ARABIC NEGATION MARKER (LAYSA) WITH BARE ARGUMENT ELLIPSIS


AND ITS ASSOCIATION WITH INFORMATION STRUCTURE

NASSER AL-HORAIS

Abstract

Within the framework of Chomsky’s Minimalist Program (1995, 2006, 2008), this
paper aims to assemble some evidence to provide additional support for the PF deletion
approach to ellipsis. To this end, data from the Arabic negation marker laysa with Bare
Argument Ellipsis constructions, a topic which is still largely terra incognita, are
investigated. The paper also aims at contributing to the ongoing debate regarding how
the remnant phrase in Bare Argument Ellipsis is derived before deletion takes place. In
this connection, it has been argued for quite extensively that the remnant phrase escapes
deletion by having moved out the elided TP into a Focus position (Depiante 2000). The
current paper, after proving the movement fact, takes this analysis as a starting point and
provides a minimalist analysis of how this movement derives by assuming the following
minimal steps. (i) A moved remnant moves overtly to the spec of TopP or FocP in order
to satisfy the Edge-feature. (ii) NegP with negative Bare Argument Ellipsis is excluded
since the negative marker with Bare Argument Ellipsis is not a head, and therefore it
must occupy the spec of TopP or FocP, waiting for a moved remnant to be merged with.
(iii) Non-negative Bare Argument Ellipsis has the same movement route, apart from
that a moved remnant is only contrastively focused and hence occupies the spec of
FocP.

1. Introduction

Ellipsis has always been one of the most challenging puzzles for linguistic
theory. “It remains difficult to classify, as it appears to involve phonology (due to its
similarity to deaccenting), syntax (by virtue of its distribution), semantics (evidenced by
its apparent licensing conditions), and pragmatics (because of the cognitive load it
imposes)” (Smith 2001: 176). Moreover, Elliptical expressions come in various forms
e.g. VP ellipsis (1), bare argument ellipsis (2), NP ellipsis (3), sluicing (4),gapping (5),
pseudostripping (6) and pseudogapping (7).

(1) They play the piano but Anna doesn't_.

(2) Peter wurde eingeschult und Anna _ auch.
Peter was sent-to-school and Anna too (German, from Winkler 2005:159).


This paper is partly based on a chapter of my Ph.D thesis. I am very grateful to my supervisor Prof.
Anders Holmberg for insightful comments, suggestions and discussions. I wish to thank an anonymous
reviewer for clarifying my ideas and for the many significant editorial improvements which I tried to put
to use here. Also I would like to thank Dr. Jason Merchant for explaining me some points about his work
on Ellipsis. Special thanks are forwarded to Ahmad Mahfouz and Khaled Kakhia for helpful discussions
about some Arabic data and to the editors of this volume, especially to Nadeem and Miri for
encouragement and patience during the writing of this paper. I am solely responsible for any mistakes or
inadequacy.

1
Al-Horais

(3) Manny played a solo with one hand and Anna with two_.
(4) Jack bought something, but I don’t know what_. (Merchant 2004:664).
(5) Ehrlichman duped Haldeman, and Nixon, Ehrlichman (Hankamer & Sag 1976:
410).
(6) Juan leyó El Quijote pero Maria no_.
Juan read El Quijote but Mary not (Spanish, from Depiante (2000: 105)).
(7) That may not bother you, but it does_ me. (Hoeksema 2006: 335).

Limiting its discussion to one of the most controversially problematic ellipsis

constructions – Bare Argument Ellipsis (2), this paper attempts, within the framework
of Minimalist syntax inaugurated by Chomsky (1995 and subsequent work such as
2000, 2001, 2006, 2008), to provide new arguments in favour of considering Bare
Argument Ellipsis constructions as a PF deletion procedure preceded by a movement
operation. The PF-deletion approach to ellipsis, advocated by Ross (1969), Sag (1976),
Hankamer & Sag (1976), Chomsky & Lasnik (1995), Heim & Kratzer (1998), Depiante
(2000), Merchant (2001, 2003, 2004) inter alia, consider all elliptical environments as
an interface phenomenon, involving a fully derived syntactic structure which is deleted
in PF, and therefore is not pronounced. Having adopted this approach, this leads the
paper to argue contra the second approach which also defines ellipsis as an interface
phenomenon. However, it is not at PF as a deletion procedure as claimed by the first
approach, but at LF as a covert movement (e.g. Bare Argument Ellipsis) (Reinhart
1991), or as a reconstruction operation (e.g. VP-ellipsis)1 (William 1977). Crucially, this
paper also attempts to propose a minimalist analysis of the mechanism of movement of
the remnant phrase outside the elided TP before deletion takes place. This analysis,
unlike the previous analyses, seeks to provide a costless explanation of how this
movement derives. These new arguments in favour of a version of the first approach
and the new movement analysis come from investigating the role of the Arabic negative
marker laysa in Bare Argument Ellipsis constructions. This, to my knowledge, is
virtually neglected by all the previous studies on Arabic negation.
The paper is organised as follows. Section 2 begins by laying out the data that
gives universal rise to Bare Argument Ellipsis phenomenon in order to explore its
syntactic characterization. Section 3 reviews previous leading approaches that attempt
to analyse ellipsis of this sort in English and other languages. Section 4 provides a brief
introduction to the syntactic nature of Bare Argument Ellipsis in Arabic. Section 5
elucidates the role of the negative laysa in Bare Argument Ellipsis and provides
increasing evidence that Bare Argument Ellipsis constructions are sensitive to overt
syntax as a PF deletion process. This section also touches on the information structure
distributed by laysa and its influence on the syntactic derivation of the negated remnant
in the remaining clause. Section 6 proposes a minimalist analysis of the movement of
the contrasted constituent out of TP, which is different from the one proposed in the
literature. Section 7 summarises the major findings.
Before launching into the discussion, it should be noted that all the Arabic data
used here, unless otherwise indicated, are from Modern Standard Arabic (Arabic, for
short).

1
That is, the elided elements are generated with empty terminal nodes at PF, but filled at LF through a
movement/copy operation.

2
 Al-Horais

2. Syntactic characterization of BAE

The term Bare Argument Ellipsis (BAE, henceforward), (also known as

“Stripping”: Ross 1969) is defined as “a rule that deletes everything in a clause under
identity with corresponding parts of a preceding clause, except for one constituent (and
sometimes a clause-initial adverb or negative” (Hankamer & Sag 1976: 409). This
phenomenon is illustrated by the following crosslinguistic data in (8-10), from Heim &
Kratzer (1998: 249), (11), (12), from (Aelbrecht 2006: 2), (13), from Busquets (2005:
7), (14), from Winkler (2002:2) and (15), from Depiante (2000: 102):

(8) Some people smoke, but not many.

(9) Laura left Texas and Lena as well.
(10) Laura drank the milk last night, or perhaps the juice.
(11) maa raʔi-tu zaid-an wa laa khalid-an
neg saw.I Zaid-Acc and neg Khalid-Acc
Intended to mean: “I saw neither Zaid nor Khalid.” [Arabic]
(12) Ik heb gisteren Klaas gezien, maar niet Kim.
I have yesterday Klaas seen but not Kim
“I saw Klaas yesterday, but not Kim.” [Dutch]
(13) Va venir al cinema [la MARTA], però no [EN MIQUEL]
It came to the movies Marta, but not Miquel [Catalan]
(14) Leon kann die Aufgabe lösen, aber nicht Peter
Leon can the task solve, but not Peter [German]
(15) Juan leyo El Quijote pero no Hamlet
Juan read El Quijote but not Hamlet [Spanish]

From the preceding data in (8-15), four salient characteristics of BAE

constructions obtain. First, they do not consist of well-formed full syntactic sentences,
but rather strings composed of a single remnant constituent. Second, they are well
formed as ellipsis only if their LF is identical to the LF of an antecedent clause in the
discourse (see Culicover & Jackendoff 2005: 234). Third, the remnant of BAE, which
can be any XP, occurs in a contrastive relationship with its ‘correlate’ (labeled first by
Reinhart (1991)) and it is usually accompanied by a pre-negative marker (not)2 with or
without but, a focusing adverb (only, also, too) or a modal adverb (perhaps). Last and
more importantly, BAE only occurs in coordination structures and cannot occur in an
embedded clause3 as illustrated by the following example, taken from McShane (2005:
143):

(16) *This is the classroom where we usually study biology, and there is the
classroom where sometimes ø chemistry.

2
If the remnant precedes the negative marker, then we have a different elliptical construction called
pseudostripping as exemplified in (6). Such a sort of ellipsis can be observed in Romance languages (see
Depiante 2000, for more discussion). I will not discuss pseudostripping in this paper since it does not
exist in Arabic.
3
Unless the coordination structure is embedded as a whole such as I think that [he talked to Peter
yesterday, but not to Charlotte] (Merchant 2003, Jones 2004).

3
Al-Horais

3. Bare Argument Ellipsis: theoretical background

Since the introduction of Bare Argument Ellipsis into syntactic theory, in Ross
(1969) and Hankamer & Sag (1976), two major lines of analysis have been predominant
regarding how the internal structure of BAE is constructed. The first one argues that the
remaining sequence in BAE is derived by two procurers: movement and deletion. That
is, ellipsis in BAE, like some other elliptical environments such as Sluicing (see
Merchant 2004), is to be defined in terms of the overt syntactic structure involving
movement and deletion. I call this analysis the Ellipsis analysis. The second line of
analysis, on the other hand, argues contra the Ellipsis analysis, claiming that the
constructions in (8-15) do not involve a deletion operation in second clauses at all. It
does involve movement but different from the one proposed by the Ellipsis analysis.
Under this analysis, the internal structure of BAE is derived by adjoining an NP in the
antecedent clause to an NP fragment by LF movement for Reinhart (1991), and D-
structure for McCawley (1991). I call this analysis the Non-Ellipsis analysis. Both these
two analyses will be discussed in more detail in the next sections, starting first with the
Non-Ellipsis analysis.

3.1. Non-Ellipsis analysis

This analysis is proposed first by Reinhart (1991) and extended later by May
(1991), McCawley (1991) and Fiengo & May (1994). Reinhart mainly built her analysis
on the observation that BAE always occurs with two-place connectives or determiners
such as but not, except for, and etc. And it is well-known that these words require two
paired constituents. Hence, it might be the case that BAE is derived by LF-movement of
the correlate (a constituent within the first conjunct) to the remnant in the second
conjunct forming a coordinated structure, where both its parts c-command each other at
LF (cf. McCawley 1991)4. Under this analysis, the interpretation of the bare argument
structure is obtained by applying lambda abstraction to the trace in the antecedent
clause. Thus, (18a) is the LF derivation for (17), and (18b) is the interpretation induced
by this LF.

(17) The critics liked your book and the public — too.
(18) a.
IP

IP NP2

NP 1 VP NP1 NP2

the critics Conj NP2

liked your book and
e1 the public
The critics
[LF-movement]
b. (the critics and the public) (λx [x liked your book]).

4
See McCauley (1991) for a slight different analysis, where he relates the stripped DP to its correlate by
movement, forming a conjoined DP at D-structure not at LF as argued in Reinhart’s analysis.

4
 Al-Horais

According to (18), the remnant is not followed by any element which needs to be
deleted in PF (see Depinate 2000). It is, therefore, plausible to posit that BAE is not a
phenomenon of ellipsis, but rather involves LF movement forming derived conjoined
phrase which binds a trace in the original position of the raised XP as shown in (18).
That is, in (17), the critics is adjoined to and the public at LF, and that newly created
conjoined DP [The critics [and the public]] is interpreted as one operator binding the
variable created by the LF movement. However, being a derived conjoined phrase
means that BAE must meet the standard conditions of conjunctions which require a
matching in category and case between the correlate and the base-generated remnant;
otherwise the derivation is uninterpretable (Reinhart 1991: 368).
The main argument that Reinhart presents to bolster the Non-Ellipsis analysis,
comes from the claim that BAE and the argument to which it corresponds in the
antecedent clause is constrained by subjacency5 which, as she assumes, holds at LF. In
this connection, she argues that LF movement, like syntactic movement, is subject to
the subjacency condition by making a contrast between (19) and (20) below.

(19) Lucie will admit that she stole the diamonds if you press her, but not the car.
(20) *We have interrogated the burglar who stole the car already, but not the
diamonds.

Reinhart’s analysis of (19) is that both the correlate the diamonds and the
remnant the car move out the complement clause and the relative clause in LF. (20),
however, is ill-formed since Subjacency constraints govern the raising of the car. The
NP cannot be raised out of the relative clause, as well as it cannot be extracted out of a
complex subject, or out of clausal adjuncts. By contrast, (19) is well-formed because
there is no Subjacency violations (no relative clause, no adjunct, etc). Hence, the
coordinated NP the car can be linked to the embedded NP the diamonds. This is
illustrated by the two following abbreviated derivations in (21) and (22) respectively:

5
In generative grammar in the 1970s, movement takes place in a series of small steps determined by the
Subjacency which states that movement cannot cross more than one cyclic (bounding) node. In English,
Bounding Category Parameters are IPs and NPs (i).
(i)Subjacency Condition (Chomsky (1977: 73).
[…X... [α ... [β ...Y ...]…]…X…] where α and β are cyclic nodes
In the 1980s, the Subjacency Condition was formulated in terms of barriers which preclude movement
from crossing two barriers (two functional heads) (Chomsky 1986). In the minimalist framework,
movement is ensured by the Phase Impenetrability Condition (ii). This condition determines that an
operation does not have access to a phase below its head; hence ‘The Phase-Impenetrability Condition
yields a strong form of Subjacency’ (Chomsky 2000: 108). Consequently, long movement successively
targets the edge of every phase.
(ii) Phase Impenetrability Condition (Chomsky 2000: 108):
In phase α with head H, the domain of H is not accessible to operations outside α; only H and its edge are
accessible to such operations.

5
Al-Horais

(21)
PF

NP IP

IP NP
the diamonds

conj NP
but not
spec VP the car
Lucie
V CP

will admit that she stole [the diamonds],if you press her

(22)
PF

NP IP

IP NP
the car

conj NP
but not
spec VP the diamonds
we
V CP

have already interrogated the burglar who stole [the car]

Subjacency, however, would not necessarily provide evidence for the Non-
Ellipsis analysis, since a number of studies have shown that Subjacency does not hold at
LF (see for example, Huang 1982, Pesetsky 1987). Moreover, Lappin (1996: 161)
doubts the possibility that BAE is sensitive to subjacency even if we agree that it holds
at LF. He cites, for example, the sentence in (23), where it is possible to interpret the
negated remnant New York Times as outside of the scope of the syntactic island
containing its correlate Daily Telegraph in the antecedent clause:

(23) John enjoyed reading the article which appeared in the New York Times last
week, but not the Daily Telegraph (Lappin 1996: 161)

This suggests that the Subjacency condition is not responsible for the ill-
formedness of the sentence in (20) above. It presumably stems from “some sort of
locality condition [which] may govern the connection between a bare fragment and the
constituent to which it corresponds in the antecedent clause” (Lappin 1996: 161-162).

6
 Al-Horais

3.2. Ellipsis analysis

This analysis has a long tradition in generative grammar. It basically defines all
ellipsis contexts as an interface phenomenon at PF (as a deletion procedure). Based on
this approach, Hankamer & Sag (1976), who apparently first proposed this analysis,
treat BAE, like other forms of elliptical expressions, as an instance of surface anaphora6
which arises by syntactic processes such as deletion in the course of the derivation.
According to this analysis, the BAE constructions mentioned in (8-15) above, is a
surface anaphor that has a linguistic antecedent and a full-fledged syntactic structure
whose surface representation is rendered opaque by PF-deletion. This implies that in the
syntax and at LF, a BAE construction like (9), here as (24), has a representation like
(25a), whereas at PF the second clause gets elided but for one remnant accompanied by
a focusing adverb as in (25b):

(24) Laura left Texas and Lena as well.

(25) a. Laura [Clause1 left Texas and [Clause 2 Lena left Texas as well]].
b. Laura [Clause1 left Texas i and [Clause 2 Lena left Texas ti as well]].

Culicover & Jackendoff (2005:234) provide, based on the Structural

Uniformity principle7, suggestive evidence that BAE is derived from an underlying
clause. They observe that there must be hidden syntactic structure that is responsible for
the full interpretation, since the remnants in BAE “are fuller than their surface syntactic
would permit. Strong Interface Uniformity says that uniform factors in interpretation
must stem from uniform syntactic sources”. Another piece of evidence comes from the
syntactic fact that the remnant has identical features to its counterpart in the antecedent
sentence such as case-marking, observed first by Ross (1969) and Merchant (2003) in
studying sluicing phenomenon. This can be illustrated by the following data from
Arabic (26) and Modern Greek (27) (from Kolokonte 2006: 4).

(26) ʔaʕTaitu zaid-an l-kitab-a laysa xalid-an/ *xalid-un

gave-1s Zaid-Acc the-book-Acc neg Khalid-Acc/ Khalid-Nom
“I give Zaid the book not Khalid.”

(27) Irthe o Yanis, oxi o Yorgos/ *oxi ton Yorgo

Came the John-Nom, not the George-Nom / not the George-Acc
“John came, not George.”

More recently, a number of studies have extended the Ellipsis analysis and
argued that BAE is a sort of PF deletion of TP, preceded by focus movement of the

6
Hankamer and Sag (1976) showed that anaphora can be subdivided into two types, which they termed
deep anaphora and surface anaphora. A deep anaphor does not require an overt linguistic antecedent
(i.e., finds its antecedent at the non-grammatical level of discourse representation); while a surface
anaphor does need an overt antecedent, either in the same sentence or earlier in the linguistically
represented discourse. BAE is an example of surface anaphora and Null Complement Anaphora is an
example of deep anaphora.
7
Structural Uniformity: “an apparently defective or misordered structure is regular in underlying
structure and becomes a distorted in the course of the derivation” (Culicover & Jackendoff 2005:7).
Structural Uniformity leads to the hypothesis that all phrases of all categories of all languages fall under
the XN-Schema (Jackendoff 1977).

7
Al-Horais

remnant to Spec of Focus before Spell-out (Depiante 2000, Merchant 2001, 2003). In
this regard, Merchant (2003: 1) provides additional support for this PF deletion
approach to BAE by considering a number of arguments. Two of these, especially
relevant to the current study, will be mentioned here. First, the possibility of sloppy
identity, which under ellipsis is contingent on c-command and generally attributed to
the pronoun being a bound variable, is an indicator that BAE is a reduction of a full
clause as shown in the following examples, taken originally from Reinhart (1983:152):

(28) You can keep Rosa in her room the whole day, but not Zelda. [sloppy ok]
(29) Her father played with Rosa the whole day, but not Zelda. [sloppy not possible]

Another indication, discussed also by Merchant, that the conjunction involved in

BAE is sentential conjunction and not DP conjunction comes from the occurrence of
certain sentential adverbs in BAE, preceding the remnant as in (30) below from
Merchant (2003: 2).

(30) Abby speaks passable Dutch, and {probably/possibly/fortunately} Ben, too.

(30) adds, also, an increasing argument that the conjunction involved in BAE is clausal
conjunction and not DP conjunction. If we assume that the right-edge and Ben is
originating as a regular DP conjunct with Abby, this semantically cannot be maintained
since (30) is not synonymous with the counterpart DP-conjunction as exemplified in
(31) below:

(31) Abby and Ben {probably/possibly/fortunately} speak passable Dutch.

Although recent studies such as Depiante (2000); Merchant (2001, 2003),

concur with Hankamer & Sag (1976)’s seminal work in considering BAE as a true case
of ellipsis, they differ in going further and arguing that the remnant phrase in BAE
undergoes movement into Spec of FocP, which is followed by phonological deletion. In
more technical terms, BAE is an instance of TP ellipsis following movement of the
remnant to a Focus Position. Both deletion and movement, according to Merchant
(2001, 2003 and related work), are motivated by an E(llipsis) feature8 that imposes
syntactic, semantic and phonological requirements of ellipsis to be licensed. More
specifically, The E-feature instructs the grammar to delete everything below the head on
which is placed. This feature consists of two formal features: an (u)ninterperable (F*)
feature that is in charge of Focus and a (Conj) feature that distinguishes coordinating
conjunctions like and, but from subordinating conjunctions like because, while, etc. In
the case of BAE, the E-feature is placed on F; the head of FP, and checked by the
remnant (DP). This operation is immediately followed by instructing PF (Focus Phrase)
not to pronounce TP [TP-ellipsis]. As for how the E-feature captures the semantic
requirement on BAE, and of course on ellipsis in general, Merchant (2001) asserts that
E-Givenness, which introduces the notion of focus in ellipsis structures, must be met.
The definition of E-Givenness is given in (32), along with the condition on IP ellipsis in

8
It should be noted that Merchant (2001) introduces the E-feature first in his analysis of Sluicing, which
is later extended to other ellipsis types such as BAE or stripping as he prefers to call it (cf. Merchant
2003).

8
 Al-Horais

(33), both taken from Merchant (2001:31). Accordingly, the bracketed BAE’s examples
in (34) and (35), with and without neg, will have the derivation in (36) and (37)
respectively.

(32) E-Givenness:
An expression E counts as e-GIVEN iff E has a salient antecedent A and, modulo
-type shifting9,
(i) A entails F-clo(E)10, and
(ii) E entails F-clo(A).
(33) Focus condition on IP ellipsis:
All IP  can be deleted only if  is e-GIVEN.
(34) John saw Mary, [but not Peter].
(35) John saw Mary, [and Peter too].
11
(36)
but NegP

not Neg'

Neg FP

DP2 F'
Peter
F TP
[E]
t2 saw Mary
(37)
and FP too

DP2 F'
Peter
F TP
[E]

t2 saw Mary

To summarize, this analysis has two different stages of its development. Stage
one considers all elliptical environments, such as BAE, as a PF deletion as argued in
Hankamer & Sag (1976). Stage two adds that BAE indeed involves a PF deletion but is
preceded by a movement operation of the remnant out the TP, as argued in Merchant
(2003).
In the remaining sections, the paper considers the syntactic features of BAE in
Arabic, focusing on the role of the negative laysa in this sort of ellipsis and provides a
minimalist analysis, sustaining that BAE involves a combination of movement and
ellipsis as argued by the Ellipsis analysis. The proposed minimalist analysis, however,

9
“-type shifting is a type-shifting operation that raises expressions to type <t> and existentially binds
unfilled arguments” (Merchant 2001:14n3).
10
The F-closure of , written F-clo(), is the result of replacing F-marked parts of  with -bound
variables of the appropriate type (modulo -type shifting) (Merchant 2001:14).
11
See Merchant (2003) for another possible derivation for the negative not with BAE.

9
Al-Horais

differs from the Ellipsis analysis in providing a different interpretation of how

movement of the contrasted constituent out of TP operates.

4. Syntactic Characterization of BAE in Arabic

BAE in Arabic and in other languages are in many ways alike. (i) It occurs only
in coordinate structures and never occurs in embedded clauses as shown in (38) below.
(ii) The remnant constituent can be preceded by a (negative or positive) polarity marker
(laysa,wa laa ,bal, laakin) (38-42). That is, the antecedent and reduced clauses can be
either affirmative (38) or negative (42) or one negative but not the other and vice versa.
This distinguishes Arabic BAE from its counterpart in other languages, where the
antecedent must bear affirmative as shown in (8-15) above. (iii) A focusing adverb or a
modal adverb can be overt in the stripped clause (38a). (iv)The remnant and its correlate
in the antecedent clause are contrasted and bear the same morphological case (38-42).

(38) a. ahmad-u ya-ktubu Ŝ-Ŝiʕr-a wa xalid-un (aydhan)

Ahmad-Nom 3m-write the-poetry-Acc and Khalid-Nom (too)
“Ahmad writes poetry and Khalid too.”
b.*zayd-un zaʕama anna ahmad-a ya-ktubu Ŝ-Ŝiʕr-a wa xalid-un
Zayd-Nom claimed that Ahmad-Acc 3m-write the-poetry-Acc and Khalid-Nom
aydhan
too.

(39) raʔaitu Zaid-an (wa) laysa xalid-an

saw.I-1s Zaid-Acc (and) not Khalid-Acc
“I saw Zaid not Khalid.”

(40) maa raʔaitu xalid-an bal zaid-an

neg saw.I-1s Khalid-Acc but Zaid-Acc
“I didn’t see Khalid but Zaid.”

(41) maa raʔaitu xalid-an (wa) laakin zaid-an

neg saw.I-1s Khalid-Acc (and) but Zaid-Acc
“I didn’t see Khalid but Zaid.”

(42) maa raʔaitu xalid-an wa laa zaid-an

neg saw.I-1s Khalid-Acc and neg Zaid-Acc
“I saw neither Khalid nor Zaid.”

10
 Al-Horais

5. The negative Laysa and BAE

5.1. Laysa: its origin and distribution

Laysa originally is made up of laa12 and the noun ʔysa ‘existence’. For
morphological reasons, the second vowel a in laa and the glottal stop in “ʔysa” were
omitted, allowing la to be merged with ysa and ending up yielding the form laysa (Al-
Farahiidi 2003)13. In Semitic languages, an equivalent negative marker is only found in
Early Classical Syraic (Suryāyā ), namely, the negative marker layt which is thought to
be a contraction of the negative laa and the copula ’it (Pat-El 2006).
In addition to the rich negation system that Arabic has14, laysa itself has a rich
variety of different strategies for the expression of sentential negation. It has, at least,
three strategies. The first strategy is that of negating a clause by means of a copular
negative marker which has the characteristics of a verb taking a sentential complement,
negating mainly adjectival or nominal attributes and the existence of something (Holes
1995). This strategy is considered to be the main use of laysa and it is limited to
negating only copular sentences with the absence of an overt verbal copula in the
present tense. The second strategy consists in negating a clause by a negation word
which has the properties of a finite auxiliary (see Elrskhawi 1982, for more discussion).
Laysa in the first and second strategies, unlike the other Arabic negatives, inflects for
agreement with the subject, carrying the same inflectional agreement as the perfective
verb. In the third interesting strategy, laysa appears in the form of a negative particle
and is limited to negate only BAE. This strategy of laysa, as mentioned in section one,
has not been discussed before in the literature. The following data, in which laysa in
bold, give an indication of this richness and the use of these three strategies
respectively:

(43) ʔl-ʔawlaad-u lays-uu sighaar-an

the-boys-Nom neg.3pm little.Acc
“The boys are not little.” [Negative copular verb]
(44) a. laysa-t zaynab-u ta-ktubu Ŝ-Ŝ iʕr-a
neg.3fs Zaynab-Nom 3fs-write-Present the-poetry-Acc
“Zaynab does not write poetry.” [Negative auxiliary verb]
b. laysa kul-u l-madʕuwe-ina *(qad) ja?uu
neg.3m all-Nom the-invited-people-Gen perf marker came.3mp
“Not all of the invited people had come.”
(45) raʔaitu zaid-an laysa xalid-an
saw.I-1s Zaid-Acc neg Khalid-Acc
“I saw Zaid not Khalid.” [Negative particle]

Leaving the two first strategies aside as they are beyond the scope the current
study, the next section considers the third strategy, namely laysa with BAE. More

12
laa is considered to be the default form of negation in Arabic (Benmamoun 2000).
13
Al-Farahiidi is the first one among early Arab grammarians who established of Arabic lexicology and
prosody. He died in 786 AD.
14
Morphologically, Arabic has six main sentential negative markers. These are laa, lam, lan, lammaa,
laysa and maa (Fassi Fehri 1993).

11
Al-Horais

specifically, the intent of the next section is to explain how syntactically and
semantically laysa with BAE is characterised.

5.2. The characteristics of laysa with BAE

Two different strategies used by laysa with BAE constructions must be

distinguished. The first one is that of negating the remnant by merging laysa with the
coordinator particle wa which conveys the meaning but in English. The second one is
via laysa alone without being merged with wa. The two strategies are spelt out in the
following examples respectively.
(46) a. aljamiʕat-u l-jadidat-u sa-tubna fii l-Gahirat-i wa laysa fii
the-university-Nom the-new-Nom will-be-built in the-Cairo-Gen and neg in
l-askaadariat-i
Alexandria-Gen
“The new university will be built in Cairo and not in Alexandria.”
b. Gabal-tu xalid-an wa laysa zaid-an
met-I.3s Khalid-Acc and neg Zaid-Acc
“I met Khalid and not Zaid.”
c. kaana fii d-dar-i xalid-un wa laysa zaid-un
was-past.3ms in the-house-Gen Khalid-Nom and neg Zaid-Nom
“Khalid was in the house and not Zaid.”
(47) a. aljamiʕat-u l-jadidat-u sa-tubna fii l-Gahirat-i laysa fii
the-university-Nom the-new-Nom will-be-built in the-Cairo-Gen neg in
l-askaadariat-i
Alexandria-Gen
“The new university will be built in Cairo not in Alexandria.”
b. Gabal-tu xalid-an laysa zaid-an
met-I.3s Khalid-Acc neg Zaid-Acc
“I met Khalid not Zaid.”
c. kaana fii d-dar-i xalid-un laysa zaid-un
was-past.3ms in the-house-Gen Khalid-Nom neg Zaid-Nom
“Khalid was in the house not Zaid.”

As can be noted, the remnant preceded by laysa has syntactic features

appropriate to its being a part of a sentence which is structurally identical to the
antecedent clause such as having case-marking appropriate to the particular verb or
predicate in the antecedent clause. This syntactic requirement, which is met in (45) and
(46), proves that the syntactic structure of the remnant is fully projected but not
phonologically realized and hence that remnant must be licensed as a part of a full
clause. That is, BAE involves clausal ellipsis as argued in the Ellipsis analysis.
Another syntactic feature that should be noted here is that laysa with BAE does
not select a particular tense or carry phi-features. Its lack of inflection singles it out from
the negative verb laysa mentioned earlier in section 5.1. The following examples in (48-
50) illustrate this major syntactic distinction, in contrast with (43) and (44) above:

12
 Al-Horais

(48) jaʔa rrijal-u (wa) laysa/*las-na nnisaʔ-u

came(PAST).3ms the-men-Nom (and) neg the-women-Nom
“The men came (and) not the women.”
(49) ya-ktubu Ahmad-u Ŝ-Ŝiʕr-a (wa) laysa/*lays-at l-qiSat-a
3m-write(PRESENT) Ahmad-Nom the-poetry-Acc(and) neg/*neg.3fs the-prose-Acc
“Ahmad writes poetry (and) not prose.”

(50) sa-ʔðhab-u illa r-rabaT-i (wa) laysa illa l-Gahirat-i

will-go (FUTURE)-I.1s to Rabat-Gen (and) neg to Cairo-Gen
“I will go to Rabat not to Cairo.”

From the above data, it can be suggested that laysa preceded by wa as in (45) is
distinct from laysa that is not preceded by wa as in (47) with respect to discourse
grammatical function. Whereas the remnant in (46) functions as a contrastive topic, the
discourse function in (47) is a contrastive focus (I am not the first to make this
distinction, see Winkler 2005, for the same distinction in German BAE). In the next
section, I briefly explain how the distribution of the two types of contrastiveness is
implemented by laysa.

5.2.1. Contrastive Topic with laysa

As topic usually is defined as the entity that the utterance is about (Reinhart
1982), Contrastive Topic shows “a combined effect of topicality and focussing”
(Molnár 1998: 135). That is, Contrastive Topic “is topical in the sense that it comes
from a potential Topic and somewhat focal in the sense that the choice of the particular
part is not known to the hearer” (Chungmin 2003a: 155).
The most important point about Contrastive Topic, which singles it out from
Contrastive Focus as will be illustrated later, is that Contrastive Topic can be preceded
by a conjunctive question (see Chungmin 2003b, for very interesting discussion). This
can be evidenced by laysa preceded by wa in BAE. Laysa in such contexts is used as a
negative contrastive topic marker. Consider the following example in (51a) where
(51b), contrary to (51c) is the right response of (51a):

(51) a. hal Gabal-ta zaid-an wa xalid-an?

Q met-you Zaid-Acc and Khalid-Acc?
“Did you meet Zaid and Khalid?”

b. Gabal-tu xalid-an wa laysa zaid-an

met-I Khalid-Acc and neg Zaid-Acc
“I met Khalid not Zaid.”

c. # Gabal-tu xalid-an laysa zaid-an

met-I Khalid-Acc neg Zaid-Acc
The reason why (51c) is not the right answering of the question in (51a) will be
illustrated more in the following section.

13
Al-Horais

5.2.2. Contrastive Focus with laysa

Cross linguistically, two types of foci must be distinguished: information focus

and contrastive focus. The former, exemplified in (52) below, represents the new
information and “marks the non-presupposed nature of the information it carries” (Kiss
1998: 248, see Winkler 2005: 30, for detailed discussion). In contrast, (53) provides an
identificational reading and requires a limited number of contextually given alternatives
(Molnár 2006:204).

(52) VOITTI Samppa Lajunen kultaa.

[Won-Past.3s Samppa Lajunen-Nom gold-PAR(TITIVE)]
“Samppa Lajunen DID win gold.”

(53) KULTAA Samppa Lajunen voitti (eika¨ hopeaa)

[Gold-PAR S.L.-Nom won-Past-3s ( not silver-PAR)]
“It was the gold that Samppa Lajunen won (not silver).”
[Both two examples are from Finnish (Kaiser 2006: 3)]15.

In Arabic, as in many other languages, both types of focus can be found.

Consider the following examples, taken from Ouhalla (1993: 279):

(54) a. ʔallaf-at Zaynab-u QASIIDAT-AN

write(PERF)-3f Zaynab-Nom Poem-Acc
“Zaynab has written a POEM.”

b. QASIIDAT-AN ʔallaf-at Zaynab-u

poem-Acc Write(PERF)-3f Zaynab-Nom
“It is a poem Zaynab has written.”

(54a) is an example of new information focus (see Moutaouakil 1989). Thus, it

represents a felicitous answer to the wh-questions: maadha?allaf-at Zaynab-u ‘what has
Zaynab written?’ and maa al-xabaru? ‘What is up/ new?’ (54b) by contrast, is an
instance of contrastive focus and can be found in contexts where the speaker gives
information which is in conflict with existing information. Thus, it does not represnt a
felicitous answer to either of these two questions. The same can be said about the
contrastive focus with laysa. it is indeed compatible with givig information which is in
conflict with questioner’s expectation.
From this it follows that contrastive focus with laysa, unlike contrastive topic,
mainly represents a felicitous answer to disjunctive yes/no questions. In this case wa
cannot be inserted. This explains why the response in (55b) is acceptable, whereas the
response in (55c) is not.

(55) a.hal Gabal-ta zaid-an?

Q met-you Zaid-Acc?

15
In the source, both examples are not glossed. In the interests of consistency, the gloss is added. Thanks
to Prof. Anders Holmberg for the help.

14
 Al-Horais

“Did you meet Zaid?”

b. Gabal-tu xalid-an laysa zaid-an

met-I Khalid-Acc neg Zaid-Acc
“I met Khalid not Zaid.”
c. # Gabal-tu xalid-an wa laysa zaid-an
met-I Khalid-Acc and neg Zaid-Acc

6. A Minimalist analysis of BAE

In this section, I show how Chomsky’s Minimalist Program is able to add

increasing evidence in favour of the Ellipsis analysis?
Within the minimalist conventions, taking BAE constructions as a combination
of movement and deletion, as argued in the Ellipsis analysis, seems more minimizing
computational cost, though I will argue that the derivation and motivation behind the
movement operation is different from a minimalist perspective.
Leaving the issue of movement aside at this moment, treating BAE
constructions as a PF phenomenon involving deletion is strongly supported by
Minimalism. In this regard, Chomsky & Lasnik (1995) suggest that ellipsis is just an
extreme form of deaccenting that is pronounciation with a low pitch and flat intonation.
It is deaccenting in as much as nothing is heard. Ellipsis also shows the importance of
the principle of Economy of Articulation. When there is a choice, we opt for the
expression which will convey the message with the least amount of articulation. This
least amount of articulation obviously, as suggested by Chomsky & Lasnik (1995), is
determined by a rule of the PF-component that deletes the phonologically redundant
information. The lexical material is only pronounced if the context alone is not
sufficient to yield the right interpretation.
Adopting the assumptions laid down by Minimalism, the non-Ellipsis analysis
cannot be maintained since the level of LF and covert movement can no longer be
invoked in the minimalist syntax. In an early version of the Minimalist Program,
Chomsky (1993) rejects the existence of LF-movement when he suggests that most
candidates for LF-movement; e.g. wh-phrases occurring in situ at Spell-out- do not in
fact undergo any rising at LF. Moreover, as noted by Simpson (2001) data-arguments
used to justify LF-movement in the pre-minimalist era loss their validity with the
framework of the new minimalist model. Hence Simpson (p.203) is led to suggest that
“a natural step forward for the Minimalist Program should now be to dispense with the
notion of LF as a syntactically distinct level of derivation formed via covert movement
and assume instead that Spell-Out is the syntactic endpoint to the derivation” (for
interesting discussion see Simpson 2001: 191-203).
Having given some minimalist theoretical conventions in favour of the deletion
analysis, Arabic BAE with laysa can provide empirical evidence in this regard. This
evidence comes from the fact that Negation with laysa, whether it is a verb or a particle,
is always sentential. More precisely, it is never used for negating constituents of
sentences, in contrast to laa which is either used in sentential or constituent negation.
This can be demonstrated by adopting Klima’s (1964) definition of sentential negation
which relies on the convergence of a set of diagnostic tests: Sentences with sentential
negation are those which permit (i) positive rather than negative tag questions, (ii) tags

15
Al-Horais

with neither rather than so associated with positive sentences, and (iii) continuation with
phrases beginning with not even. Although these tests are specific to English, they
(particularly the not even test) can be modified to other languages. Let us use, for
example, the neither-tag test and not even test to see whether negation with laysa in
BAE constructions is a sentential negation or not.

(56) raʔaitu Zaid-an (wa) laysa xalid-an wa laa-hata omar-a

saw.I-1s Zaid-Acc (and) neg Khalid-Acc and neg-even Omar-Acc
“I saw Zaid not Khalid and not even Omar.”

(57) daʕaw-tu Zaid-an illa lʕaŜaʔ-i (wa) laysa xalid-an wa laa omar-a
invited-I. Zaid for the-dinner but neg Khalid-Acc and neg/neither Omar-Acc
“I invited Zaid for dinner but not Khalid and neither Omar.”

The convergence of both tests in (56) and (57) is a strong indication that laysa
there is a sentential negative marker and thus precedes a construction which involves a
process of deletion, resulting in structure that is not phonetically realized.
There is also syntactic evidence to bolster the Ellipsis analysis. This evidence
comes from the fact that, although in English and related languages there is no well-
formed overt counterpart as exemplified in (58), Arabic and its varieties provide this
well-formed overt counterpart, though it is not commonly used16. This can be illustrated
by (59b) the overt counterpart of (50), repeated here as (59a) and by (60b), from Najdi
Arabic17. It should be noted laysa is not used by this dialect or any other Arabic dialect
today. Instead, maa with some morphological changes is used.

(58) a. Sammy played the piano, but not Lisa.

b. *Sammy played the piano, but not Lisa played the piano (Lones, 2004:1).

(59) a. sa-ʔðhab-u illa r-rabaT-i (wa) laysa illa l-Gahirat-i

will-go (FUTURE)-I.1s to Rabat-Gen (and) neg to Cairo-Gen
“I went to Rabat not to Cairo.”
b.sa-ʔðhab-u illa r-rabaT-i (wa) laysa illa l-Gahirat-I
will-go (FUTURE)-I.1s to Rabat-Gen (and) neg to Cairo-Gen
sa-ʔðhab-u
will-go (FUTURE)-I.1s

(60) a. shfit omar bas mu xalid

saw.I Omar but neg Khalid
I saw Omar but not Khalid”

b. shfit omar bas mu xalid (ana) shifit

saw.I Omar but neg Khalid ( I ) saw.I

16
It should be noted that the overt counterpart only articulates to add extra emphasis to the statement if
denied by the hearer.
17
Najdi Arabic is a variety of the Arabic language spoken in the desert regions of central and eastern
Saudi Arabia.

16
 Al-Horais

Turning to the issue of movement, the current paper also concurs with the
Ellipsis analysis in that BAE involves movement of the remnant out of the domain of
TP, before deletion takes place. A close look suggests that movement prior to the PF
deletion evidently occurs with BAE constructions. Cross linguistically, data from
Islandhood phenomena and Preposition-Stranding fact are compatible with an analysis
in which the remnant is moved to the left edge of its clause before the TP gets elided
(later laysa provides striking evidence for this movement).
As for Islandhood phenomena (see Ross, 1967, Jones 2004), it is obvious that
there is an island in the second conjunct because its antecedent (the first conjunct), is
within an island and more precisely, the structure of the second conjunct is parallel to
the first. Having illustrated that, like other types of movement, BAE respects all types of
island, including Sentential Subject Constraint (61), Complex DP Constraint (62), Wh-
islands (63) and Relative clauses (64) (all examples, below, are taken from Jones 2004:
2).

(61) a. *Who was that Mary kissed t obvious?

b. *That Mark wanted to kiss Emily was obvious, but not Jessica.
(62) a. *Who was the claim that Dan liked t ludicrous?
b. *The claim that Dan liked Chris was ludicrous, but not Danielle.

(63) a. *Bob, I speculated which politician would give t a prize.

b. *I speculated which politician would give Bob a prize, but not Sally.

(64) a. *Who did Adam buy the book that Bill had recommended to t for five dollars?
b. *Adam bought the book that Bill had recommended to Susan for five dollars,
but not Betty.
The above data clearly shows that BAE constructions show sensitivity to
islands, like other types of movement. That is, BAE cannot occur, if the remnant
corresponds to a position from which movement is not permitted.
Further evidence for movement is found in P(reposition) stranding18
requirements (Depiante, 2000). Although P-stranding is possible in BAE in some
languages like English where prepositions can be part of an elided remnant or be absent
as shown (65), some languages like Arabic (50) here as (66), Spanish (67), Greek (68),
prepositions cannot be left stranded (or deferred). In other words, a stripping remnant
from PP must be accompanied by the preposition which selected it. This means that
“languages which don't allow stranding with movement don't allow it in stripping
[BAE] either” (Jones 2004: 2).
(65) a. John talked about Mary but not (about) Susan.
b. This article appeared in the New York Times but not (in) the Daily Telegraph.
(Depiante, 2000: 107)

(66) sa-ʔðhab-u illa r-rabaT-i (wa) laysa *(illa) l-Gahirat-i

will-go (FUTURE)-I.1s to Rabat-Gen (and) neg *(to) Cairo-Gen

18
Preposition stranding is the syntactic construction in which a preposition appears without an object.
(The preposition is then described as stranded or hanging or deferred). For extensive discussion about this
phenomenon, see Takami (1992).

17
Al-Horais

“I will go to Rabat not to Cairo.”

(67) Maria quiere leer en el patio y no *(en) la sala de estar
Maria wants to read in the garden and not *(in) the family room
(Depiante, 2000: 107)
(68) Milisa me ton Yorgo, ala oxi *(me) ti Maria
Talked-1SG with the George, but not *(with) the Mary
‘I talked with John, but not with Mary’ (Kolokonte 2006: 5)

Having proved by the above data that the stripping remnant has moved out of
the elided TP, this movement is assumed to be triggered in order to satisfy the Edge-
feature (Ed-feature19, for short) which, following Winkler (2005: 33), contrastive focus
and contrastive topic are endowed with. That is the Ed-feature which is assigned by the
two types of contrastiveness, consists of two features: F(oc) and T(op) features. The
only way of satisfying the Ed-feature requirement is for a nominal constituent to be
moved into the specifier of Top or Foc, when BAE involves a negative marker or into
the specifier of Foc, when BAE does not.
It should be noted that the negative with BAE is not a head neg. This can be
evidenced by laysa with BAE. As previously illustrated in section 5.1, laysa functioning
as a particle as in BAE does not inflect for phi-features or select a particular tense, in
contrast to its function as a verb (see section 5.1 above). Thus, it is plausible to assume,
in line with Ouhalla (1993)’s analysis of the negative marker maa , that laysa, like maa ,
does not project a NegP, but acts as a morphological marker of what we have called in
section 5.2 negative contrastive focus/ or topic. This means that since laysa is not a
head, it must be then base-generated in the spec of FocP or TopP and the Ed-feature in
head Foc or Top must be erased via movement of the remnant XP: NP, AP, PP to adjoin
to laysa in the spec of FocP/TopP.
The question one may ask at this point is how the head Foc/Top can have Ed-
feature since all the relevant features are generated in the phase heads: C and v*
(Chomsky 2008). To resolve this paradox, I adopt the theory of Feature
Inheritance proposed recently by Chomsky (2006, 2008). In this theory, Chomsky
suggests that phasal heads transmit their Inheritance Features to lower heads. From this
it follows that Ed-feature is transmitted from the phasal head C to the immediate lower
head Foc or Top, as a process of inheritance. This can be illustrated in (69) below.
(69)
CP

C’

C FocP/TopP

Foc’/ Top’

Foc/Top

Features Inheritance [Edge feature]

19
Edge feature is the current version of the "generalized EPP" of Chomsky (2000).

18
 Al-Horais

Movement of the remnant to the spec of Foc/Top can be evidenced by the

ungrammaticality of (70a) where the object xalid-an remains in situ. The object must
move and adjoin to laysa that must select a nominal element as shown in (70b).

(70) a. * Laysa raʔaitu xalid-an bal zaid-an

neg saw.I-1s Khalid-Acc but Zaid-Acc
b. Laysa xalid-an raʔaitu bal zaid-an
neg Khalid-Acc saw.I-1s but Zaid-Acc

(71) below, as well, adds increasing evidence that the moved remnant moves higher
than the spec of TP.

(71) ahmad-u raʔa xalid-an laysa zaid-an

Ahmad-Nom saw.3ms Khalid-Acc neg Zaid-Acc
“Ahmad saw Kalid not Zaid.”

In (71) above, the missing material in the second conjunct can be, in contrast with the
first one, read as follows in (72):

(72) ahmad-u raʔa xalid-an laysa zaid-an ahmad-u raʔa

Ahmad saw.3ms Khalid-Acc neg Zaid-Acc Ahmad.Nom saw.3ms

From (72) it follows that the remnant Zaid-an in (71) moves into a projection higher
than TP since it becomes obvious that the spec of TP is already occupied by the subject
Ahmad. If what I have argued for is right, the negative BAE construction in (62) will
have a derivation along the lines of (73) below:
(73)
CP

spec C'

C FocP

Spec Foc
laysa+Zaid-an
Foc TP
[Edge feature]
[+F]
Aahmad-u raʔa Zaid-an

Needless to say, when laysa is a contrastive topic (wa+laysa), the same

derivation in (64) applies with the difference that the remnant must move into the spec
of TopP to merge with laysa, sitting there.
Now we turn to the analysis of BAE constructions that are not preceded by a
negative marker. Non-negative BAE is derived through the same procedure in (64) with

19
Al-Horais

a slight difference that the landing site of a moved remnant, following Depiante (2000:
133-134) is only the spec of FocP. Accordingly, the remnant clause wa xalid-un
(aydhan) in (38a), produced once again here as (74), will have the derivation shown (in
simplified form) in (75) below:
(74) ahmad-u ya-ktubu Ŝ-Ŝiʕr-a wa xalid-un (aydhan)
Ahmad-Nom 3m-write the-poetry-Acc and Khalid-Nom too.
“Ahmad writes poetry and Khalid too.”
(75) ahmad-u ya-ktubu Ŝ-Ŝiʕr-a wa [xalid-un i [TP t ya-ktubu Ŝ-Ŝiʕr-a] (aydhan)]

7. Conclusion

To recapitulate briefly, on the basis of data from Arabic BAE constructions with
the negative marker laysa, the primary aim of this paper has been to provide a
minimalist analysis to support a PF deletion approach to ellipsis. Two types of evidence
have been provided in this regard. The first one is theoretical, taken from adopting the
minimalist conventions. The second type of evidence is empirical and comes from the
fact that negation with laysa, whether it is a verb or a particle, is always sentential
proved by two sentential negation tests and by the possibility of having a well-formed
counterpart of BAE. This strongly supports that BAE is a PF reduction of a full clause.
Crucially the paper has further discussed the proposal that prior to phonological
deletion takes place, the remnant phrase in BAE undergoes movement into Spec of
FocP to satisfy what is called E-feature (Merchant 2003). Although the paper concurs
with this proposal to a certain extent, it has its own minimalist analysis of how this
movement derives by assuming minimal steps explained in section 6 above.
If the present analysis is on the right track, two immediate advantages can be
gained. First, the proposed landing site of a moved remnant shows the mapping between
syntax and information structure in BAE. Second, in addition to providing a costless
explanation of the internal structure of BAE, this analysis is compatible with the interest
of reducing the range of functional projections since it excludes the NegP from the
building structure of negative BAE.

References
Aelbrecht, L. (2006). Stripping the gap: Stripping and gapping in Dutch. Ms.,
University of Tilburg.
Al-Farahiidi, A. (8th century/2003). kitab alʕayn. Beirut: Dar al-Kutub al-ʕilmiahti.
Benmamoun, A. (2000). The Feature Structure of Functional Categories: A
Comparative Study of Arabic dialects. . Oxford and New York: Oxford
University Press.
Busquets, J. (2005). Stripping vs VP-Ellipsis in Catalan. RR5616, INRIA: Universite
Bordeaux 1. 1-21.
Chomsky, N. (1977). On WH-Movement. In Culicover, P., Wasow, T. & Akmajian, A.
(eds.), Formal Syntax. New Your Academic Press. xx–yy.
Chomsky, N. (1986). Barriers. Cambridge, MA: MIT Press.
Chomsky, N. (1993). A minimalist program for linguistic theory. In Hale, K. & Keyser,
S. (eds.), The view from Building 20. Cambridge, Mass: MIT Press. 1–52.
Chomsky, N. (1995). A Minimalist Program for Linguistic Theory. The Minimalist
Program. Cambridge, Mass: MIT Press. 167–217.

20
 Al-Horais

Chomsky, N. (2000). Minimalist inquiries. In Martin, R., Michaels, D. & Uriagereka, J.

(eds.), Step by step: Essays on minimalism in honor of Lasnik Howard.
Cambridge, Mass: MIT Press. 83–155.
Chomsky, N. (2001). Beyond Explanatory Adequacy. MIT Occasional Papers in
Linguistics 20. Cambridge, MA: MITWPL.
Chomsky, N. (2006). Approaching UG from below. Ms., MIT, Cambridge, Mass.
Chomsky, N. (2008). On Phases. In Freidin, R., Otero, C. P. & Zubizarreta, M. L.
(eds.), Foundational issues in linguistic theory. Cambridge, MA: MIT Press.
Chomsky, N. & Lasnik, H. (1995). The Theory of Principles and Parameters. The
Minimalist Program. Cambridge, Mass: MIT Press. 13–127.
Chungmin, L. (2003a). Contrastive Topic and Proposition Structure. In Di Sciullo, A.
M. (ed.), Asymmetry in Grammar Volume 1: Syntax and semantics.
Amsterdam: John Benjamins. 345–371
Chungmin, L. (2003b). Contrastive Topic and/or Contrastive Focus. In McClure, W.
(ed.), Japanese/Korean Linguistics 12. Chicago: The University of Chicago
Press.
Culicover, P. W. & Jackendoff, R. (2005). Simpler Syntax. Oxford: Oxford University
Press
Depiante, M. (2000). The syntax of deep and surface anaphora: A study of null
complement anaphora and stripping/bare argument ellipsis. Ph.D
Dissertation, University of Connecticut.
Elrskhawi, T. (1982). Aspects of sentential Negation in Arabic: a Contribution to The
Typology of Negation. Ph.D Dissertation, University of Birmingham.
Fassi-Fehri, A. (1993). Issues in the Structure of Arabic Clauses and Words. Kluwer:
Dordrecht
Fiengo, R. & May, R. (1994). Indices and identity. Cambridge, MA: MIT Press
Hankamer, J. & Sag, I. (1976). Deep and surface anaphora. Linguistic Inquiry 7(3).
391– 426.
Heim, I. & Kratzer, A. (1998). Semantics in generative grammar. Oxford: Blackwell
Holes, C. (1995). Modern Arabic: Structures, functions, and varieties. London:
Longman
Huang, C.-T. J. (1982). Logical relations in Chinese and the theory of grammar. Ph.D
Dissertation, MIT.
Jackendoff, R. (1977). X-Bar Syntax: A Study of Phrase Structure. Cambridge,MA:
MIT Press
Jones, J. (2004). Negation and movement in stripping constructions. Ms., University of
Santa Cruz.
Kaiser, E. (2006). Negation and the left periphery in Finnish. Lingua 116(3). 314-350.
Kiss, K. E. (1998). Identificational versus information focus. Language 74(2). 45–273.
Klima, E. (1964). Negation in English. In J, F. & J, K. (eds.), The Structure of
Language. N.Jersey Prentice-Hall. 246-323.
Kolokonte, M. (2006). Bare Argument Ellipsis and Polarity in Modern Greek. First
Newcastle Postgraduate Conference in Theoretical and Applied Linguistics.
Newcastle University, United Kingdom.
Lappin, S. (1996). The interpretation of ellipsis. In Lappin, S. (ed.), Handbook of
Contemporary Semantic theory. London: Blackwell. 145–175.
May, R. (1991). Syntax, Semantics and Logical Form. In Kasher, A. (ed.), The
Chomskyan Turn. Oxford: Basil Blackwell. 334–59.

21
Al-Horais

McCawley, J. (1991). Contrastive negation and metalinguistic negation. In Dobrin, L.,

Nichols, L. & Rodriguez, R. (eds.), CLS 27: The parasession on Negation.
Chicago: Chicago Linguistic Society. 189–206.
McShane, M. (2005). A Theory of Ellipsis. Oxford: Oxford University Press
Merchant, J. (2001). The syntax of silence. Oxford: Oxford University Press
Merchant, J. (2003). Remarks on stripping. Ms., University of Chicago.
Merchant, J. (2004). Fragments and Ellipsis. Linguistics and Philosophy 27. 661–738.
Molnár, V. (1998). Topic in Focus. Acta Linguistica Hungarica 45. 89-166.
Molnár, V. (2006). On Different Kinds of Contrast. In Molnár, V. & Winkler, S. (eds.),
Architecture of Focus (= Studies in Generative Syntax 82). Berlin: De
Gruyter. 97–233.
Moutaouakil, A. (1989). Pragmatic Functions in a Functional Grammar of Arabic.
Dordrecht:: Faris Publications
Ouhalla, J. (1993). Negation, Focus, and Tense: The Arabic maa and laa. Rivista di
Linguistica 5. 275-300.
Pat-El, N. (2006). Syntactical Aspects of Negation in Syriac. Journal of Semitic Studies
51(2). 329–348.
Pesetsky, D. (1987). Wh-in-Situ: Movement and Unselective Binding. In Reuland, E. J.
& ter Meulen, A. G. B. (eds.), The Representation of (In)definiteness.
Cambridge, MA: MIT Press. 98–129.
Reinhart, T. (1982). Pragmatics and linguistics: An analysis of sentence topics.
Bloomington: Indiana University Linguistics Club.
Reinhart, T. (1983). Anaphora and Semantic Interpretation. Chicago: University of
Chicago Press.
Reinhart, T. (1991). Elliptic Conjunctions- Non-Quantificational QR. In Kasher, A.
(ed.), The Chomskyan Turn. Oxford: Blackwell. 360–384.
Ross, J. R. (1969). Guess who? Proceedings of the fifth regional meeting of the
Chicago Linguistic society, Ill. University of Chicago.
Sag, I. (1976). Deletion and Logical Form. Ph.D Dissertation, MIT.
Simpson, A. (2001). On Covert Movement and LF. In Alexandrova, G. & Arnaudova,
O. (eds.), The Minimalist Parameter. The Netherlands: John Benjamins.
191-204.
Smith, N. (2001). Ellipsis Happens, and Deletion Is How. Working Papers in
Linguistics 11. 176-191, University of Maryland.
Takami, K. (1992). Preposition Stranding: From Syntactic to Functional Analyses.
Berlin: Walter de Gruyter
Williams, E. (1977). Discourse and logical form. Linguistic Inquiry 8. 101–139.
Winkler, S. (2005). Ellipsis and Focus in Generative Grammar (Studies in Generative
Grammar). Berlin/New York: Mouton de Gruyter

Nasser Al-Horais
School of English Literature, Language and Linguistics.
University of Newcastle upon Tyne
NE1 7RU
United Kingdom

Nasser.Al-Horais@newcastle.ac.uk

22