Bare nominals, information structure and word order

Bare nominals, information structure and word order

Murad Salem

Wayne State University, Department of Classical and Modern Languages, Literatures and Cultures, Detroit, MI 48202, United States

1. Introduction

One observation often made about bare nominals (henceforth, BNs) concerning their distribution has been the subject ofsome lively discussion in the literature. BNs in some Romance languages, such as Spanish and Italian, exhibit an asymmetryin distribution, where they appear to be licensed in postverbal positions only, unless they are in focus, modified, orcoordinated, in which case they can appear preverbally.1 English BNs, on the other hand, are not so constrained and candistribute freely in pre- as well as postverbal positions.

Another observation is that the informational status of English BNs follows from their interpretation (Laca, 1990; cf.Cohen and Erteschik-Shir, 2002): existential BNs are focused and generically interpreted BNs are topical.

The present paper pursues a solution to the first observation in terms of the second: the distributional differences betweenBNs in a number of languages can be argued to emanate from the interplay between the two systems of focus andword order.

One language where BNs have not received attention thus far is Arabic, in which BNs distribute, and are interpreted, on apar with their counterparts in Spanish/Italian (however, see below for more about Italian). The current study seeks to shedlight on Arabic BNs and the conclusions in the discussion to follow can be summed up in a,b.

a. Arabic BNs, which are always existential, are in focus, in linewithwhat has been argued to be the case for existentiallyinterpreted BNs in English.

b. Differences in the properties of the systems of focus and word order in Arabic and English can explain the surfacedifferences in the distribution of BNs. English has a rigid word order, but a fairly flexible focus system; the oppositeholds true in Arabic. The interaction of these two systems results in the observed constrained distribution of ArabicBNs and the freer distribution of their English counterparts.

Lingua 120 (2010) 1476–1501

A R T I C L E I N F O

Article history:

Received 29 December 2007

Received in revised form 4 November 2009

Accepted 6 November 2009

Available online 16 December 2009

Keywords:

Bare nominals

Focus

Topic

Word order

NSR

Interface

A B S T R A C T

Existentially interpreted bare nominals in English have been argued to be in focus; generic

bare nominals, topical. In distributional terms, these nominals distribute more freely as

compared to, e.g. Romance bare nominals. Arabic bare nominals pattern with the latter in

interpretation and distribution. This paper proposes that Arabic bare nominals, which can

only be interpreted existentially, are foci, on a parwith their English counterparts, and that

this plays a key role in their licensing. More importantly, the present paper seeks to derive

the surface differences seen in the distribution of these nominals in Arabic and English

from deeper differences between these two languages in the focus system facts and word

order possibilities. Arabic has a freer word order, but is more constrained in its focus

system. The reverse holds in English, a fact that explains why English bare nominals are

more freely distributed than their Arabic counterparts.

� 2009 Elsevier B.V. All rights reserved.

E-mail address: [email protected] See, among others, Casielles-Suarez (1997), Contreras (1986), and Suner (1982) for Spanish, and Longobardi (1994, 1996, 2000, 2001) for Italian.

Contents lists available at ScienceDirect

Lingua

journa l homepage: www.e lsev ier .com/ locate / l ingua

0024-3841/$ – see front matter � 2009 Elsevier B.V. All rights reserved.

doi:10.1016/j.lingua.2009.11.002

The rest of the paper is structured as follows: in the next section I review the argument that existential BNs in English are infocus and generic BNs are topical. In section 3, I demonstrate thatArabic BNs are foci, on a parwith existential BNs in English. Inthe same section, Spanish BNs are briefly examined to show that they too are amiable to the same treatment. Section 4 looks atthe argument put forth to explain the licensing of (Italian) BNs in terms of lexical government and it is concluded that thisargument is problematic at best. An alternative proposal is made in the same section, based on the idea that BNs are licensedeither by being marked as foci, in the case of Arabic BNs and their existential English counterparts, or as topics, in the case ofgeneric BNs in English. Section 5 explores the distributional differences between Arabic and English BNs and it is argued thatthese differences are explainable in terms of the two systems of focus and word order in these languages. Section 6 is theconclusion.

2. Existential BNs in English and focus

In this section, I briefly review the evidence brought to bear in support of the argument that the information status ofobject BNs in English is determined by the interpretation a BN receives (Laca, 1990).2 Additionally, more data will be offeredto generalize the argument to BNs in other structural positions as well.

Laca (1990) proposes that inclusive (i.e. generic) BNs are presupposed, but that non-inclusive (that is, existential) BNs areasserted or focused. The author notes that there will be a blocking effect, seen in (1), when the BN is interpreted generically:that the Gwamba-Mamba eat salmonmeans that they eat fish, since salmon in (1a) is interpreted existentially; however, if theGwamba-Mamba hate salmon, it does not necessarily hold that they hate fish. Salmon in (1c) is therefore interpretedgenerically.

(1) a. The Gwamba-Mamba eat salmon.

b. The Gwamba-Mamba eat fish.

c. The Gwamba-Mamba hate salmon.

d. The Gwamba-Mamba hate fish.

Consider now the effect predicative adjectives have on the interpretation of a BN. In (2a), if Mary wears silk blouses, itfollows that she wears blouses; the inference goes through and the BN is interpreted existentially. However, if she wears silkblouses unironed, it does not follow that she wears her blouses unironed, which says that silk blouses in (2c) is interpretedgenerically. The predicative adjective in (2c,d) attracts the focus in the sentence and the generic BN is presupposed.3

(2) a. Mary wears silk blouses.

b. Mary wears blouses.

c. Mary wears silk blouses unironed.

d. Mary wears blouses unironed.

(3) The Gwamba-Mamba eat salmon with their fingers.

(4) The Gwamba-Mamba eat fish with their fingers.

Manner adverbials behave in a similar fashion as seen in (3)–(4). Notice that the inference in (4) does not follow from (3):that the G-M eat salmon with their fingers does not entail that they eat all their fish that way. Salmon in (3) is interpreted aspart of the presupposition whereas the manner adverbial is in focus.4

Laca’s argument can be generalized to BNs in non-object positions as in these examples from Partee (1991)5:

(5) a. TICKETS are available.

b. Tickets are AVAILABLE.

2 Cohen and Erteschik-Shir (2002) propose that the opposite also holds true; the focus structure determines if the bare plural is existentially or generically

interpreted. Focused bare plurals receive an existential interpretation whereas topical bare plurals the generic interpretation.3 That the BN, silk blouses in (2c,d), is presupposed becomes clear if these sentences are negated as in (i)–(ii). The presupposition is retained under

negation and it is the predicative adjective that is negated.(i) Mary does not wear silk blouses unironed.(ii) Mary does not wear blouses unironed.

4 Laca notes that the so-called ‘‘affective verbs,’’ verbs such as like, admire, hate, pose a problem to her general argument in that their objects can be

focused, despite their generic interpretation. The reader is referred to Laca (1990) for the argument she puts forth w.r.t. affective verbs. What I can say,

however, is that the focused objects are interpreted as contrastive topics. So, in John hates athletes, there seems to be implicit contrast between the people or

things that John hates and those that he does not hate. I would speculate, however, that the contrast is inherent in the meaning of the predicate, to which a

reviewer objects, saying that we could answer the question what does John like? with Nothing!, which argues that contrast is not inherent in the lexical

meaning of the predicate. Again, it could be assumed that even here, there seems to be an implicit contrast between (liking) something and (liking) nothing.

Finally, that the object of an affective verb is a contrastive topic is further illustrated by the unacceptability of BNs with a generic reading in languages such

as Arabic, and Romance, with these predicates, unlike English BNs which can serve as generic objects to these predicates (see Dobrovie-Sorin, 2009).5 The examples illustrate, as suggested also byDiesing (1992) and Kratzer (1995), that subjects of stage-level predicates are ambiguous between a generic

and an existential reading. For Carlson (1977a,b), on the other hand, only existential interpretations are possible with stage-level predicates.

M. Salem / Lingua 120 (2010) 1476–1501 1477

(6) a. *TICKETS are expensive.

b. Tickets are EXPENSIVE.

(5) and (6) differ in terms of the predicate used, a stage-level (s-level) predicate and an individual-level (i-level) predicate,respectively. In (5), both the BN and the predicate can be focused. In an out of the blue utterance, the BN in (5a) is interpretedexistentially, whereas it is generic in (5b) with the predicate in focus. In (6), on the other hand, the predicate alone can benon-contrastively focused and Tickets is read generically.6,7

Krifka et al. (1995:13) make an observation similar to Laca’s concerning the blocking effect seen with generic BNs. Theynote that in upward-entailing (that is, non-negative) contexts, object-referring (existential) DPs can be replaced with lessinformative DPs as in (7), without affecting the truth-conditional status of the sentence. Kind-referring (generic) DPs, bycontrast, do exhibit such a difference in truth-condition, as seen in (8).

(7) Berber lions escaped from the zoo. ) Lions escaped from the zoo.

(8) Berber lions are extinct. #) Lions are extinct.

Since an s-level predicate is used in (7), the BN can readily be interpreted existentially. An appropriate context question(CQ) for (7) is (9), in which Berber lions is focused as it replaces the wh-word.

(9) What escaped from the zoo?

However, (10), is not an appropriate CQ for (7): the BN in (10) is likely to be interpreted generically (kind-referring),resulting in an awkward (perhaps, unacceptable) question due to the clash between kind reference and the s-level predicateescape.8

(10) # Where did Berber Lions escape from?

Consider next that (11), unlike (12), is not a possible CQ for (8) (reproduced below for convenience), which suggests thatthe BN in (8) is not amiable to a reading where it is (non-contrastively) focused9,10:

(8) Berber lions are extinct.

6 The BN in (6a) can be focused as a contrastive topic, meaning, for example, that ‘‘TICKETS, not POSTCARDS, are expensive.’’7 Kiss (1998) distinguishes between information focus and identificational (contrastive) focus on both semantic and syntactic grounds. Semantically,

identificational focus identifies a subset among a set of elements given in the situation or context of discourse which can potentially hold of the predicate

phrase and that subset is exhaustively identified as the one for which the predicate holds. So, for example, of all the individuals present in the domain of

discourse, Mary is identified exhaustively as the one whom I introduced Peter to in (i).(i) Tegnap este Marinak mutattam be Petert. (Hungarian)

last night Mary.DAT introduced.I PERF Peter.ACC‘it was to Mary that I introduced Peter last night’

Unlike identificational focus, information focus fails tomake any exhaustive identification of a subset in a set of potential discourse elements, but rather

communicates new or nonpresupposed information as in (ii).(ii) Tegnap este be mutattam Petert MARINAK.

‘last night I introduced Peter TO MARY’Some of the other differences noted by Kiss between identificational and information focus can be summed up in (iii)–(vi).

(iii) Every sentence has information focus, but not identificational focus.

(iv) Constituents that can function as identificational focus are more restricted but those that can be information focus are not restricted as such.

(v) Identificational focus involves movement to the specifier position of a functional phrase but information focus does not.

(vi) Identificational focus takes scope whereas information focus projects.

Contrary to Kiss, I will assume that both kinds of focus involve movement to a specifier position in Arabic.

For a different view of focus and focus movement (in Hungarian), see Horvath (2006).8 In (10), Berber Lionsmay still be open to an existential interpretation in which case the question would have two focuses,where and the BN, triggering a

pair-list reading.9 Notice, however, that (12a) and (12b) are not equivalent. (12a) gives rise to a taxonomic reading where Berber Lions is interpreted as a contrastive topic

(i.e. BERBER LIONS are extinct (as opposed to some other animals)). In (12b), on the other hand, the predicate extinctwill be focused and the BN is clearly topical

as in (i); no taxonomic reading is available in (i) though.

(i) Berber Lions are EXTINCT.10 A similar treatment can be given to (i), which can be interpreted either as in (ii) or (iii), also from Krifka et al. (1995:24).

(i) Dogs must be carried. (Sign in front of the escalator)

(ii) A dog must be carried to use the escalator.

(iii) Any dogs on the escalator are to be carried.

To get the funny reading in (ii), and the intended meaning in (iii), the relevant CQs are, respectively:

(iv) What must be carried (to use the escalator)?

(v) What about dogs? What do you know about dogs?

If dogs in (i) is read existentially, it will be focused and the relevant CQ is (iv). If it is interpreted generically, it will be unfocused as indicated by the fact

that dogs in (v) is part of the presupposition and we get the intended reading in (iii).

M. Salem / Lingua 120 (2010) 1476–15011478

(11) #/*What/Who is extinct?

(12) a. What kind(s) of animal(s) is/are extinct?/Which animal is extinct?

b. What do you know about Berber lions? What about them?

Rooth (1992) points out that the focus structure of a sentence can affect its truth conditions. (14) is not a good answer to(13) since Bill is presupposed in the latter but in focus in the former (the diacritic F indicates focus).

(13) Who cut Bill down to size?

(14) *Mary cut [Bill]F down to size.

Now consider the following set of data:

(15) a. *Who cut boys down to size?11

b. Who cut the boys down to size?

(16) ?*Where did boys go?

(17) ?*Why did girls leave early?12

(18) ?*When will boys leave?

The predicates in (15)–(18) are s-level, which should allow for existential interpretations of the BNs. The question in (15a)is ungrammatical. An appropriate answer to (15a) should take the form x cut boys down to size, where the BN boys ispresupposed, a fact that leads to ungrammaticality as is further confirmed by the grammaticality of (15b) that contains adefinite DP, which can be (in fact, is, see section 4.3) presupposed.With i-level predicates, on the other hand, where a BNwillbe interpreted generically, the CQs in (19)–(21) are felicitous because there is only one focus per sentence, namely the wh-word; the generic BNs are presupposed.

(19) Why do boys hate school?

(20) What do girls like?

(21) What do boys smoke?

There is, then, sufficient evidence to support the claim that existential BNs in English are focused and genericallyinterpreted BNs are topical. In the next section, I demonstrate that BNs in Arabic are also in focus. This will in turn serve as astepping stone to mymain argument in section 5, which is that the differences seen in the distribution of BNs in English andArabic can be explained in terms of differences in focus and word order.

3. Arabic BNs are foci

To minimize cross-linguistic variation, the null assumption should be that BNs in Arabic, which are existential, ought tobehave on a par with existentially interpreted English BNs with respect to focus. Themain task in this section is to show thatthe null assumption actually holds and that Arabic provides further evidence in support of Laca’s (1990) general argumentdiscussed in section 2. However, I first turn to a brief discussion of the Nuclear Stress Rule in Zubizarreta (1998), which willbe crucial in the investigation of Arabic BNs.

11 This question is acceptable on a pair-list reading as John cut boys down to size, Mary cut girls down to size, etc.12 A reviewer notes that the question in (i) is acceptable, even though the BN is focused, alongside when.

(i) When do we throw confetti?

It seems to me that the prominent reading for confetti in (i) is the generic reading. Consider a possible answer as in (ii).

(ii) We throw confetti when we celebrate New Year’s.

I find the existential reading hard to get in the Arabic counterpart of (ii) and a definite nominal with the definite article shows up indicating that only a

generic reading is available. In the event where confetti is intonationally made prominent, it will be interpreted as a contrastive topic and the question will

have a pair-list reading.

M. Salem / Lingua 120 (2010) 1476–1501 1479

3.1. The Nuclear Stress Rule (NSR) and prosodically motivated movement (p-movement)

The Nuclear Stress Rule (NSR) in Zubizarreta (1998) has two versions, S(electional)-NSR, and C(onstituent)-NSR,depending on how prominence is calculated, either based on the selectional ordering or c-command, respectively. InRomance, it is only the C-NSR that is operational and phrasal prominence is assigned to the lowest constituent in terms ofc-command.13 The C-NSR can be defined as follows (Zubizarreta, 1998:124):

(22) C-NSR

Given two sister nodes Ci and Cj, the one lower in the asymmetric c-command ordering is more prominent.

Phrasal prominence can also be assigned by another rule, namely the Focus Prominence Rule (FPR), which Zubizarretadefines as in (23) (1998:21).

(23) FPR

Given two sister nodes Ci (marked [+F]) and Cj (marked [�F]), Ci is more prominent than Cj.

Romance andGermanic differ in how each resolves the conflict thatmay result from the operation of the NSR and the FPR.Unfocused material is treated as invisible in Germanic whereas the conflict is resolved between the NSR and the FPR inRomance bymoving the unfocusedmaterial leftwards thus placing the focused constituent in the lowest position in terms ofc-command. This movement, which is not motivated by feature checking (as in Chomsky, 1995), but rather by prosodicconsiderations, Zubizarreta calls prosodically motivated movement (p-movement).14

The discussion raises two questions: what module of the NSR is operational in Arabic, the C-NSR or the S-NSR? Secondly,can p-movement be attested in this language? The examples in (24) suggest that Arabic follows the Romance pattern in onlyallowing the C-NSR module.15

(24) a. *l-Tifil bibki.

the-baby (is)-crying

b. l-Tifil bibki

(25) The baby is crying.

(26) L-TIFIL bibki (miʃ IMMO)

THE BABY (is)-crying (not HIS MOTHER)

(27) #bibki l-Tifil (miʃ IMMO)

L-Tifil ‘the-baby’ in (24a) cannot be non-contrastively focused in a sentence-internal position (that is, a position that is notthe lowest in terms of c-command). In otherwords, neutral, non-contrastive focus, assigned by the C-NSR, is only available inthe lowest position in the c-command ordering as in (24b), unlike what we see in the English example in (25). What (26)illustrates is that focusing a constituent sentence-internally has to be contrastive, assigned by a prominence rule other thanthe C-NSR (capitalization indicates contrastive, sentence-internal stress). Finally, (27) demonstrates that the stress assignedby the C-NSR is ordinarily neutral. The answer to the first question, therefore, is that the C-NSR alone is operational in Arabic.As for the second question, whether p-movement is attested in Arabic or not, the answer is in the positive.

As remarked earlier in reference to Romance, the conflict between the C-NSR and the FPR in Arabic is resolved bymovingthe defocalized constituents to a higher position in the syntactic tree. Therefore, a felicitous answer to (28), where the subjectis focused, with a full sentence, is (29), rather than (30).

(28) miin ʔaʕ Tatak l-xubiz?

who gave-3FS-2MS the-bread

‘who gave you the bread?’

13 Zubizarreta finds evidence for the operation of the C-NSR alone in Spanish by noting the following contrast with English (focused constituents

underlined in Spanish, accented in English):(i) a. *El bebe llora. vs b. El bebe llora

c. The baby’s crying.14 If focusing a constituent, ormarking it as a topic, involves checking a focus and a topic feature, respectively, as argued by Rizzi (1997), then, p-movement

can still be viewed as functionally motivated by feature checking. A topical BN is shifted leftwards to check its topical feature in the specifier position of a

topic phrasewith a Top head and a focused BN checks its focal feature in a spec-head configurationwith a focus head.More on this will be said in section 4.3.15 I follow Zubizarreta in indicating a constituent stressed by the C-NSR by underlining it.

M. Salem / Lingua 120 (2010) 1476–15011480

(29) ʔaʕTatni l-xubiz Salma

gave-3FS-1MS the-bread Salma

‘Salma gave me the bread’

(30) #Salma ʔaʕ Tatni l-xubiz

(31) *SALMA ʔaʕ Tatni l-xubiz

S. gave-3FS-2MS the-bread

The question in (28) places narrow focus on the subject of the verb a Tatak ‘gave-3FS-2MS’, which entails that the subjectin the answer should be non-contrastively focused in the lowest position in terms of c-command where the C-NSR isoperational; this is exactly what we see in (29). The word order in (29) is VOS, whichmeans that p-movement has applied tothe underlying structure, be it either SVO or VSO, placing Salma in a position to be assigned stress by the C-NSR.16 Theconflicting demands of the C-NSR and the FPR are resolved by p-moving the defocalized constituent, the object l-xubiz ‘the-bread’, leftwards, leaving the subject Salmawithin the scope of the C-NSR.17 The unresolved conflict between the C-NSR andthe FPR leads to ungrammaticality, as seen in (31), where SALMA is contrastively focused but l-xubiz ‘the-bread’ is stressed bythe C-NSR. Finally, the SVO sentence in (30), where the object l-xubiz ‘the-bread’ is assigned stress by the C-NSR,marking it asfocus, cannot be an answer to (28) since the latter, as observed above, places narrow focus on the subject, and not the object(hence the sentence is marked with ‘‘#’’).

In conclusion, there is ample evidence that the C-NSR is operational in Arabic and that p-movement is attested in thislanguage. Next, I will use the preceding discussion of the C-NSR to demonstrate that BNs in Arabic are in focus.

3.2. Arabic BNs and focus

The notion of focus Imake use of in this work is the one assumed in Zubizarreta, 1998 (based on Chomsky, 1971, 1976 andJackendoff, 1972), in which focus is understood in terms of the discourse notion of presupposition as the non-presupposedpart of the sentence. The question/answer pairing is used to determine the information structure of the sentence, as in (32)and (33). The focused constituent in the answer replaces the wh-word in the context question (CQ).

(32) [What did John do?]

[John [F ate [the pie]]]

(33) [What did John eat?]

[John [ate [F the pie]]]

Let us now consider (34), where the BN banaat ‘girls’ occurs in a postverbal position.

(34) *rib -o banaat s-sibaq

won-3P girls the-race

‘girls won the race’

My claim is that (34) is unacceptable if banaat ‘girls’ is read under normal conditions of stress. Note that s-sibaq ‘the-race’is assigned stress by the C-NSR,marking it as focus, while banaat ‘girls’ is part of the presupposition. To anticipate, I will showthat the cause of the unacceptability of (34) is that the BN resists a reading in which it is presupposed.

Notwithstanding the ungrammaticality just noted, let us attempt to formulate an appropriate CQ for (34). Since s-sibaq

‘the-race’ is assigned stress by the C-NSR, it would be represented by a wh-word in the relevant CQ and a variable in thelogical representation, as seen in (35) and (36), respectively.

(35) *ʃuu ribħ -o banaat?

what won-3P girls

‘what did girls win?’

16 Zubizarreta argues that in Spanish the VOS order is derived from VSO by moving VP2 that contains the object to left adjoin to VP1 as in (i).

(i) [TP me regalo [VP1[VP2 la botella de vinok [ek [V2 ek]]]i [VP1 Maria [V1 [ei]]]]]

I leave the question as to what the underlying order of VOS is in Arabic open since it is not crucial for the argument pursued here.17 If we adhere to Kayne’s (1994) assumptions allowing only left adjunctions, to answer a concern raised by a reviewer, then the defocalized element will

move leftwards (to TopicP) then followed by the leftward movement of the focused element to the specifier position of a focusP (see section 4.3 for more

discussion).

M. Salem / Lingua 120 (2010) 1476–1501 1481

(36) there is an x, such that girls won x

the x, such that girls won x = the race

(37) ʃuu ribħ-o l-banaat?

what won-3P the-girls

‘what did the girls win?’

But (35) is ungrammatical. That its ungrammaticality can be attributed to the fact that the BN, banaat ‘girls’, is part of thepresupposition of the question, is further confirmed by contrasting it with theminimally different but grammatical (37)withthe definite article.

The initial attempt to produce an appropriate CQ for (34) has not proved to be a success thus far. I believe that noappropriate CQ can be produced for (34), due to the fact that a BN will always resist being presupposed or topical. To gathermore support for this conclusion, let me continue with the endeavor to generate a CQ for this sentence. Examine (38).

(38) miin ribiħ-o s-sibaq?

who won- won-3P the-race

Although grammatical, (38) could not be the right CQ for (34) due to the clash seen in the focus status of s-sibaq ‘the-race’in (38) and its status in (34): it is presupposed in the former but focused in the latter. In fact, a good answer to (38), with a fullsentence, would be (39), where the subject banaat ‘girls’ is assigned narrow focus as it substitutes for the wh-word in (38).18

(39) ribħ-o s-sibaq banaat

won-3P the-race girls

‘girls won the race’

Alternatively, (38) can be answered by left-dislocating s-sibaq to mark its topical status as in (40).

(40) s-sibaq, ribħ-o-o banaat

the-race, won-3P-3S girls

‘the race, girls won it’

The discussion lends support to the conclusion that a BN has to be focused and can never be part of the presupposition ofan utterance. The ungrammaticality of (34) follows automatically from this requirement. In fact, grammaticality is restored ifcontrastive stress is placed on the BN banaat ‘girls’ as in (41) which can be an answer to a CQ such as (42).

(41) ribħ-o BANAAT s-sibaq (miʃ WLAAD)

won-3P GIRLS the-race (not BOYS)

‘girls won the race, not boys’

(42) ribħ-o s-sibaq WLAAD?

won-3P the-race BOYS

‘boys won the race?’

In the same way, the ungrammaticality of (35), reproduced here for convenience, is also explained as stemming from thefact that a BN cannot be part of the presupposition.19

18 In (39), the focused constituent banaat ‘girls’ would be replaced by a variable x in the logical representation to indicate its focal status as in the x, such as x

won the race. Accordingly, there is no conflict between the status of s-sibaq ‘the-race’ in (39) and the CQ in (38) since it is thematic in both, nor is there

conflict in the focus status of banaat ‘girls’ because it is focused in both.19 V. Figueroa (p.c.) informs me that the CQs in (i) and (ii), which are structurally similar to (35) in Arabic, are unacceptable in Spanish either.(i) *

?

Que construyeron esclavos?what built slaves

(ii) *

?

Cuando llegaron muebles?when arrived furniture

Additionally, the following examples, with two focuses, are acceptable with the pair-list reading only, as has been noted for Arabic on a number of

occasions.(iii)

?

Quien quiere tortillas?who wants tortillas

(iv)

?

Quien construyo piramides?who built pyramids

These judgments suggest that what holds for Arabic BNs also holds for their Spanish counterparts. The facts in Spanish will be further discussed in

section 3.3.

M. Salem / Lingua 120 (2010) 1476–15011482

(35) *ʃuu ribħ-o banaat?

what won-3P girls

‘what did girls win?’

The question in (35) is ungrammatical because it has two focuses: the wh-word and the BN, banaat ‘girls’. Contrastivelyfocusing banaat would not resolve the problem. This line of reasoning is substantiated by the possibility of contrastivelyfocusing a BN in a question that contains no wh-word as seen in example (42) above.

If a BN in Arabic is indeed always in focus, this predicts that no two ormore BNs should co-occur since there is atmost onefocus per sentence (see section 5). The empirical evidence bears the correctness of this prediction out as in (43) (some of theexamples in (43) have an acceptable reading with a pair-list interpretation).

(43) a. *ʃaaf-o WLAAD banaat

saw-3P BOYS girls

b. *ʃaaf-o BANAAT wlaad

c. *WLAAD ʃaaf-o banaat

I conclude that Arabic BNs patternwith existentially interpreted BNs in English in being focused. In the next section, I willshow that Spanish BNs, withwhich Arabic BNs correspond in interpretation and distribution, are open to a similar treatment.

3.3. Spanish BNs

Suner (1982:125) argues that a presentational sentence (of the order VS) asserts the existence of the NP (DP) referent as in(44)–(45).

(44) A parecio un homre

appeared a man

(45) A somo el sol radiante

peeked (3sg) the sun radiant

‘the radiant sun peeked out’

That the DP referent is what is being asserted (or, more specifically, its existence) is seen when the sentence is negated(192 fn. 2):

(46) No aparecio un hombre, en realided no aparecio absolutamento nada

no appeared (3sg) a man, in reality no appeared absolutely nothing

The existence of the DP referent in (46) of a man and its participation in the act of appearing is negated. In other words,presentational sentences are existential. Both the verb and the DP in presentational sentences are in the scope of assertion, soboth can be rhematic/focal. In declarative sentences (with the order SVO), S is thematic while the VP is rhematic.

Suner (1982:212) further argues that only semantically intransitive verbs can be presentational. If BNs in Spanish areunacceptable in preverbal positions (under normal conditions of stress), and since in presentational sentences the subject isrhematic (and the verb may be), the expectation would be that BNs are perfectly acceptable in presentational (i.e.intransitive) sentences. What remains to be decided, however, is the status of transitive sentences and their interactionwithBNs: can transitive sentences be presentational? And, if so, can BNs freely occur in them? The answer to the first question,Suner argues, is negative. If the ultimate purpose of presentational sentences is to introduce the DP referent into theworld ofthe discourse, uniquely and exclusively, then the predictionwould be that transitive sentences, which characteristically havetwo DP referents, cannot be presentational. Suner also answers the second question in the negative: postposing a BN in atransitive sentence (that is, creating a presentational context) does not license the BN. This is exemplified by the followingexamples she provides:

(47) a. *Rumores llenaban el cuarto

Rumors filled the room

b. *Llenaban rumores el cuarto

c. *Llenaban el curato rumores

Since transitive sentences, Suner points out, have two arguments (V NP NP), ambiguity may result under normalconditions of stress and intonation (1982:213). It is not sufficient to say that (BN) subjects are barred from postverbalpositions in transitive (non-presentational) sentences, but rather that they are unacceptable under normal conditions of

M. Salem / Lingua 120 (2010) 1476–1501 1483

stress. Postverbal BN subjects are allowed if the focus structure of the sentence is clearly marked, as seen in her examples in(48)–(49).

(48) La cosecha la destuyeron langostas

the harvest (f.sg.) it (f.sg.) destroyed (3 pl.) locusts

‘the harvest, locusts destroyed it’

(49) El cuarto lo llenaban rumores

the room (m.sg) filled rumors

‘the room, rumors filled it’

The topical status of La cosecha ‘the harvest’ and El cuarto ‘the room’ is made abundantly clear and the focal status oflangostas ‘locusts’ and rumores ‘rumors’ is likewise marked.20

Suner goes on to point out thatmodification of a preverbal subject results in grammaticality, as seen in (50), and she notesthat such examples are contrastive in that (50) does not refer to the totality ofmen but to a subset of them ‘‘who are like that’’(1982:220–227).

(50) Hombres ası no debieran exister.

‘men like that ought not to exist’

The same argument can also be made about (51) and (52).

(51) Hombres de calidad estudian el asunto.

‘men of quality are-studying the matter’

(52) Cosas como esta desaniman a cualquiera

‘things like this discourage anyone’

Importantly, however, Suner (1982:256 fn. 24) notes that the subjects in (51) and (52) are all focal, as it is partially shownby the intonational pattern she provides for one of the examples. This means that Spanish BNs, even if modified, are stillfocalized.

As far as coordination is concerned, Suner (1982:229) notes the grammaticality of the following Spanish examples21:

20 Suner posits the Naked Noun Constraint (NNC) to filter out preverbal BNs in Spanish. The NNC could be recast in terms of the current study as follows: if

stressed constituents in Spanish (and Arabic) are positioned lowest in terms of c-command, then it would be safe to assume that the preverbal domain is

reserved for topicalmaterial. BNs are unable to showup in such a domain because they are focused. Postverbally, BNs are allowed since this domain iswhere

focused material is positioned anyway. However, if a BN is positioned postverbally, but not lowest in c-command, it gets a contrastive reading.21 A reviewer wonders why the coordinated bare singulars in (53)–(54), along with examples such as (i) and (ii), are acceptable, while uncoordinated bare

singulars, or modified bare singulars, are out.(i) Studenti e professori accompagnatori hanno effettuato visite guidate in varie localita

‘students and accompanying professors have made guided visits in several places’(ii) Ogni giorno dale 10 alle 12 studenti e insegnanti hanno potuto ritirare materiale divulgativo

‘every day between 10 and 12 o’clock students and professors could come to collect reading materials’A possible answer may be provided by the analysis proposed by Heycock and Zamparelli (H&Z) (2003). These authors argue that coordinated bare

singulars are in fact quasi definites and that these nominals require uniqueness to be licensed. They suggest that the coordinated phrase moves to the

specifier of an empty D which activates the DP resulting in the proper interpretation for definites. Bare plurals, along with their indefinite readings, can be

argued to have definite interpretations as in (iii).(iii) Una nave di rifugiati e arrivata in Puglia. Marinai *(e passeggeri) sono albanesi, il capitano e italiano

a ship of refugees has arrived in P. Sailora and passengers are Albanians the captain is ItalianH&Z also point out (p. 449) that unmodified bare plurals in Italian cannot be interpreted as kinds. However, the kind reading is available with

coordinated bare plurals since these have the semantics of definites.(iv) Cani *(e gatti) diventano piu comuni come si passa dal Sud al Nord Italia

dogs (and cats) become more common as one moves from South to North ItalyEven in French, which disallows bare plurals in argument positions, bare plurals are licensed by coordination with definite and kind readings.

(v) a. [Chiens *(et chats)] avaient tous l’air tres sale[dogs (and cats) had all the appearance very dirty

b. Le chien, le chat, et le lapin sont parmi les animaux domestiques les plus repandus; en particulier,the dog the cat and the rabbit are among the animals domestic most widespread in particular[chiens *(et chats)] son tune/des especes qui s’adapte(nt) facilement a la campagniedogs (and cats) are a/sm species which adapts easily to the companydes etres humainsof human beings

If H&Z’s analysis is correct, coordinated bare nominals (in English and Romance) may be different from uncoordinated (modified) bare nominals and

have licensing conditions different from those the latter nominals are subject to. Modified nominals, however, are subject to the focus requirement noted in

the text as seen in the discussion of examples (50)–(52) fromSpanish, where the examples set up contrasts between allmen andmen like that (50) ormen of

quality (51) and all things and things like this (52).

M. Salem / Lingua 120 (2010) 1476–15011484

(53) Cielo y tierra te contempian.

‘heaven and earth watch you’

(54) Tıa y sobrina IIegaron sin novedad anoche

‘aunt and niece arrived without problems last-night’

She points out that each of the conjoined elements receives strong stress. In otherwords, the act of conjoining BNs per se isnot the licenser and heavy accent is sill required.22,23

A question that is pertinent to both Arabic and Spanish BNs, and perhaps to BNs in other languages, is: if these nominals arein focus, how, or whether, being focused is instrumental in their licensing and/or distribution? I believe this to be the case. Inthe next section, I would like to take up the first issue of licensing and I argue that BN licensing is achieved through focus. Insection 5, I discuss the second issue of distributionwhere the similarities and differences seen betweenArabic vis-a-vis Englishare highlighted and the differences are attributed, in part, to the variation seen in how these languages assign focus.

4. Licensing of Arabic BNs

To shed light on how Arabic BNs are licensed, it is perhaps instructive to go over an approach to BN licensing that hasdominated the discussion of Romance BNs in the last two decades, namely, the lexical government approach, to which I nowturn.

4.1. Italian BNs

As noted in section 1, it is often observed that Romance BNs are ungrammatical in preverbal positions under normalconditions of stress, or if unmodified or uncoordinated.24 However, in postverbal positions, a BN is acceptable, even if it is notfocused, modified, or coordinated. The following examples, taken from Italian, illustrate this asymmetry (Longobardi,1994:616):

(55) *Acqua viene giu dalle colline

water comes down from the hills

(56) Viene giu acqua dalle colline

comes down water form the hills

(57) Ho preso acqua dalla sorgente

I took water from the spring

Longobardi (1994) is basically an attempt to motivate an N-movement to D in the syntax in Italian. Insofar as that ispossible, Abney’s (1987) DP hypothesis would turn out to be correct. In view of their apparent behavior, BNs could be arguedto be DPs inwhich D is subject to a government requirement by a lexical head.25 In preverbal positions, D fails to be governedas seen in (55). Postverbally, however, this structural requirement of D seems to be satisfiedwhich leads to grammaticality asin (56) and (57). The empty D is taken to be an existential operator imposing a plurality (count ormass) interpretation on thehead noun, hence the ungrammaticality of bare singular count nouns in Italian.26

Such a strictly syntactic account of the facts has its drawbacks. Even though it is a common observation that focusing a BNpreverbally licenses it, an explanation of this fact has remained elusive. Longobardi suggests a reconstruction analysis (1994fn. 10), where, after reconstructing the BN to a postverbal position, the lexical government requirement is fulfilled. Note,however, that the licensing of a BN in this case would not take place at S-Structure, but at LF, which, within Longobardi’s setof assumptions, is not the relevant level for the licensing of Italian BNs because the government requirement in this languagehas to be satisfied in the syntax. This is in fact what propels Longobardi’s argument of N-to-D raising of proper nouns overtly

22 Recall also the CQs mentioned in fn. 19 which all point to the fact that Spanish BNs are foci.23 However, see the discussion in fn. 21.24 For references, see fn. 1.25 Contreras (1986) explains the distribution of BNs in Spanish in terms of the Empty Category Principle (ECP). Spanish BNs are assumed to have the

structure [NP [QP e] N’] where the empty quantifier phrase has to be properly governed. Since the preverbal subject position is governed by Infl in Spanish,

Contreras suggests that preverbal subjects in Spanish have to be adjunctions rather than occurring in the canonical specifier positions. If the subject position

was governed in Spanish, Contreras argues, examples such as (i) would be predicted to be grammatical, which it is not. The ungrammaticality of (i) is given a

natural explanation as a violation of the ECP.(i) * [S [NP [QP e] cafe] INFL [VP me gusta]]

‘I like coffee.’Notice that Longobardi does not need to make the claim that Italian subjects are adjunctions since he assumes that government has to be satisfied by a

lexical head.26 This situation does not hold for Arabic since singular count nouns occur bare. Assuming that singular bare nominals in Arabic are also introduced by an

empty D, which seems to be conceptually reasonable, then the empty D in Arabic does not force such plurality requirement on the N it quantifies over.

M. Salem / Lingua 120 (2010) 1476–1501 1485

in the syntax in Italian.27 If the licensing of the empty determiner of a focused BNwas allowed to take place at LF this wouldbeg the question why, for example, unfocused preverbal BN subjects are not allowed to reconstruct back into the VP(assuming that subjects originate inside the VP). Moreover, the reconstruction analysis becomesmore problematic once thefacts in English are taken into consideration.

The commonly held view is that English BNs, unlike their Italian or Spanish counterparts, are unrestricted when it comesto their syntactic distribution or to their semantics. In terms of the semantics, English BNs are ambiguous between a genericreading and an existential one. In terms of their syntax, they appear to be free to occur in any structural position.

(58) I saw boys in the park. Existential

(59) Dogs were chasing after me. Existential

(60) Cheese comes in so many varieties. Generic

The facts in English are different from what we have seen in the Italian examples (55)–(57) above, or in the followingSpanish examples from Casielles-Suarez (1997)28:

(61) Jugabans ninos en la calle Existential

played-3pl children in the street

‘children were playing in the streets’

(62) *Ninos jugaban en la calle Existential

children played-3pl in the street

‘children were playing in the streets’

27 In fact, Longobardi (1996) generalizes this raising to a category of common nouns in Italian such as casa ‘home’, again in the syntax and not LF.28 A reviewer points out that BNs in Romance can be generic, as argued in Longobardi (2001). Schmitt and Munn (2003) observe that the interpretive facts

argued for in Longobardi (2001) donothold for Spanish,whichmeans that Longobardi’s claims cannot carry over to all Romance. Longobardi’s claims have also

been challenged by Zamparelli (2001) who points out that a kind reading in Italian is only possible with the definite article. More importantly, the facts in

Longobardi (2001) do not carry over to Arabic where BNs can only be interpreted existentially. The following examples are modeled after those used by

Longobardi.

In characterizing sentences with s-level predicates, Arabic BNs cannot be read generically as in (i a,b).(i) a. *ʔalaat ħadiiθe yimkin tθiir kθ iir min l-fuDuul Generic

tools modern may raise a lot of curiosity‘modern equipment can raise a lot of curiosity’

b. *sayyaraat beiDah yimkin tkallif θakθar Generic

cars white may cost more‘white cars can cost more’

Arabic BNs cannot be generic in episodic sentences with a generalizing adverb with s-level predicates that Longobardi claims to give rise to generic

readings in Italian (although some of these examples in Italian, as pointed out by a reviewer, might be acceptable with a potentially taxonomic

interpretation; see fn. 48 for more details):(ii) *ʔalaat ħadiiθe/ fiyala beiDah dayman/ Ɣaaliban ʔθaarat kθiir min l-fuDuul

tools modern/elephants white always/ often raised a lot of curiosity‘modern tools/ white elephants always/ often raised a lot of curiosity’

Longobardi (p.341)claimsthat ItalianBNsbehavenon-uniformlywith i-levelpredicateswheregeneric readingsarepossiblewith ‘moreeventive’predicates.(iii) Cani da guardia di grosse dimensioni sono piu efficienti / aggressivi.

‘watchdogs of large size are more efficient/ aggressive’As a matter of fact, V. Figueroa (p.c.) informsme that the Spanish equivalents of these examples are unacceptable with a generic reading of the BN as in

(iv a,b).(iv) a. *Estados de gran tamano son peligrosos

states of great size are dangerousb. *Perros guardines de gran tamano son mas eficientes/ agresivos

watchdogs of large size are more efficient/ aggressiveMoreover, Arabic BNs are unacceptable with i-level predicates either and the examples require the definite article (note that definite relatives require

ʔilli ‘that’):(v) a. *(l-) wilayaat *(ʔilli) ħadʒim-hin kbiir xiTraat

(the) states (that) size-3FP big dangerous-3FPb. *(l-) wilayaat *(ʔilli) ħadʒim-hin kbiir mintiʕ ʃaat

(the) states (that) size-3FP big prosperousArabic BNs in object position are also unacceptable with a generic reading, contrary to what Longobardi claims to be the case in Italian.

(vi) a. Una buona legge esclude solo cittadini stranieri dal diritto divoto. Gen/ Ex

‘a fair law only excludes foreign citizens from the right to vote’b. Amo/ Adoro/ Mi piacciono arance di grandi dimensioni. Gen

‘I love/adore/like oranges of large size’In fact the only way to express a generic reading in Arabic is with the definite article:

(vii) l-qanuun l-ʕaadel byistaθni bas l-muwaTiniin l-ʔdʒaneb.....the-law the-just excludes only the-citizens the-foreign....

(viii) ʔana baħibb l-burtqaal l-kbiirI like the-oranges the-big

If Longobardi’s empirical claims do hold, I have nothing to say, concerning this difference seen between Italian, which seems to allow generic or

taxonomic uses of BNs, and Arabic (andGreek, according to a reviewer), which does not allow BNs to be used in generic contexts, beyond, of course, pointing

out that this would be simply a case of parametric variation.

M. Salem / Lingua 120 (2010) 1476–15011486

Longobardi (1994:641–643) suggests that the default existential interpretation of the empty D head and the requirementthat it be lexically governed have to be satisfied as soon as possible, in the spirit of Pesetsky’s (1989) Earliness Principle. Theassignment of the existential interpretation to the DP and the satisfaction of the lexical government requirement on theempty D head take place in the syntax for Italian but not in English. But notice that the Earliness Principle is not respected forItalian in the case of focused BNs since these have to wait to LF to satisfy the requirement that their D head be lexicallygoverned. However, unfocused, postverbal BNs in Italian achieve their licensing at S-Structure. This disjunction in treatingfocused and unfocused BNs is unappealing.

Another problematic aspect of Longobardi’s analysis has to do with the effect modification and coordination are said tohave on licensing a BN (1994 fn. 12).

(63) Meravigliose foreste/Foreste meravigliose si aprivano davanti ai nostri occhi

beautiful forests opened in front of our eyes

(64) Cane e gatto si erano gia addormentati

dog and cat had already fallen asleep

Longobardi speculates that modification provides some feature(s) to the empty head D, allowing it to somehow satisfythe lexical government restriction. As for coordination, he tentatively points out that it could accomplish the same functionthat D would in turning a nominal phrase into an argument. These suggestions, however, remain hypothetical and it is notclear how, or if, they can be implemented.

In view of these problemswith the lexical government approach, the next section presents a different proposal to explainthe licensing of Arabic BNs based on the idea put forward in section 3 that these nominals are in focus.

4.2. Arabic BNs and the semitic construct state

In this section, I propose that focus is crucial to the licensing of Arabic BNs. To achieve this end, I examine a constructionthat has been widely discussed in the Afro-Asiatic literature, the so-called Construct State (CS). BNs in CS constructionsdistribute freely, unlikewhatwe see in the general casewhere Arabic BNs are subject to certain restrictions.29 Understandingwhat features CS BNs possess, and how those features are acquired, a topic I have dealt with elsewhere (Salem, 2008),promises to be quite telling as to why BNs in general are so restricted. In what follows, I give a brief overview of the mainargument in that work.

The main argument in the mentioned work is that BNs in Arabic are unmarked or unspecified for � Definiteness featuresand that those features are best viewed as a property of the DP. For the nominal to be licensed, D has to be saturated, in the sensethat its features checked. One option is to lexically realize an article in D, in which case the nominal is interpreted as definite.30

Alternatively, D can be saturated bymoving a nominal into it, an option arguably available to BNs in CS constructions.31 If neitherof these two options is available, a BN is licensed through focus, the only option I will argue to be available to BNs in non-CSconstructions. The difference between CS BNs and non-CS BNs would therefore reduce to how and why the former could belicensed by moving into D whereas the latter could not. Combining these two assumptions, that nouns are unspecified for �Deffeatures and that D has strong matching features that have to enter in a checking relationship, we are able to see why BNs in CSstructures are licensed but those in non-CS structures are not. The assumption of the existence of strong �Def features of the Dposition can be recast as a non-emptiness requirement of D at LF driven by the necessity to achieve an interpretation (see, forexample, Delfitto and Schroten, 1991 and Zamparelli, 2001). The BN is licensed by achieving the necessary interpretation at LF. Inthe case of CS BNs, the nominal is interpreted as either definite or indefinite, depending on the definiteness value of the genitiveDP/possessor (see below). As for non-CS BNs, the nominal is likewise licensed by achieving an interpretation at LF. For Arabic BNs,the interpretation is only indefinite; English BNs can be interpreted as indefinites or as definite (topics).

Turning our attention to the Semitic CS, we see that this construction is quite productive in Arabic and Hebrew (see,among others, Borer, 1999; Danon, 2001, 2006; Dobrovie-Sorin, 2000; Fassi-Fehri, 1999; Longobardi, 1996; Ritter, 1988,1991; Siloni, 1994, 1996, 1997, and references cited there).32

29 It is important to note that a BN in CS structures is not in the same structural position a BN in non-CS structures is in. However, I continue to use the label

‘‘bare’’ for ease of reference only with this caveat in mind.30 Recall that Arabic does not have an indefinite article.31 See immediately below in the text for references.32 The Construct State has the following well-documented characteristics:

a. The head N is first in the construction which has been argued to be a DP.

b. The head N carries the Case assigned to the whole DP.

c. An obligatory genitive DP argument, or possessor, that carries genitive Case follows N.

d. The head N lacks the (definite) article, whether the head N is interpreted as definite or indefinite.

e. The head N is interpreted as definite or indefinite depending on the (in)definiteness of the genitive argument or possessor, the so-called (In)

definiteness Inheritance.

f. The head N must be adjacent to the genitive argument.

M. Salem / Lingua 120 (2010) 1476–1501 1487

For concreteness, (65)a and b illustrate, respectively, definite and indefinite CS constructions in Arabic.

(65) a. (*l) beet (*l-kbiir) *(l-mara) (l-kbiir)

(the) house-MS (the-big-MS) (the-woman) (the-big-MS)

‘the woman’s big house’

b. beet *(mara) (*l) kbiir

house-MS (woman) (the) big-MS

‘a woman’s big house’

An aspect of CS structures that is usually overlooked is that N is not restricted to singular BNs; plural BNs are also possible.Consider (66).

(66) a. tullaab l-madrase ʃarako b-l-mubaaraa

students-MP the-school participated-3P in-the-match

‘the school’s students participated in the match’

b. ʃufit Tullaab l-madrase b-l-mubaaraa

saw-1S students-MP the-school in-the-match

‘I saw the school’s students at the match’

These plural BNs are acceptable preverbally or postverbally; whatever restrictions BNs are said to be subject to generally,CS BNs seem to be exempt from.

CS structures have been standardly assumed to involve N to Dmovement and that N originates in a position lower than itssurface position. Furthermore, it is assumed that the genitive DPmoves to a position higher than that of attributive adjectivesbut lower than N in D.

I assume, with Longobardi (1996) and Siloni (1994) that the possessor moves to AgrGP. I also assume with these authorsthat the genitive DP moves into AgrGP to check its Case against an AGR head, with which it will be in a spec-head structuralconfiguration. I similarly take the Case features of AGR to be strong and that they trigger overt movement of the genitive DP.The resulting structure for (67) and (68) will be as in (69).

(67) byuut l-walad

houses the-boy

‘the boy’s houses’

(68) byuut walad

houses boy

‘a boy’s houses’

M. Salem / Lingua 120 (2010) 1476–15011488

If N in Arabic lacks the features necessary to check the strong�Def features of D, those features are acquiredwith the help ofan obligatory genitive DP that has to check its structural genitive Case in a spec-head configuration with the AGR head.Consequently, the genitive DP, which is specified for �Def features, raises to AGRP landing in its specifier. AGR acquires therelevant Def features of this DP. Raising of N to D becomes possible once it acquires the necessary �Def features of the genitiveDP.33,34 An immediate advantage of this analysis is that it provides uswith a simple and a straightforward answer as towhy BNs innon-CS structures do not raise to D. These nominals cannot move into D because they lack the necessary �Def and there is nogenitive DP to provide them with these features. D in non-CS BNs has to be licensed differently, through focus.

The same facts appear to hold in English BNs as well. English lacks singular bare nouns in argument positions as seen in(70)–(71).

(70) *(The/A) boy won the prize.

(71) I gave *(the/a) boy a prize.

But if D has strong Def features in English, how can the acceptability of plural BNs be explained, as exemplified in (72)?

(72) a. Boys won the prize. Existential

b. Boys never win any prizes. Generic

I believe the answer to this question lies in the argument reviewed in section 2 that existential BNs in English are focusedand generic BNs are presupposed. But how does being focused or topical license a BN?35,36 In addition, if Arabic/Spanish BNs,and existentially interpreted BNs in English, are focused, why is there a difference in distribution between these languages?Let me begin with the first question.

4.3. How does focusing a BN license it?

The left periphery of the clause, Rizzi (1997:285) points out, has traditionally been said to involve dichotomies such astopic and comment as exemplified in the following topicalization example from English:

33 If the strong �Def features of D were checked by generating a definite article (and the non-emptiness requirement of D at LF to achieve an interpretation is

met), this would preclude the raising of N as the morphological trigger would no longer be available; it is precisely such a circumstance that gives rise to the

observed complementarity between the definite article and the occurrence of N in D in CS structures. If the lexical article is not realized in D, D’s strong �Def

features will still be unchecked, triggering N raising (however, see next footnote). But recall that nouns in Arabic are, by argument, not specified for �Def in the

base position. At this point the genitive DP proves instrumental. N has to adjoin to AGR on its way to D, as a consequence of the HeadMovement Constraint, and it

inherits the Def feature specification of the possessor sitting in the Spec position of AGRP. This is precisely where and how (in)definiteness inheritance in the CS

takes place. (See Longobardi (1996) for a similar, albeit not identical, proposal.)34 A reviewer raises a non-trivial question: if N is not specified for �Def features, and since it is these strong features on D that motivate N-movement and its

subsequent substitution into D, howwould N be visible to D in the first place? I believe the answer lies in the observation that it is true that N lacks�Def features,

but it is still specified for agreement features that need to be checked against an Agr head. This is essentially the catalyst for N movement and its adjunction to

AGR. Once N has acquired the �Def feature specification from AGR and has become visible, it moves to D. This analysis allows us to explain why a genitive DP is

obligatorily present in the CS structure. The reader is referred to (Salem, 2008) for more details.35 An anonymous reviewer asks whether the analysis put forward by Delfitto and Schroten (1991) (D&S) can instead be applied to BNs in Arabic. D&S

propose that the distribution of bare plurals in English and Dutch vis-a-vis Spanish and Italian can be explained in terms of the morphological differences

between these two language groups. The number affix in English, for example, selects an N0, unlike what happens in Spanish and Italian where it selects an

N�1 (a bound morpheme). Raising the number affix to D at LF in the latter languages will result in a Relativized Minimality violation (Rizzi, 1990), because

N�1 will act as a potential governor thereby blocking Num, which is presumably an X�1-level category, in D from governing its trace. Since the number affix

in English-type languages selects an X0 category, noMinimality violations ensue, andNumcan raise to D at LF. The non-emptiness of D at LF is necessary and

‘‘forced by the restricted quantification requirement on natural language’’ and the excorporation of the Num affix at LF and moving it to D is necessary

because it acts as a plural quantifier. This analysis is problematic when Arabic is taken into account. First, Arabic bare plurals behave exactly like their

Spanish counterparts in terms of their semantics and distribution. Num in Arabic should then be expected to select an N�1-level category andmoving Num

to D should therefore be banned. However, remember that in Construct State structures, N is arguablymoved to D at S-Structure resulting in a configuration

different from that assumed by these authors to hold in Spanish and Italian BPs. Assuming with D&S that Num moves to D at LF, Num and N will not be in

distinct syntactic positions at LF, unless we assume a lowering of N from D to its base-generated position. Second, singular bare count nouns are possible in

Arabic, a result predicted to be ungrammatical within D&S’s set of assumptions. Third, to assume, as D&S argue is the case for Spanish and Italian, that Num

in Arabic attaches to an X�1-level category is problematic. Consider the so-called ‘‘sound’’ plurals in Arabic, such as muhandis�muhandisuun/muhandisiin

‘engineer-engineers’. The singular, just like in English, does not require any numbermarking and appears to be a freemorpheme; it can hardly be claimed to

be a bound morpheme. Moreover, and even more problematic, is the case of the so-called ‘broken’ plurals, such as manzil�manazil ‘house-houses’, in the

formation of which internal vowel change (ablaut) is involved. Obviously, a closer look at the plural system in Arabic is warranted, but the observations

made so far seem to suggest that Arabic, partially, aligns itself with English, rather than Spanish, in terms of the plural facts, and with Spanish and Italian, in

terms of the distribution of bare plurals.36 The same reviewer mentioned in the previous footnote points out that D&S’s analysis is able to explain the fact that French disallows BNs in argument

positions. A closer, more detailed, investigation of BNs in French is necessary, and, importantly, a closer look at the two systems of focus and word order in

this language is called for. It has to be said, however, that the present paper is a rough attempt to explain certain idiosyncrasies in the distribution of BNs in a

limited number of languages, where the data are available. In other words, the general purpose of this study is to explain why BNs, when they are actually

attested and allowed by the language, behave the way they do. It is beyond the purview of this study to explain why BNs are not attested in certain

languages, such as French, in the first place; that would have to be a topic for future research.

M. Salem / Lingua 120 (2010) 1476–1501 1489

(73) Your book, you should give t to Paul (not to Bill).

The topic-comment dichotomy in Italian is traditionally expressed by Clitic Left Dislocation (CLLD), where a resumptiveclitic, co-referential with the topic, is contained in the comment as in (74).

(74) II tuo libro, lo ho letto

‘your book I read (,not his)’

The left periphery of the clause is also said to involve the focus-presupposition articulation as seen in (75).

(75) YOUR BOOK you should give t to Paul (not mine).

Rizzi provides the following structures for the left periphery of the clause:

Accordingly, Rizzi assumes the structure in (77) of the complementizer system. Notice that the FocP is flanked by a TopPthat can be reiterated on each side; hence the asterisk.

Rizzi (p. 287) assumes that a constituent possessing a topic or focus feature ends up in a spec-head relationshipwith a Tophead or a Foc head, respectively. Movement of a focused or topicalized constituent can then be subsumed under movementfor feature checking, in the spirit of Chomsky (1995). Rizzi also argues (1997:290) that a clause can havemany topics but onlyone unique focus as seen in (78) and (79).

(78) II libro, a Gianni, domani, glielo daro senz’altro

‘the book, to John, tomorrow, I’ll give it to him for sure’

(79) *A GIANNI IL LIBRO daro (non a Piero, l’articolo)

‘TO JOHN THE BOOK I’ll give, not to Piero, the article’

M. Salem / Lingua 120 (2010) 1476–15011490

Furthermore, Rizzi argues that a focus and one or more topics can co-occur as in (80).

(80) A Gianni, QUESTO, domani, gli dovrete dire

‘to Gianni, THIS, tomorrow, you should tell him’

With the structure in (77) in mind, we can explain the structure of an Arabic SVO sentence in which S is contrastivelyfocused preverbally by assuming that S sits in the specifier of a FocP and is licensed in a spec-head configurationwith the Foc8(more on how the licensing works will be said below). The potentially problematic case is (81), with WLAAD ‘BOYS’ in apostverbal position contrastively focused.

(81) liʕb-o WLAAD f-l-ħadiiqa (miʃ BANAAT)played-3MP BOYS in-the-park (not GIRLS)

‘boys played in the park (not girls)’

The structure in (81) is incompatible with the clause structure assumed by Rizzi where the FocP is in a left peripheralposition. A closer look at the relevant CQ for (81) is enlightening though. (81) is a good answer to (83), not (82), sinceWLAAD

‘BOYS’ in (81) is contrastively focused; (82), however, places narrow, neutral focus on S.

(82) #miin liʕb-o f-l-ħadiiqa?who played-3P in-the-park

(83) liʕb-o BANAAT f-l-ħadiiqa?played-3P GIRLS in-the-park

If (83) is indeed the relevant CQ, liʕb-o ‘played-3P’ in (81) will be part of the presupposition, not the assertion; in otherwords, it will carry a topical feature and can be assumed to sit in a TopP flanking the FocP on the left. WLAAD ‘BOYS’ wouldmove into FocPhrase and F-l-ħadiiqa ‘in-the-park’ in (81) would presumably be in a TopP on the right edge of FocP, again, areasonable assumption in terms of Rizzi’s analysis, since it carries a topical feature. The structure can be represented asfollows (unimportant details omitted).

(85) liʕb-o f-l-ħadiiqa wlaad

played-3MP in-the-park boys

‘boys played in the park’

The C-NSR assigns stress towlaad ‘boys’ and it does not seem descriptively correct to say that the BN sits in a FocP in a leftperipheral position. But recall that the CQ for (85) is (86), which suggests that liʕbo f-l-ħadiiqa ‘played-3MP in-the-park’ in(85) is topical.

(86) miin liʕib f-l-ħadiiqa?who played in-the-park

‘who played in the park?’

M. Salem / Lingua 120 (2010) 1476–1501 1491

Since liʕb-o f-l-ħadiiqa has a topical feature, it must sit in a TopP flanking the left edge of FocP and the structure willcontain two TopPs, one for each topical phrase as in (87).37

We can thus maintain that a focused BN in Arabic moves to a specifier position of a FocP overtly in the syntax.But how does focusing a BN license it? It is a fairly uncontroversial and well-established, much-quoted, fact about BNs in

languages such as Spanish or Italian that they are allowed in preverbal positions if semantically focused and therefore madeprosodically prominent. Arabic BNs behave in a fashion identical to that observed of BNs in these languages. What I amproposing is to generalize this idea to apply to all existentially interpreted BNs in all structural positions.

Recall from our discussion of the Arabic Construct State (CS) that Arabic BNs in CS structures are licensed by landing in D,having obtained the necessary �Def features from the genitive DP. This way, the BN is able to achieve an interpretation at LF,either indefinite or definite. Ideally, focusing a BN in a non-CS construction ought to perform a similar function to that performedby the genitive DP in a CS structure, which is to provide the BN with the necessary features to be interpreted at LF. I believe thatfocusing a BN does just that. It makes it possible for the BN to check its features against Foc8 in a specifier-head configuration so asto achieve an interpretation. Once focused, the BN can move into a specifier position that carries a [+focus] feature. Focus isstandardly taken to encode new information. It can thus be said to encode indefiniteness features (i.e. �Def). If the DP thatcontains the BN sits in a specifier position of a FocP in a checking configuration with Foc8, and since focus is inherently indefinite,indefiniteness features are passed on to the DP containing the BN in the specifier position. It is then predicted that the onlyfeatures that can be inherited by a BN in the specifier of a FocPwould be indefiniteness features. This prediction is exactlywhatwefind in Arabic and Spanish and BNs in these languages can only be interpreted as indefinite. Furthermore, this prediction is alsoconfirmed for English, since the interpretation an existentially read BN has is indefinite.38,39,40

If BNs in Arabic, and existentially interpreted BNs in English, have to be in focus since they are ordinarily unable to moveinto D, this raises the question how, in the case of English generic BNs, which are unfocused by argument, the features of Dare checked.

To begin with, generically interpreted BNs have been argued to be non-asserted, or topical, unlike existentiallyinterpreted BNs. It is reasonable, then, to assume that generic BNs should behave differently from their existential

37 Rizzi rejects the idea that (multiple) topics are adjunctions which explains their reiterability while assuming standard x-bar schema for focus. The

reader is referred to Rizzi (1997) for arguments and details.38 A reviewer explains that it is conceivable for indefinite BNs that focus will impart an indefinite reading and Topic a generic reading, but with pure kind

readings, such as in ‘‘DINOSAURS (and not lions) have become extinct,’’ focus does not imply indefiniteness. But notice that in this example DINOSAURS is a

contrastive topic and not informationally focused (on the difference, see fn. 7). Therefore, it is natural for it not to be indefinite as indefiniteness goes hand in

hand with informational focus.39 But recall that in fn. 21 I stated that Heycock and Zamparelli (2003) analyze instances of coordinated singular bare nouns in English and Italian as having

the semantics of definites. Unmodified bare plurals are argued to have a definite reading, alongside their usual indefinite readings. Note, though, that the

definite readings for bare plurals are only possible with coordination and after certain uniqueness conditions are met. The reader is referred to H&Z (2003)

for details.

On a different note, H&Z (2003:457) also propose what they call the ‘‘Activation Principle’’ which they state as follows:

‘‘An empty functional projection can be licensed, and the semantic content activated, by the presence of appropriate material in its specifier.’’

This activation principle, which H&Z propose to explain the definite semantics of coordinated bare singulars and unmodified bare plurals, can perhaps

be used to help us explain how focus is able to license BNs in Arabic. The movement of the focused BN into the specifier of a Focus Phrase will create the

necessary spec-head configuration with the empty functional head F8. Such movement will presumably activate the focus projection, giving rise to the

interpretation proper for focused material, encoding new information or indefiniteness.40 Rizzi’s analysis and its application to the argument pursued here can be recast in purely minimalist terms, as in Gutierrez-Rexach and Villar (G-R&V)

(1999), by adopting the minimalist hypothesis that a functional category can be associated with multiple specifiers (Multiple Specifiers Hypothesis). G-R&V

explain that according to this hypothesis, a specifier will be created if there is a feature that needs to be checked for convergence. So, if there is a [+Focus]

feature to be checked, a [+Focus] specifier will be projected. Likewise, if there is a [+Topic] feature to be checked, a [+Topic] specifier will project. These

specifiers will be associated with the functional head D. In the case of BNs, G-R&V argue, as I have also done in this paper, the focused BN does notmove into

D (in non-CS structures, in my terms). To be licensed, only the feature [+Focus] of the BN will move and enter into a spec-head agreement with a specifier

with a [+Focus] feature and, since it is an interpretable feature, it does not erase but survives to the semantic component to be interpreted. A similar account

can be hypothesized for checking [+Topic] features. Modified BNs receive a similar treatment. These latter nominals can check their [+Focus] features in the

appropriate specifier. Notice that these nominals do not necessarily have to move, only their [+Focus] feature raises.

M. Salem / Lingua 120 (2010) 1476–15011492

counterparts. Generically interpreted BNs carry a topical feature that needs to be checked in the proper configuration, that is,in the specifier of a TopP. Supposing that these nominals are structurally located in the specifier position of a TopP, the DPs inwhich these nominals are contained enter into a checking relationwith Top8 of that projection. If we allow existential BNs tobe licensed by an empty functional head F8, then for Top8 to license BNs that carry a topical feature should be quite plausible;these nominals will be licensed and also interpreted as definite topics.41,42

In terms of information structure, the current analysis inevitably lumps together bare nominals (plurals andmass nouns)with generic and kind interpretations in English, as well as definite expressionswith the definite article (definite expressionsare presupposed in that they are part of the background information of the speaker and addressee) and nominals with thedefinite article (the-plurals, the-singular mass nouns) with generic or kind interpretations, or as definite expressions, inArabic. These nominals will always have topical features andwill be presupposed. By contrast, bare nominals in Englishwithexistential interpretations (EX indefinites) and bare nominals generally in Arabic (EX indefinites) are likewise lumpedtogether and will always be in focus.

What is the significance of this division of labor? This appears to suggest that a language will first resort to bare nominalsto express a given interpretation. In other words, there could be an economy principle similar to Chierchia’s (1998) ‘‘AvoidStructure’’ operational as a universal principle seeking to achieve an interpretation with the most minimal structurepossible. English, in adherence to this principle, will express both existential and generic interpretations by using BNs. Thisprinciple will in fact prohibit English from achieving the interpretation in question by using the definite article, which is noteconomical. Arabic, on the other hand, will not be barred by this economy principle from utilizing the definite article sinceusing BNs to achieve a generic interpretation is not allowed. But why is English able to express both the generic/kind andexistential interpretations using BNs, but Arabic is unable to do that?

Dayal (2004) explains this variation between languages by proposing a scale of definiteness with languages choosingdifferent cut-off points for lexicalizing the definite article.43 Dayal points out that a language may choose to lexicalize thenominalization operation \ ‘down’ and that the latter does not have to be a covert type shift. If \ is not a covert type shift in agiven language it will be lexicalized as the definite determiner encoding ı. Now, both \ and ı are points on a scale ofdefiniteness, with ı being higher on the scale than \. Variation emerges since a language cannot lexicalize \ beforelexicalizing ı first as the latter is higher on this scale of definiteness than the former. This could be responsible for thevariation seen between English and Arabic in expressing the generic interpretation via bare nominals and plural/massnominals determined by the definite article, respectively.44,45

In the next section, I take up the issue of modified and coordinated BNs. As it turns out, modified BNs are still subject tobeing focused in that, as mentioned in section 3.3 about Spanish, they set up a contrast and they fail all topic-hood tests.46

Conjoined BNs also fail topic-hood tests in Arabic, even though I have pointed out in fn. 21 that the acceptability of coordinate(singular, and to some extent, plural) BNs in English and Italian could be explained in terms of Heycock and Zamparelli’s(2003) analysis as having the semantics of definites. More on the last point will be said later.

41 It could be assumed with Vallduvı (1990) that the information structure of a sentence is checked at an intermediate abstract syntactic level, analogous

to but distinct form LF, called IS (Information Structure). This level mediates the relationship between S-Structure and Informatics (i.e. the information

packaging component), the way the LF level mediates the relationship between S-Structure and the logico-semantic component. Focused and topical

materials would be represented at the abstract IS level and an interpretation is assigned. In some languages such as Catalan, according to Vallduvı, there is

direct correspondence between S-Structure and IS. Other languages, such as English, the relationship is not direct and Move-α is involved. For our present

concerns, it could then be assumed that checking and interpreting focus and topical features could take place at such an abstract level.42 Some of the discussion to follow owes itself to the suggestion of an anonymous reviewer.43 Thanks to a reviewer for bringing this point to my attention.44 Chierchia (1998) explains why BNs in Italian cannot co-occur with individual-level predicates by assuming that individual-level predicates are licensed

by a generic operator Gn. This operator occurs in an Aspect Phrase and is neither lexical nor a focus head (the two being the only licensers of the empty D of

bare nominals in Italian, according to Chierchia). For English, which differs from Italian in not having an empty D with the BN in need of licensing, this

problem does not arise in the first place. Clearly, Chierchia’s assumptions are different than mine as I assume that even English BNs have an empty D that

needs licensing, in the case of generic nominals by a Topic head and by a Focus head in the case of existential nominals.45 In accordance with the principles of the Discourse Representation Theory (DRT), as in Heim (1982), one could hypothesize, even though at the risk of

the argument being circular, that BNs in Arabic, being focused, cannot have a generic interpretation since generics are, by argument, topical. According to

the DRT, indefinites do not have any quantificational force of their own; the quantificational force they have is brought about by other expressions in the

linguistic environment. So, adverbs of quantification (such as always) act as operators that bind two arguments according to the schema in (i).

(i) Q-Adv (Ф,Ψ)

The first argumentФ of the quantifying adverb acts as a restriction on the domain of quantification, hence the term ‘‘restrictive clause.’’ The second term in

the quantification corresponds to what Heim labels ‘‘nuclear scope.’’ Indefinites, Heim argues, act more like variables than quantifiers and have this

variable reading in virtue of their lexical meaning. The mapping from S-Structure to LF is achieved via what Heim calls ‘‘rules of construal,’’ thereby cutting

up the logical form into a tripartite structure. Partee (1991) takes this tripartite structure as her starting point to argue that there is a correlation between

the tripartite structure posited for the logical representation of sentences with indefinite nominals and the topic-focus articulation of the sentence. More

specifically, Partee points out that the restrictive clause of the tripartite structure correlates with the topic or the presupposed part of the sentencewhile the

nuclear scope can be shown to coincide with the focus of the sentence. By argument, BNs in English can either be in focus, if existential, or topics, if generic.

In other words, English BNs can be defocalized and will accordingly be mapped into the restrictor, or focused, and bemapped into the nuclear scope. Arabic

BNs, on the other hand, cannot be marked as topics and will therefore always be mapped into the nuclear scope where they will be caught by existential

closure.46 Suner (1982:254 fn. 9) points out that ‘‘one of the problems which face the researcher [is that] very few people bothered to separate consistently

noncontrastive from contrastive environments.’’ In other words, it is sometimes observed that a sentence with a BN in such and such position is

grammatical without being mindful of its topical versus focal status. I think this has been the tendency when it comes to the grammaticality judgments

w.r.t. the phenomenon of BNs.

M. Salem / Lingua 120 (2010) 1476–1501 1493

4.4. Modification and Arabic BNs

A number of topic-hood tests suggested in Reinhart (1981) can be used to demonstrate that modified and coordinatedBNs in Arabic fail to be interpreted as topics. One such test is ‘‘left-dislocation.’’ For example, Felix in (88) andMatilda in (89)would be topics.47

(88) Felix, it’s been ages since I’ve seen him.

(89) As for Matilda, she can’t stand him.

Another test is topicalization, which is distinguished from left-dislocation in lacking an anaphoric pronoun. In this casethere is a gap, co-indexed with the topicalized argument.

(90) Your second proposal, the board found t unfeasible.

Now take the following example from Arabic:

(91) *muħamiin maʃhuriin ribħ-o l-dʒaaʔizehlawyers famous won-3MP the-prize

‘famous lawyers won the prize’

Under normal conditions of stress, I want to argue, (91) is ungrammatical. Testing the topic-hood ofmuħamiin maʃhuriin‘famous lawyers’ confirms this conclusion. Consider first how the BN may not be part of the presupposition of a CQ:

(92) *ʃuu ribħ-o muħamiin maʃhuriin?what won-3MP lawyers famous

Second, muħamiin maʃhuriin may not be left-dislocated as in (93) or (94).

(93) *bilnisbe la muħamiin maʃhuriin, homme ribħ-o l-dʒaaʔizehas for lawyers famous they won-3MP the-prize

(94) *muħamiin maʃhuriin, homme ribħ-o l-dʒaaʔizehlawyers famous they won-3MP the-prize

Topicalizing the BN is not possible either as (95) attests.

(95) *muħamiin maʃhuriin, ma-ʃuft-iʃ t fi-l-maħkame

lawyers famous neg-saw-1MS-neg in-the-courthouse

‘famous lawyers, I didn’t see in the courthouse’

Modification with a non-intersective adjective does not rescue the BN either under normal conditions of stress48:

(96) *muħamiin saabeqiin ribħ-o l-dʒʔizehlawyers former won-3MP the-prize

‘former lawyers won the prize’

47 (88)–(90) are taken from Reinhart (1981).48 The same can be said about modification by ethnic and cardinal adjectives or modification by gerundive reduced relatives. The modified BNs are read

under normal conditions of stress.(i) *θalaθ muħamiin (ʔurdiniyiin) ribħ-o l-dʒaaʔizeh

three lawyers (Jordanian) won-3MP the-prize‘three (Jordanian) lawyers won the prize’

(ii) *wlaad gaʕdiin fi l-gahwe xisr-o kull maSarii-homboys sitting in the-coffee shop lost-3MP all money-3MP‘boys sitting in the coffee shop lost all their money’

Applying any of the usual topic-hood tests would yield ungrammatical results as expected.(iii) a. *ʃuu xisr-o wlaad gadʕiin fi l-gahwe?

what lost-3MP boys sitting in the-coffee shopb. *bilnisbe la wlaad gaʕdiin fi l-gahwe, Omar ħaka maʕhom

as for boys sitting in the-coffee shop, O. talked to themc. *wlaad gaʕdiin fi l-gahwe, (homme) xisr-o kull maSarii-hom

boys sitting in the coffee shop (they) lost-3MP all money-3MPd. *wlaad gaʕdiin fi l-gahwe, Omar ma-laaħað-iʃ t

boys sitting in the-coffee shop O. neg-noticed-3MS-neg‘boys sitting in the coffee shop, O. did not notice’

M. Salem / Lingua 120 (2010) 1476–15011494

The same topic-hood tests can be applied to (96) in (97)–(99).

(97) *ʃuu ribħ-o muħamiin saabeqiin?

what won-3MP lawyers former

(98) *bilnisbe la muħamiin saabeqiin, homme ribħ-o l-dʒaaʔizehas for lawyers former they won-3MP the-prize

(99) *muħamiin saabeqiin, homme ribħ-o l-dʒaaʔizehlawyers former they won-3MP the-prize

Topicalizing the BN is not possible either as (100) clearly illustrates.

(100) *muħamiin saabeqiin, ma-ʃuft-iʃ t fi-l-maħkame

lawyers former neg-saw-1MS-neg in-the-courthouse

‘famous lawyers, I did not see in the courthouse’

We can then conclude that modified BNs do not count as topics. I believe that this is a welcome result for two reasons.First, this will guarantee us a greater degree of uniformity in dealing with the distribution of BNs. Focusing a BN wouldtherefore be an across-the-board requirement BNs are subject to. Second, the consensus in the literature (for example,Carlson, 1977a,b; Longobardi, 1994) takes BNs to be those nominals that are determinerless (that is, not determined by, forexample, an article or quantifier). BNs with adjectival premodification are still considered to meet the ‘bareness’ criterion. Itwould be surprising for the same not to carry over to postmodification. Thatmodified BNs are allowed in preverbal positionssuggests that they domeet the focus requirement. I findmyself in agreementwith Suner who suggests thatmodified BNs arestill foci in an important way since they trigger a contrast. So, for example, in (96),muħamiin saabeqiin ‘former lawyers’ doesnot refer to the entirety of lawyers, but isolates a certain class of lawyers, i.e. former lawyers, not the current ones. Thiscontrast provides the BN with the necessary prominence to satisfy the focus requirement.49

Modification by reduced relatives or indefinite relatives cannot license BNs under normal conditions of stress.(iv) a. *ʕabiid maʕrufiin b-guwwit-hom ban-o l-ahraamaat

slaves known for-strength-3P built-3P the-pyramids‘slaves known for their strength built the pyramids’

b. *ʃuu ban-o ʕabiid maʕrufiin b-guwwit-hom?what built-3P slaves known for-strength-3P

c. *bilnisbe la ʕabiid maʕrufiin b-guwwit-hom, (homme) ban-o l-ahraamaatas for slaves known for-strength-3P (they) built-3P the-pyramids

(v) a. *ʕulamaaʔ ribħo l-dʒaaʔizeh ħiDr-o l-ʔidʒtimaaʕscientists won3-MP the-prize attended the-meeting

b. *ʃuu ħiDr-o ʕulamaaʔ ribħ-o l-dʒaaʔizeh?what attended-3MP scientists won-3MP the-prize

c. *bilnisbe la ʕulamaaʔ ribħ-o l-dʒaaʔizeh,as for scientists won-3MP the-prize(homme) ħiDr-o l-ʔidʒtimaaʕ(they) attended-3MP the-meeting

A marginal reading of a contrastive nature is available in (iv)a with a pair-list reading where we have a juxtaposition between what slaves who are

known for their strength built contrasted with what slaves who are known for their weakness built, for example.49 A reviewer provides the following Italian example to suggest that existential BNs can be topics:(i) Minerali ħad alto contenuto di uranio sono stati trovati solo in questa miniera

Minerals with a high uranium content have been found only in this mineJudging from the English translation, it seems to me that the reviewer’s claim that the BN is interpreted existentially is dubious. In fact, if any of the topic-

hood tests is used with the BN in (i), as in (ii), the natural interpretation is for the BN to be read generically.(ii) As for minerals with high uranium content, they have been found only in this mine.Therefore, even though it is true that themodified BN is used as a topic in (i), it is not used existentially, notwithstanding the s-level predicate it is being used

with. The BN is used as a contrastive topic, and has a taxonomic interpretation tomean that instances ofminerals with high uranium content have only been

found in a certain mine. A similar taxonomic reading can be found in (iii) on a par with (i).(iii) Dogs with two heads have been documented only in one town in the country.Obviously, these nominals do not constitute a kind, if a kind is judged in terms of frequency or regularity, but they are generic in that they create a taxonomy.

The same reviewer also notes that although the BN in (iv) is interpreted generically, it is nevertheless focused.(iv) Sono pericolosi soprattutto cani da guardia di grandi dimensioni

Are dangerous especially watchdogs of large sizeI am informed by R. DeBenedictis (p.c.) that the predicate pericolosi soprattutto, and not the BN, in (iv) is focused, just like in the English exampleWatchdogs

are particularly dangerous, whether (iv) is interpreted as a question or a statement. I am also informed that the BN here, just as I claimed to be the case for

Minerali ad alto contenuto di uranio in (i) above, is interpreted generically. It is clear, however, that what we have here is a taxonomic reading which can be

used generically, as pointed out in fn. 28. Clearly, this particular usage of BNs in Italian has to be investigated for a better understanding. What I can say at

this point, as also indicated in fn. 28, is that these examples are not possible in Arabic with a generic reading.

This reviewer also raises the questionwhy examples such as (v) are ungrammatical, where a BN appears in a post-copular, supposedly focused, position

unless it is modified.(v) I colpevoli dei disordini sono ribelli *(che non hanno accettato la tregua)

the culprits of the riots are rebels (who did not accept the truce)I do not have an immediate answer to this observation, but what I can say is that the equivalent in Arabic is perfect as in (vi), with or without modification.(vi) l-muttahamiin fii l-muð aharaat kan-o mutamarridiin (ma wafag-oʕala l-hodneh)

the-accused in the-riots were rebels (neg agree to the-truce)

M. Salem / Lingua 120 (2010) 1476–1501 1495

4.5. Coordination and BNs

Coordinated Arabic BNs also fail the topic-hood tests. Consider (101).

(101) *zlaam w niswaan ħiDr-o l-ʔidʒtimaaʕmen and women attended-3P the-meeting

The conjoined BNs may not appear part of the presupposition of a question:

(102) *ʃuu ħiDr-o zlaam w niswaan?

what attended men and women

Coordinated BNs also fail the ‘‘as for’’ test as in (103).

(103) *bilnisbe la zlaam w niswaan, homme ħiDr-o l-ʔidʒtimaaʕas for men and women they attended-3P the-meeting

Coordinated BNs do not pass the left-dislocation or the topicalization tests:

(104) *zlaam w niswaan, ʃuft-hom fi l-ʔidʒtimaaʕmen and women saw-1MS-3P in the-meeting

‘men and women, I saw them at the meeting’

(105) *zlaam w niswaan, ʃufit t fi l-ʔidʒtimaaʕmen and women saw-1MS in-the-meeting

‘men and women, I saw in the meeting’

Remember that these coordinated BNs are to be read unfocused. If these nominals are focused, contrastively as in (106)–(107) or not as in (108), the result is grammatical.

(106) ZLAAM W NISWAAN ʃufit fi l-ʔidʒtimaaʕ, miʃ WLAAD W BANAAT

MEN AND WOMEN saw-1MS at-the-meeting, not BOYS AND GIRLS

‘MEN AND WOMEN I saw at the meeting, not BOYS AND GIRLS’

(107) ʃufit ZLAAM W NISWAAN fi l-ʔidʒtimaaʕ, miʃ WLAAD W BANAAT

saw-1MS MEN AND WOMEN at-the-meeting, not BOYS AND GIRLS

‘MEN AND WOMEN I saw at the meeting, not BOYS AND GIRLS’

(108) ʃufit fi l-ʔidʒtimaaʕ zlaam w niswaan

saw-1MS at-the-meeting men and women

‘I have seen men and women at the meeting’

My claim is that coordinate BNs in Arabic are still subject to a focus requirement, on a par with non-conjoined BNs. This isobviously a conclusion that is different from the one I have speculated about in fn. 21 concerning coordinate singular and pluralBNs in English and Italian. The latter nominals are argued to have the semantics of definites byHeycock and Zamparelli (2003).So, why should we find such parametric difference between Arabic and English/Italian in this respect? It seems to me that inattempting an explanation of this parametric difference, another difference, which has to dowith the indefinite article and theavailability of bare singular nouns in these two sets of languages, becomes relevant. English and Italian have indefinite articles,unlike Arabic. Bare singulars in Arabic are therefore possible, even though they are subject to certain restrictions as argued forin the present study, but not in English/Italian. Again, this could be a consequence of an economy principle such as Chierchia’s‘‘Avoid Structure’’: Arabic does not have an indefinite article and singular BNs are allowed. By contrast, English and Italian haveindefinite articles and bare singulars are not generally allowed. Since Arabic allows bare arguments generally, it allowscoordinate BNs with indefinite interpretations, with the proviso that they be focused. English and Italian do not allow baresingular arguments, which is reasonable considering that they have singular indefinite articles.When these latter languages doallowbare singular arguments in coordinate structures, they assign to these nominals the semantics of definites. These remarkshave to remain speculative at this stage and a more detailed investigation is necessary, which I leave for future research.

To sum up, I have argued that BNs in Arabic are licensed through focus, on a par with their existential counterparts inEnglish, a point of convergence between these two languages. What is yet to be determined, however, is why these two

M. Salem / Lingua 120 (2010) 1476–15011496

languages diverge when it comes to the distribution of BNs. In the next section, I argue that the divergence is a consequenceof the differences between English and Arabic in terms of the focus system facts and the options made available in terms ofword order.

5. Focus, word order, and the distribution of BNs

I have so far maintained that BNs in Arabic are licensed through focus. However, these BNs behave quite differently whenit comes to their distribution as against English BNs. BNs in English are unrestricted in their distribution and are not confinedto a certain structural position to be licensed. Arabic BNs, I have argued, are acceptable if they are made prosodicallyprominent in the lowest c-command position by the C-NSR or, if they are sentence-internal, by a contrastive stress rule.These distributional differences, I would like to argue, are merely surface manifestations of much deeper differences in theproperties of focus and word order between English, on one hand, and Arabic, on the other hand.

We have seen above how the C-NSR assigns stress in Arabic to the constituent lowest in terms of c-command.Alternatively, another analysis can be readily applied to Arabic and English, namely Cinque’s (1993) proposal regardingnuclear stress. Cinque’s (1993) proposal (as reviewed in Reinhart, 1995) basically revives Chomsky’s (1971) analysis of focusas determined solely with reference to the PF structure itself without reference to discourse (Reinhart, 1995:20). Sentencestress assignment, as defined on PF structures, and as assigned by the NSR, does not need to be parameterized in order tocapture the varied stress assignment patterns across languages. Rather, the NSR can be assumed to apply cyclically tosyntactic constituents assigning sentence stress to the most embedded element in the sentence. The most embeddedconstituent would be determined by taking the direction of recursion (that is, syntactic branching) in the language inquestion into consideration. Neutral focus, for Cinque, is assigned in the sentence according to the sentential stressassignment on the PF structure in accordance with the following focus rule:

(109) The Focus Rule

The focus of IP is a(ny) constituent containing the main stress of IP, as determined by the stress rule.

The Focus Rulewould predict that in a sentence such as (110) any of the three constituents bearing sentential stress couldserve as focus.

(110) [My neighbor [is building [a desk]]]

* * *

a) NP cycle: [* ]

b) VP cycle: [ * ]

c) IP cycle: [ * ]

Any of the three constituents, NP, VP, or IP, can serve as focus, as Reinhart’s following example illustrates (stress isindicated by bolding, focus by underlining):

(111) a. What’s this noise?

My neighbor is building a desk.

b. What’s your neighbor doing?

My neighbor is building a desk.

c. What’s your neighbor building?

My neighbor is building a desk.

d. Has your neighbor bought a desk already?

#My neighbor is building a desk.

e. Who is building a desk?

#My neighbor is building a desk.

Notice that the subjectmy neighbor and the verb alone, building, are not possible neutral foci according to the Focus Rule.The subject and the verb can be assigned focus by another rule, namely the Marked Focus Rule.

(112) Marked Focus Rule

Relocate the main stress on a constituent you want to focus.

M. Salem / Lingua 120 (2010) 1476–1501 1497

The Marked Focus Rule, when applied, generates (113) and (114), for (111d) and (111e), respectively.

(113) My neighbor is building a desk.

(114) My neighbor is building a desk.

The examples in (111a–c) can be reproduced in Arabic:

(115) a. ʃuu haaði d-dadʒdʒe?what this the-noise

-b-yibni dʒaari b-Tawla 50

is-building neighbor-1S (a)-table

‘my neighbor is building a table’

b. ʃuu b-yiʕmal dʒaarak?what is-doing neighbor-2MS

‘what is your neighbor doing?’

-dʒaari b-yibni b-Tawla

neighbor-1MS is-building (a)-table

‘my neighbor is building a table’

c. ʃuu b-yibni dʒaarak?what is-building neighbor-2MS

what is your neighbor building?’

-dʒaari b-yibni b-Tawla

neighbor-1MS is-building (a)-table

‘my neighbor is building a table’

(115a) represents the option where the whole sentence is in focus. In this case, the NS is assigned to the most embeddedconstituent, Tawla ‘a table’, although the whole sentence is in focus. In (115b), only the VP is focused and the NS is stillassigned to themost embeddedword in the sentence, namely, Tawla ‘a table’. Finally, when only theNP (DP), Tawla ‘a table’ isfocused, it also gets the NS, as in (115c). So far, English and Arabic behave similarly.We have noted how English, in examples(113) and (114), resorts to the so-called ‘‘Marked Focus Rule’’ to shift the stress onto the verb or the subject in the sentence.Consider Arabic:

(116) a. ʔiʃtara dʒaarak Tawla?

bought-3MS neighbor-2MS (a)-table

‘has your neighbor bought a table?’

*/#dʒaarai b-yibni b-Tawla

neighbor-2MS is-building-3MS of-table

‘my neighbor is building a table’

b. miin b-yibni b-Tawla?

who is-building of-table

‘who is building a table?’

*/#dʒaarai b-yibni b-Tawla

neighbor-3MS is-building-3MS of-table

‘my neighbor is building a table’

Notice how the acceptability judgments between the English examples in (111d,e) and the Arabic examples in (116a,b)coincide. However, these two languages go about resolving the unacceptability of these sentences in completely differentways and English and Arabic implement Cinque’s marked focus rule differently. English destresses a desk, shifting the stressto the focused element, building and my neighbor, respectively, and crucially, in situ. Arabic destresses the non-focalconstituent by p-moving it to a position outside of the scope of the C-NSR, and the focused constituent is placed in the most

50 In a so-called out-of-the-blue answer, where the whole sentence is in focus, the more preferred word order is VS(O), rather than SVO, as in (115)a.

M. Salem / Lingua 120 (2010) 1476–15011498

deeply embedded position as in (117b) (the Arabic counterpart to the English (114)). If this proves to be unavailable, andp-movement does not turn out to be an option as in (117a) (the Arabic counterpart of the English (113)), the focused elementin Arabic would be read contrastively (since it is focused internal to the sentence), unlike what happens in English.51

(117) a. dʒaarai B-YIBNI b-l-Tawla

neighbor-2MS IS-BUILDING-3MS the-table

‘my neighbor is building the table’

b. b-yibni b-l-Tawla dʒaaraiis-building-3MS the-table neighbor-3MS

‘my neighbor is building the table’

English, therefore, does not have the option of using p-movement, an option quite available to Arabic, and Spanish. Arabiccan avail itself of using the marked focus option; in fact, this language cannot resort to such an option since a ‘‘less marked’’option is available via p-movement.52

This fundamental difference between English and Arabic in terms of the availability of the marked focus option asopposed tomovement is in fact quite expected. These two languages differ in another important respect, which is the optionsmade available to them in terms of word order possibilities (cf. Reinhart, 1995). Word order in English is generally rigid, notas flexible as in Arabic or Spanish. English word order is more or less a fixed SVO. Arabic, by contrast, allows SVO, VSO, VOS,among other permutations.53 It should come as no surprise, then, that English would allow in situ focus (i.e. marked focus)since this is the only option that is available to this language. P-movement, as has been just noted, is generally unavailable toEnglish. If English had a freer word order, on a parwith Arabic, then the option ofmarked focuswould be quite unacceptable.By the same token, using the more marked focus option, i.e. marked focus, in Arabic is unacceptable. Using marked focus inthis language comes with a cost, which is the contrastive reading a BNwill have if such an option is used. On the other hand,using the marked focus rule in English is not a costly option and this language can stress constituents non-contrastively insentence-internal positions. The same contrast can be seen between the English examples in (118) and the Italian ones in(119) (the examples are cited in Reinhart (1995:32) from Cinque (1993)) (bolding indicates stress).

(118) a. Johnson died.

b. Johnson died.

(119) a. Johnson e morto.

b. E morto Johnson.

c. #Johnson e morto.

But how does this discussion apply to BNs? Consider (120) and (121) from English and Arabic, respectively.

(120) a. Boys played in the park.

b. Boys played in the park (not girls).

(121) a. liʕb-o f-l-ħadiiqa wlaad

played-3MP in-the-park boys

‘boys played in the park’

b. *wlaad liʕb-o f-l-ħadiiqaboys played-3MP in-the-park

c. WLAAD liʕb-o f-l-ħadiiqa (miʃ BANAAT)BOYS played-3MP in-the-park (not GIRLS)

d. liʕb-o WLAAD f-l-ħadiiqa (miʃ BANAAT)played-3MP BOYS in-the-park (not GIRLS)

51 P-movement is unavailable in (117a) since the outcome of stranding the verb in the right periphery is very marginal as in (i).(i) *? dʒaarai b-l-Tawla b-yibni

neighbor-2MS the-table is-building-3MS52 Reinhart (1995:32) makes similar observations in her exposition of Cinque’s (1993) analysis of the differences between Italian and English. She notes

that focusing in situ, as is the case in English, is amore marked choice (hence, marked focus) thanmoving elements around to express neutral focus as is the

case in Italian. The notion of markedness, and the related notion of economy, is a thorny issue since movement, especially in terms of Chomsky’s (1995)

minimalist assumptions, is always the less economical, hence more marked, option.53 It must be mentioned that dialectal Arabic is not as rich as Standard Arabic in terms of word order possibilities due to the absence of Case endings.

M. Salem / Lingua 120 (2010) 1476–1501 1499

Boys, in (120), is, by argument, in focus, due its existential interpretation. This is basically the only word order optionavailable to English, other permutations being unacceptable. Crucially, the focus on Boys does not have to be contrastive(cf. (120b)). The Arabic word-for-word equivalent of (120), (121b), is ungrammatical because the BN wlaad ‘boys’, which isread existentially typical of all Arabic BNs, is not focused. As p-movement is the only option available to Arabic to achieveneutral focus, forwlaad ‘boys’ to be stressed, it has to be lowest in c-command, or in themost embedded position in Cinque’sterms. This would get us (121a), which is the acceptable equivalent of the English (120a), where, just like boys,wlaad is non-contrastively focused in that position. Crucially, forwlaad to occur in any other position, i.e. outside the scope of the C-NSR, itwould only get the contrastive reading as in (121c,d).54

6. Conclusion

I have argued that BNs in Arabic are similar in importantways to BNs in English and that the surface differences these twolanguages exhibit in the distribution of these nominals are expected since these languages diverge in their focus and wordorder systems. On the one hand, English has a flexible focus system, while it is tied down by rigid word order possibilities.Word order in Arabic is freer; however, the focus system in this language is quite constrained. In fact, one is temptedhypothesize that, perhaps, because English has such an impoverished or restricted word order it compensates for that by arich focus system (or, perhaps, vice versa). We can additionally and equally hypothesize that because Arabic has animpoverished or restricted focus system, it makes up for it by permitting different word order possibilities (or, vice versa).

References

Abney, S., 1987. The English noun phrase in its sentential aspect. Doctoral Dissertation. MIT Press, Cambridge, MA.Borer, H., 1999. Deconstructing the construct. In: Johnson, K., Roberts, I. (Eds.), Beyond Principles and Parameters. Kluwer Publications, Dordrecht.Carlson, G., 1977a. A united analysis of the English bare plural. Linguistics and Philosophy 1, 413–457.Carlson, G., 1977b. Reference to kinds in English. Doctoral Dissertation. University of Massachusetts, Amherst.Casielles-Suarez, E., 1997. Topic, focus, and bare nominals in Spanish. Doctoral Dissertation. University of Massachusetts, Amherst.Chierchia, G., 1998. Reference to kinds across languages. Natural Language Semantics 6, 339–405.Chomsky, N., 1971. Deep structure, surface structure and semantic interpretation. In: Steinberg, D., Jakobovits, L. (Eds.), Semantics: An Interdisciplinary

Reader in Philosophy, Linguistics and Psychology. Cambridge University Press, Cambridge.Chomsky, N., 1976. Conditions on rules of grammar. Linguistic Analysis 2, 303–351.Chomsky, N., 1995. The Minimalist Program. MIT Press, Cambridge, MA.Cinque, G., 1993. A null theory of phrase and compound stress. Linguistic Inquiry 24 (2), 239–298.Cohen, A., Erteschik-Shir, N., 2002. Topic, Focus, and the Interpretation of Bare Plurals. Natural Language Semantics, 10. Kluwer, The Netherlands, pp. 125–

165.Contreras, H., 1986. Spanish bare NPs and the ECP. In: Bordelois, I., Contreras, H., Zagona, K. (Eds.), Generative Studies in Spanish Syntax. Foris, Dordrecht,

pp. 25–49.Danon, G., 2001. Syntactic definiteness in the grammar of Modern Hebrew. Linguistics 39 (6), 1071–1116.Danon, G., 2006. Caseless nominals and the projection of DP. Natural Language & Linguistic Theory 24 (4), 977–1008.Dayal, V., 2004. Number marking and (in)definiteness in kind terms. Linguistics and Philosophy 27, 393–450.Delfitto, D., Schroten, J., 1991. Bare Plurals and the Number Affix in DP. Probus, 32. Walter De Gruyter, Berlin, pp. 155–185.Diesing, M., 1992. Indefinites. MIT Press, Cambridge, MA.Dobrovie-Sorin, C., 2000. (In)definiteness spread: from Romanian genitives to Hebrew construct state nominals. In: Motapanyani, V. (Ed.), Comparative

Studies in Romance Syntax. Elsevier, Oxford.Dobrovie-Sorin, C., 2009. Existential bare plurals: from properties back to entities. Lingua 119 (2), 296–313.Fassi-Fehri, A., 1999. Arabic modifying adjectives and DP structures. Studia Linguistics 53 (2), 105–154.Gutierrez-Rexach, J., Villar, L.S., 1999. Spanish Bare Plurals, Multiple Specifiers and the Derivation of Focus-Related Features XXXIII/3-4, pp. 355–387.

54 A reviewer asks what implications the argument that BNs in Arabic are licensed through focus has on the debate started in Carlson (1977a) concerning

the differences between BNs and indefinites. Clearly, an adequate answer to this question would go beyond the modest purpose of the current paper.

However, it is instructive to remember that the differences between bare plurals and indefinites first recognized by Carlson (1977a), that bare nominals take

narrow scope w.r.t. intentional operators, negation, temporal adverbials, and anaphora, have been claimed by some authors, such as Zamparelli (2001), as

resulting from a lack of referentiality, and not because bare plurals are names of kinds. Zamparelli (2001:14) suggests that perhaps referentiality is linked to

the presence of some lexical material in D at LF. Consider (i).(i) a. John is looking for a piece of paper.

b. John is looking for pieces of paper.In (ib) the bare plural has narrow scope since there is no material in D at LF, unlike what we see in (ia) where the indefinite could have wide scope

because the indefinite article could move into D at LF. Empirical evidence can be brought forth from Arabic in support of this analysis. Bare singulars are

allowed in Arabic, as noted in the text above; the prediction should then be that these nominals should behave on a par with their plural counterpart in

having narrow scope because of the unavailability of an article that canmove into D at LF. I believe that this prediction is confirmed sincemumaθθ il ‘actor’

in (ii), for example, has only narrow scope with respect to negation. (ii) does not mean that ‘‘there is an actor such that Omar did not meet that actor’’ with

the existential quantifier having scope over negation.(ii) Omar ma ʃafiʃ mumaθθil

O. neg saw actor(iii) Omar bifattiʃ/ bidawwer ʕala ʃaggah

O. looking for apartment‘Omar is looking for an apartment’

(iv) Omar bidd-o ygabil mumaθθilO. want to-meet actor‘Omar wants to meet an actor’

(v) kull l-wlaad gar-o ktaab ʕan eldaynasuraatall the-boys read book about dinosaurs‘all the boys read a book about dinosaurs’

M. Salem / Lingua 120 (2010) 1476–15011500

Heim, I., 1982. The semantics of definite and indefinite noun phrases. Doctoral Dissertation. University of Massachusetts, Amherst.Heycock, C., Zamparelli, R., 2003. Coordinated bare definites. Linguistic Inquiry, vol. 34, no. 3. MIT Press, Cambridge, MA, pp. 443–469.Horvath, J., 2006. Separating ‘‘Focusmovement’’ from focus. In: Karimi, S. (Ed.), Phrasal and Clausal Architecture: Syntactic Derivations and Interpretation in

Honor of Joseph E. Emonds. John Benjamins Publishing Company, Amsterdam, NLD.Jackendoff, R., 1972. Semantic Interpretation in Generative Grammar. MIT Press, Cambridge, MA.Kayne, R.S., 1994. The Antisymmetry of Syntax. MIT Press, Cambridge, MA.Kiss, K.E., 1998. Identificational focus versus information focus. Language 74 (2), 245–273.Kratzer, A., 1995. In: Carlson, G., Pelletier, F. (Eds.), Stage-Level and Individual-Level Predicates. The Generic Book.Krifka, M., Pelletier, F., Carlson, G., ter Meulen, A., Link, G., Chierchia, G., 1995. Genericity: an introduction. In: Carlson, G., Pelletier, F. (Eds.), The Generic

Book.Laca, B., 1990. Generic objects: some more pieces of the puzzle. Lingua 81, 25–46.Longobardi, G., 1994. Reference and proper names: a theory of n-movement in syntax and logical form. Linguistic Inquiry 25 (4), 609–665.Longobardi, G., 1996. The syntax of N-Raising: a minimalist theory. OTS Working Papers. Utrecht University, Utrecht.Longobardi, G., 2000. Postverbal subjects and the mapping hypothesis. Linguistic Inquiry 691–702.Longobardi, G., 2001. How comparative is semantics? A unified parametric theory of bare nouns and proper names. Natural Language Semantics, 9. Kluwer,

The Netherlands, pp. 335–369.Partee, B.H., 1991. Topic, Focus and Quantification. In: Proceedings of SALT 1, Cornell University, Ithaca, NY, pp. 257–280.Pesetsky, D., 1989. The Earliness Principle. Unpublished paper presented at GLOW.Reinhart, T., 1981. Pragmatics and linguistics: an analysis of sentence topics. Philosophica 27, 53–94.Reinhart, T., 1995. Interface strategies. OTS Working Papers TL-95-002, Utrecht University.Ritter, E., 1988. A head-movement approach to construct-state noun phrase. Linguistics 26, 909–929.Ritter, E., 1991. Two functional categories in noun phrases: evidence from modern Hebrew. In: Rothstein, S. (Ed.), Syntax and Semantics, 25. Academic

Press, New York, pp. 37–62.Rizzi, L., 1990. Relativized Minimality. MIT Press, Cambridge, MA.Rizzi, L., 1997. The fine structure of the left periphery. In: Haegeman, L. (Ed.), Elements of Grammar. Kluwer, The Netherlands.Rooth, M., 1992. A Theory of Focus Interpretation. Natural Language Semantics, 1. Kluwer, The Netherlands, pp. 75–116.Salem, M., 2008. The Construct State Revisited. MS. Wayne State University.Schmitt, C., Munn, A., 2003. The syntax and semantics of bare arguments in Brazilian Portuguese. In: Pica, Pierre (Ed.), Linguistic Variation Yearbook.Siloni, T., 1994. Noun phrases and nominalization. Doctoral Dissertation. U. de Geneve.Siloni, T., 1996. Hebrew noun phrases: generalized noun raising. In: Belletti, A., Rizzi, L. (Eds.), Parameters and Functional Heads. Oxford University Press.Siloni, T., 1997. Noun Phrases and Nominalization: The Syntax of DPs. Kluwer, Dordrecht.Suner, M., 1982. Syntax and Semantics of Spanish Presentational Sentence-Types. Georgetown University Press, Washington, DC.Vallduvı, E., 1990. The informational component. Doctoral Dissertation. University of Pennsylvania.Zamparelli, R., 2001. Definite and bare kind-denoting noun phrases. In: Frank, D., et al. (Eds.), Proceedings of Going Romance 2000. John Benjamins

Publishers, Amsterdam, Philadelphia.Zubizarreta, M.L., 1998. Prosody, Focus, and Word Order. MIT Press, Cambridge, Mass.

M. Salem / Lingua 120 (2010) 1476–1501 1501