rule I have proposed. Sequences like
(39-40) are unacceptable unless the classifier appears 1n a
nominalizat1on. as 1n the relative clauses 1n (41-42).
(37) CAR ,CL:3.
'lhe car is parked there.'
(38) CHU~CH iCL:A.
'Th~ church is up there.'
(39) ',CL:3 CAR.
'Th~ car 1s parked there.'
(40) ·iCL:A CHURCH.
'The church is up there.'
------rc-~ (41) THAT ,CL:3, a·u-s, NOT T-R-U-C·K.
'loihat was parked there was a bus, not a truck,'
------rc---(42) WAT ,CL:A, CHURCH.
'What is up there is a church.'
Whereas the monoclausal analysis of complex classif1er sequences
required a second. separate s1gn order rule to account for the order of
185
constituents, the present multi-clausal analysis needs no adJ1tional
rule. The sign order rule presented above correctly accounts for the
facts about ordering of const1tuents within the ASL clause. including
turn now to the question of what status indices bear in sentences like
(95~97), repeated below:
(95) BOY jINDEX. GIRL jINDEX. jKICKj •
'The boy kicked the girl.'
(96) GIRL jINDEX. BOY jlNDEX, jXICKj •
'The boy k1cked the girl.'
(97) jXICKj BOY jINDEX. GIRL jiNDEX.
'The boy kicked the girl.'
If the indices in the above sentences do not head P-arcs. then
classifier verbs cannot be SUbstituted for them. 8ut, as (115-117) show,
classifier verbs may be substituted for the indices:
(115) BOY jCL:V, GIRL jCL:V,jKICKj'
'The boy kicked the girl.'
(116) GIRL ICl:V, BOY jCl:V, IKICKj.
'The boy kicked the girl.'
(117) jKICKj BOY jCL:V, GIRL jCL:V.
'The boy kicked the girl.'
On the basis of the above facts, 1 propose the following multl-
200 1
J
clausal analys1s for sentences like (95-97) 1n which the indtces bear
the predicate relation to the clause:
'The boy 1s here; the g1rl 1s there; he kicked hera '
'The boy was standing herei the g1rl standing there; he kicked
her.'
7.4.3 S1gn order
On the basis of the above data. I propose that no speCial
cor,dition on s1gn order 1s needed for clauses containing Inflecting
verbs, but that Sign order 1s, as previously stated:
(llS) Sign ~
Order of constituents 1n ASL 1s: Flnat 1 - Pred1cate - 2 _ t\onterms.
Sentences like (95-97) do not present counter_examples to the
above constraint; I have argued that the sequences of Noun + Index in
those sentences are not nominals. but clauses. and within the clauses
themselves, order 1s constrained according to (liB).
207
(118) predicts that 1n sentences ltke (119) below which do not
v10late the sign order rule, indices may e1ther be predicates or
determiners. In some sentences, we find that classifier verbs cannot be
substituted for tndices. as in (I20). but if the sequence 1s
multi-clausal, and a clear boundary pause or marker (the sentence
adverbial, D:WRONG[+1dtomat1c] 'Suddenly') appears between clauses. then
classifier verbs can be substituted for indices, as in (I2l).
(119) BOY jlNOEX jKICKj.GIRL jINOEX.
'The boy k1cked the girl.'
(120) 1'BOY jCL:V jKICKj GIRL jCL:V.
'The boy kicked the g1rl.'
(121) BOY jCL:V. O:WRONG[+idjom.tjc). JKICK j • GIRL jCL:V
INNOCENT.
'The boy was standing there then suddenly he kicked her. The
girl was just standing there,'
In ASL grammar. the status of indices has been problematic.
Previous analyses have suggested that they are either determiners or
some type of specifier. 1 presented data here showing that at least
some sequences of Noun + Index c~nnot be analyzed as nominals. but must
be clauses. In complex sequences containing several i(ldices, Such as
(95-97). it 1s not clear from the sentence itself whether the indices
are determiners or predicates. If some indices are predicates, then
208
substitution of similar predicates should be possible, and the fact that
such substitutions are allowed supports the analysis. Further, In cases
where lnd1c:es are not predicates. then substitution should be
disallowed. This is also supported. The test of substit~tion provides a
means of distinguishing between indices which bear the predicate
relation and those which do not. As a consequence, facts about sign
order could be re-examined. once the clause structure of sentences
conta1(11n9 indices Is established. 1 propose that no special condition
Is ne~ded to account for sign order In clauses containing Inflecting
verbs. but that a general sign order rule applies across different
sub-classes of verbs.
This chapter has examined a little-studied area: the syntactic
structure of sentences containing classifier verbs and indices. Using
facts about I-hood and predicatehood established 1n Chapters 4 and 5.
various tests were applied to sentences containing classifier verbs and
indices, I have argued that. on the basis of these tests, the position
that certain complex sequences containing classifiers and indices are
monoclausal cannot be supported. If these Isentences are reanalyzed as
multi-clausal sequences. then facts about sign order within the clause
can be stated more generally. across different verb classes.
209
Footnotes
1 Sentences (5-7) do not contain topic marking and are unambiguous in meaning.
2 If the former is the case. the individual clause nodes would head arcs with a superordinate node as tail. i.e the structure below: (i)
3 In his description of classifier roots, Supal1u (1982) distinguishes between "stative" and "contact" (location) roots. In this chapter, the following classifier forms: Cl:V, Cl:3. Cl:V. Cl:l, Cl:A contain contact roots.
4 Another argument for subjecthood. antecedents for SELF pronouns, cannot be used with these sequences for the reason that the pronouns, 'to do by oneself' cannot be used with statives. e.g.
(i) 'DOOR SELF ODOR-CLOSED.
'The door was closed by itself.'
(ii) "\lOMAN SELf HAPPY.
'The woman was happy by herself.'
(iii) 'CAR Sl:LF iCL :3•
'The car was parked on its own.'
5 While structures with tOhicalized obliques are ruled out. as shown in (63~6a), structures Wlt lett dislocation are allowed:
6 The gloss: INDEX encompasses at least two distinct index forms: those ~hlCh involve a short directional and contact movement, and those which do not. It appears that typically indices which bear the predicate relation involve the directional and contact movement, wh1le determiners do not. In free conversation, however, the distinction 1s often neutralized. For this reason, this chapter examines evidence for distinctions among tndices other than dHferences in form.
211
.. ~
ASK BAWL-OUT BEAT BEG BID BITE BLAME "BORROW BOTHER GIFT GIVEI CAPTURE CATCH COMMAND CONTACT CONVINCE "COpy CRlTlCIZE DEFEAT *EXTRACT FEED FINGERSPEll FLATTER FORCE HATE IlELP IGNORE INFLUENCE INFORM INSULT -INVITE KISS LOCK-IlORNS LOSE-CONTACT
DROP 2 EhT n:COURAGE fNJOY D[RCISE fORGET GliE 55 HW::: BOPE INSERT INTERPRET Jl!fJGE UIUGU LIKE llPRiAO
Appendix a
Examples of Plain Verbs
LIVE LOVE MEMORIZE PROMOTE REQUEST SERVE SET -UP SIGN SING STOP SUGGEST SUPERVISE SUSPECT TEMPT THANK THINK THROW-AWAY VOTE WALK WANT WASH WONDER WORK YAWN YELL
: :::::;;::: ...:::':: ::: 213
\
BRING/CARRY CARRY -B Y -HAND EAT -UP /CORROOE EXAMINE GO-AWAY GO/COME INSERT MOVE MOVE-AWAY POINT -TO SCRUB TRACE WRITE
l CL :A l CL :C l CL : CC lCL :1 l CL :Il ·CL:V ~Cl:V lCL : VV l CL : VV l CL : 3
Alssen. Judith 1979. Possessor Ascension tn Tzotzl1. In l. Harttn (ed.) Papers ~ Mayan Linguistics. Columbia. MO: lucas.
A1ssen. Judith 1983. Indirect Objpct Advancement in Tzotzi1. In D. Perlrr.utter (ed.) Studies in Relational GriUI'.mar L Chicago. Ill1nois: Un1vers~CfiTcago Press. ----
Allen, Barbara J. and Donald Frantz 1983. Advancements and Verb Agreerr.~nt 1n Southern Tiwa. In n. Perlmutter (ed.) Studies 1n RE'lational Grammar 1. Chicago. Illinois: University"""Of""'CfiTcago Press. -
Fillmore. Charles J. 1968. The Case for Case. Universals 1n Linf~1stic The0!l. ed. by Emmon Bach and Rohert I. Harms. Yor : Holt,liii1nchart, and Winston.
Fillmore, Charles J. 1971. Some Problem!> for Case Grammar. C~orgetown t"onograph Series in language and Linguistics 24,
New
215
ed. by Richard J. O'Brien. Wasillngton: GeOrgetown Unlvers1ty Press.
Fischer, Susan 1975. Influences on Word Order Change in ASL. Word OrUer and Word Order Change., ed. by C. l1. Austin, TX: Unlvers1ty "01'""TexasPress.--
Fischer, Susan and Bonnie Gough 1978. Verbs in Plnerican Sign language. Sign language Studies 18.17-4R.
Friedman. lynn 1975. Space, Time. and Person Reference in Pmerlcan Sign language. language 51.940-961.
Friedman, lynn 1976. The Manife$tation of Subject. Object, and Topic 1n fmerican Sign lanyuage. Suhject and Topic, ed. by C. l1. Nevf York: Academic Press. --- --
-,_.'
Gruber. Jeffrey 1976. North-Holland linguistic Series 25: lexical Structures in Syntax and semantics. hUsteraaiii'!llofIJ\-HO'TlTriO pub\ishlng to.~pany.
Jackendoff. Ray S. 1972. Semantic Interpretation in Generative ~. Cambridge. ~A: MIl Press.
Keg1. Judy 1978a Indexing and pronominalization in ASL. unpublished paper.
Kegl, Judy 1978b ASL agreement. Unpublished paper.
Klima, Edward and Ursula Bellugi 1979. The Signs of language. Cambridge. M.A.: Harvard University Press. -
Kuno, Susumu 1973. The Structure of the Japanese language. Cambridge, MA: MIT Press. -
lacy. Richard 1974. Putting Some of the Syntax Back into Semantics. Paper presented at the linguistics Society of America Annual Meeting, New York.
Ltddell. Scott 1978. M Introduction to Reli!tlve Clauses tn ASl. Understandin,9, lrt9Uyge Thp.0ut Si9'P Language Rescarch. ed. !::ly Patnca SIple. ew ark: ca emlC ress.
Liddell, Scott 1977. M Investigation into the Syntactic Structure of Pmerlcan Sign lan9uage. Unpublished doctoral dis'iertation. University of California, San Diego.
McIntire, Marina 1980. locatives in ASL. Unpublished doctOf<1l disseration, University of California, los Angeles.
116
Meier. Rich<lrd 1982. Icons, Malo9ues nnd ~'orphf"mes: The Pc.quisitlon of Verb Agreement in ASl, Unpublished doctoral dissertatio'1, University of California, San Diego.
Newport, Elissa 1931. Constraints on structure: Evidence trom Pmerican Sign language and language learning, ed. by W.A. Collins Minnesota Symposia on Child PsycholoQY 14 .'l1lsdale. NJ: Lawrence Erlbaum Assoclates.--- -
Perlmutter, David M. 1979. 'Predicate': A Grammatical Relation. linguistics Notes ~.!:! ~ 6.
Perlmutter, Oa ... id M. 1980. fie 1 ilt ional Grammar. Synt(lx and s..~mant its 13: CUl'rent Approaches to Syntax, ed. by Ed1th~ravcsn.-3iia--JeSs "iCa""1ffrth. New York: AcademIC Press.
Perlmutter. David M. and Paul H. Postal 1983. $oml:! Proposed laws of Basic Clause Structure. Studies in Relational Grammar I, ed. by David M. Perlmutter. University or Chlcago Press. -
Perlmutter, Oavid M. to appear. Syntactic Representation. Syntactic ltvels, and the Notion of Subject. The Nature of Syntactic Representation, ed. by Pauline Jacobson ana""GCoTfrey Pullum. Boston: Reldel.
Ross. John R. 1967. Constrafnts on Variables in SYntax. Bloomington: Indiana University l1ngulstics Club.
Supalla. Ted 1982. Acquisition of Verbs of Motion and location 1n ASL. Unpublfshed doctoral dissertation. University of California. San Diego.
Supal1a, Ted and Elissa Newport 1978. How Many Seats in a Chair? The Der)vation of Nouns and Verbs in Pmerican Sign language. Unrlerstandinq .~an91.Jage Thr'OuQh Si 9p Language Research. ed. by l'otrlca Jlpje.~~mlc ,ass.
Wilbur, Ronnie B. H179. Anerfcan _Stgn. Language and ~ Systems. Baltimore. MD: Universlty Pa~ress. .
Woodward. James 1973. Interrule Impl1cation in hner1can S1go Language. Sisn languag~~ 3.47.56.