Negative Polarity Items and Negative Concord in Modern Standard Arabic

by
Muqbil Alanazi

A Thesis Presented in Partial Fulfillment

of the Requirements for the Degree
Master of Arts

Approved April 2013 by the

Graduate Supervisory Committee:
Elly Van Gelderen, Chair
Carrie Gillon
Roy Major

ARIZONA STATE UNIVERSITY

May 2013

ABSTRACT
This thesis explores the distribution of certain lexical items in Modern Standard
Arabic (MSA) and their relationship with two linguistic phenomena, negative concord
(NC) and negative polarity items (NPIs).
The present study examines two central questions: the first question investigates
whether or not MSA shows the patterns of negative concord languages. The second
question concerns the distribution of N-words and NPIs in MSA, and in which
environments they appear. To answer the research questions, the thesis uses the
framework of generative grammar of Chomsky (1995) and The (Non)veridicality
Approach by Giannakidou (1998, 2000, 2002).
The data reveal that MSA shows the patterns of strict negative concord languages
that are suggested by Giannakidou (2000) in the sense that the negative particle
obligatorily co-occurs with the N-words which strengthen the degree of negation, and
never lead to a double negation interpretation.
Moreover, the data show that there is only one pure NPI which appears optionally
in two environments, antiveridical and nonveridical environments, and it is disallowed in
veridical environments. On the other hand, the investigated indefinite nouns show a
mixed picture since they work differently from their counterparts in Arabic dialects.
Their descendants in Arabic dialects appear as NPIs while they tend to be indefinite
nouns rather than NPIs in MSA.
Keywords: negative polarity items, negative concord, Arabic

ACKNOWLEDGMENTS
First and foremost, I would to thank almighty God who gives me strength, power
and patience to complete this work. This thesis would not have been possible without the
efforts of some people who provide me with priceless help.
I give my uttermost gratitude to my chair Professor Elly Van Gelderen for her
invaluable encouragement, support and motivation. She introduced me to the field of
Syntax and the current issues in this field, so thank you Professor Elly. In addition, I
would like to thank my committee members Professor Carrie Gillon and Professor Roy
Major. Without your precious time and help, this thesis would not be possible.
I would like to express my sincere gratitude to my beloved family in Tabuk, Saudi
Arabia especially my parents whose patience and support will not be forgotten.
Last but not least, thank you to those who participated one way or another in this
thesis. Thank you all!

ii

TABLE OF CONTENTS
Page
LIST OF TABLES ........................................................................................................... vi
LIST OF FIGURES .......................................................................................................... vii
LIST OF ABBREVIATIONS ..........................................................................................viii
CHAPTER
1

INTRODUCTION ..................................................................................................1
An Overview of the Thesis ...............................................................................1
The Language of the Study ............................................................................... 2
The Significance of the Study ........................................................................... 5
Research Questions and Methodology.............................................................. 6
Definitions of Key Terms ................................................................................. 6
Conclusion ...................................................................................................... 12

THE THEORETICAL FRAMEWORK OF THE STUDY .................................. 13

Generative Syntax ........................................................................................... 13
The Syntactic Structure and Some Principles ................................................. 15
The (Non)Veridicality Approach .................................................................... 17
Negative Concord (NC) .................................................................................. 21
Conclusion ...................................................................................................... 23

NEGATION AND NC IN MSA ........................................................................... 24

Introduction ..................................................................................................... 24
The Position of NegP in MSA ........................................................................ 24
The Negative Particles in MSA ...................................................................... 25

iii

CHAPTER

Page

Laa and its Variants ........................................................................................ 26

The Negative Particle Maa.............................................................................. 29
Laysa ............................................................................................................... 30
NC in Arabic Dialects ..................................................................................... 32
N-words in MSA ............................................................................................. 35
NC in MSA .................................................................................................... 41
Conclusion ...................................................................................................... 43
4

NPIS IN MSA ....................................................................................................... 44

Introduction ..................................................................................................... 44
aad and ay in Arabic dialects ................................................................... 45
The indefinite Nounay ................................................................................ 46
The indefinite Nounaad ............................................................................. 49
The Determiner NPI ay ................................................................................ 51
Conclusion ...................................................................................................... 54

CONCLUSION ..................................................................................................... 56
Conclusion of Thesis....................................................................................... 56
Questions for Further Research ..................................................................... 57

REFERENCES ................................................................................................................ 59

iv

LIST OF TABLES
Table

Page

1. The Lexical Items That Form NPIs and NC in MSA .................................................. 2

2. N-words in MSA ......................................................................................................... 36
3. The indefinite nouns and the determiner NPI in MSA ............................................... 44

LIST OF FIGURES
Figure

Page

1. The Political and Geographical Distribution of MSA in Southwest Asia and

Africa ...................................................................................................................... 5

vi

LIST OF ABBREVIATIONS
ADV

Adverb

N-word

Negative word

Complementizer

SI

Sensitive item

CP

Complementizer phrase

Spec

Specifier

Determiner

PF

Phonetic form

DN

Double negation

PPI

Positive Polarity items

DP

Determiner phrase

Tense

LF

Logical form

T-bar

MSA

Modern Standard Arabic

TP

Tense phrase

NegP

Negative phrase

VP

Verb phrase

NC

Negative Concord

NP

Noun phrase

NPI

Negative polarity item

vii

Abbreviations used in example glosses

1

First

Masculine

Second

Nom

Nominative

Third

Plural

Acc

Accusative

Question

Gen

Genitive

Singular

Imp

Imperative

Pst

Past

Indic

Indicative

Fut

Future

Ungrammatical sentence

Phonetic transcription of Some Arabic Letters

Glottal stop

Voiced dental fricative

Voiced pharyngeal fricative

Voiceless pharyngeal fricative

Voiced velar fricative

Voiceless postalveolar fricative

Voiceless uvular fricative

Voiceless dental fricative

Pharyngealized (emphatic)

viii

Chapter 1
INTRODUCTION
1.1 An Overview of the Thesis
In the last three decades, the notions of negative concord and negative polarity
items have been a subject of intense debate in the field of contemporary linguistics.
Several influential studies have focused on these two phenomena. The studies include
Baker (1970), Giannikadou (1998, 2000, 2002), Klima (1964), Ladusaw (1980),
Progovac (1994) and Szabolsci (2004).
This thesis is primarily intended to discover the distribution of specific lexical
items in Modern Standard Arabic (MSA hereafter) that are generally sensitive to
negation. These lexical items are mutlaqan, abadan, albattah, bad, ay, aad and ay
(see table 1). They play a prominent role in two linguistic phenomena in relation to
negation: negative concord (NC hereafter) and negative polarity items (NPIs hereafter).
This study consists of five chapters. The first chapter is an introductory chapter
that includes the research questions, the target language, and a definition of the key
terms. The second chapter presents the general framework and the methodology of the
study. In chapter 2, some syntactic and semantic theories are employed as a framework of
the study such as The (Non)Verdicailty Approach by Giannakidou (1998, 2002). The
third chapter describes the syntax of negation and NC in MSA. Chapter 4 explains the
distribution of NPIs in MSA. The thesis is concluded by the final chapter where the
results and conclusion appear.

Table1
The Investigated Lexical items in MSA
The Arabic Script

The lexical item

Parts of speech

Translation

Mutlaqan

Adv

Never

Abadan

Adv

Never

Albattah

Adv

Never

bad

Adv

Not Yet

ay

Any

aad

A person

ay

A thing

1.2 The Language of the Study

Arabic, a Semitic language, is one of the most wide-spread languages in North
Africa and Southwest Asia. It is a part of a language family that is officially called AfroAsiatic family (previously classified as Hamito-Semitic languages) (Ryding, 2005;
Watson, 2002). Besides Arabic, the Semitic language family encompasses other living
languages. Namely, they are Hebrew, Aramaic, Syriac and Amharic language. In
addition, Arabic is the most spoken language from the area of Arabian Gulf to the
Atlantic Ocean in Africa. Arabic is natively spoken by more than 200 million people in
the Middle East and North Africa and more than a billion Muslims who are religiously
motivated to learn Arabic as a second or third language throughout the world because it
represents the medium of their religious teachings and worshiping practices (Ryding,
2005).

Even though there is an ongoing debate of Arabic development and evolution,

there is a consensus among the historians that Arabic generally passed through three
different phases, Classical Arabic, Middle Arabic, and Modern Standard Arabic. The
emergence of Modern Arabic has been accompanied by the growth of Arabic dialects that
are seen as a different form of the main stream language, i.e. MSA. Some historians refer
to these different phases as ancient, medieval and modern Arabic (Chejne, 1969).
In Middle Arabic, there is an ambiguity around the Middle Arabic stage because
there was inadequate documentation of the spoken aspects of that era (Ryding, 2005).
Classical Arabic begun in the sixth century and ended in about the thirteen century, and
evolved as a result of several social, political and religious factors that carried Classical
Arabic to a remarkable level of its own history (Aoun, Benmamoun, Choueiri, 2010;
Ryding, 2005). In this context, the literary growth before and after the Islamic era
witnessed an expansion of rhymed poetry and prose among Arabs most of whom
embraced Islam right after Prophet Mohammad (570-632) was sent. In addition, the
revelation of Holy Quran plays a crucial role to document Arabic. This was successfully
initiated by the third Caliph, Uthman bin Affan (644-656) who started gathering the Holy
Quran from its own resources in one book as it exists today and made about seven copies
which were distributed to the Islamic empire (Ryding, 2005, p. 3).
Aoun, Benmamoun and Choueiri (2010) indicated that at the beginning of the
eighteenth century, the modern period or Modern Standard Arabic (MSA) emerged as an
absolute necessity to preserve Arabic from the adversely dialectal and foreign impact
which began to affect Arabic linguistically, so the process of modernization of Arabic
began to be formalized by the political and educational institutions. As a result, many

Arabic language institutions have been established in many cities as in Damascus, Syria
and Cairo, Egypt by Arab scholars to bridge the gap between the different Arabic dialects
and unify them under one common language (Ryding, 2005).
Scholars have struggled to find a precise definition for MSA. According to
Ryding (2005), there is no a comprehensive MSA definition that governs all aspects of
MSA, so the literature indicates the general lines and features of MSA without providing
an exact definition. Ahmed (2004) states that MSA comes to modernize Classical Arabic
structure. This definition is in accordance with Holes (1994) who said that MSA and
Classical Arabic show almost the same syntax, phonology and morphology, but differs in
the style, usages and lexicon since the MSA is flexible enough to coin new words and
new grammatical constructions. Holes (1994) points out that the contemporary linguistic
situation of MSA is reinforced by outputs of ubiquitous news, airwaves, televisions, and
all levels of education that adopt the MSA as an official medium of instruction. MSA, in
addition, can be defined as the language of formal education, media, newspapers, the
formal religious discourse and the formal communication among the league of Arab
states which politically includes twenty two countries whose first official language is
Arabic (Ryding, 2005). Figure 1 illustrates the political and geographical distribution
where MSA is spoken natively today in Southwest Asia and Africa. Since 1974, Arabic
has been recognized as an official language of the United Nations alongside English,
French, Spanish, Chinese and Russian (Holes, 1994).

Figure 1
The political and geographical distribution of MSA in Southwest Asia and Africa

1.3 The Significance of the Study

This thesis aims to investigate the distribution of two linguistic phenomena, NPIs
and NC in MSA because most studies that deal with these phenomena were conducted
dialectally or in non-Arabic contexts, and not enough attention has been paid to such
phenomena in MSA. Consequently, this research might bridge this gap, and give insights
into their ways of distribution. Second, it is an attempt to figure out which theory,
approach or account can provide a plausible analysis of NPIs and NC in MSA. Finally,
this study is seen as an integration of MSA into the linguistic theory at least in these
linguistic phenomena.

1.4 Research questions and Methodology

This study is an attempt to answer the following questions:
1. Is Modern Standard Arabic a negative concord language?
2. What is the distribution of the lexical items in table 1 in Modern Standard Arabic?
Do they appear exclusively in specific environments?
To achieve the goals of this thesis, the data will be collected from two main
sources: first, the authors Arabic as a representative of Arabic. Then, the invented data
will be double-checked by another native speaker of Arabic to make sure that the
generated sentences are grammatical. The Second source is the previous studies that
addressed the same linguistic phenomena.
1.5 Definitions of Key Terms
Cross-linguistically, there are many lexical items that require certain
environments in order to be valid and felicitously used, and if they occur out of the
appropriate environments, they will be judged ungrammatical. The environments, in
addition, may be various according to their grammatical mood such as positive, negative
or interrogative. Hence, the semantic and syntactic sensitivity of these lexical items is
highly influenced by the environments that allow them to occur. These sets of lexical
items have been called sensitive items (SIs) or polarity items (PIs), which only occur in
specific environments (Baker, 1970; Giannakidou, 2002; Klima, 1964; Szabolcsi, 2004).
More specifically, there are several sub-branches of sensitive items that may be
influenced by their grammatical scope. The main categories of sensitive items that

consistently appeared in the literature are NPIs, positive polarity items (PPIs) and Nwords (Laka, 1990)1.
Historically, the notion of polarity items can be traced back to the works of Klima
(1964) and Baker (1970) who proposed and introduced this new term into the literature,
and the latter is the first one to coin the term negative polarity items in English.
Baker (1970) states that "there are a handful which might be termed polaritysensitive, in that they may occur only in affirmative, or only in negative sentence" (p.
169). As a consequence, his pioneering work triggered other influential NPIs studies that
treated them intensely such as The (Non)veridicality Approach by Giannakidou (1998,
2002).
First of all, let us discover what is meant by polarity conditions in natural
languages (human languages). Radford (2009) states that there are certain types of lexical
items which have an inherent polarity in sense that they are exclusively limited to specific
environments, and these lexical items must be licensed by an effective such as negative,
interrogative or conditional constituents (p. 60, 61). The following examples in (1-3)
explain the polarity conditions:
(1) The student does not play any musical instrument.
(2) Did you bring anything?
(3) If anyone should ask for me, say I have gone for lunch
(Radford, 2009, p. 60)
The examples in (1-3) reflect that there are certain types of words (any, anything and
anyone) that are exclusively used under certain environments, i.e. negative, interrogative
or conditional environments. Klima (1964) indicates that the environments in (1-3) show
1

N-words also called negative indefinites in Haspelmaths terminology (2005)

a shared grammatical-semantic feature which is called later as effective (as cited in

Radford, 2009, p. 60).
In natural languages, the polarity of lexical items can be classified into two
categories: NPIs and PPIs. The former seems to be typologically available in most natural
languages (Giannakidou, 2002). Haspelmath (1997) finds that the indefinite pronouns
such as some and any consistently appear in more than forty languages that he surveyed.
In an earlier work, Baker (1970) defines NPIs as follows: negative-polarity items
are appropriate in structures within the scope of negations, whereas affirmative-polarity
items are appropriate elsewhere (p. 179). The PPIs are inextricably linked to NPIs in
terms of their distribution. They are in complementary distribution where the NPIs appear
clearly in negation environments, and the PPIs found in the opposite environments, i.e.
the affirmative ones. Penka (2011) defines NPIs as words or expressions that can only
occur in contexts that are in some sense negative (p. 102). Surprisingly, negative
environments are not the only contexts which are responsible for the availability of NPIs.
There are other contexts that allow them to occur. For instance, NPIs can be licensed by
questions and conditional sentences (Giannakidou, 1998).
Giannakidou (1998, 2002) shows that Greek NPIs are sensitive to other
environments which are not as strong as the overt negation, and she calls these
environments nonveridical environments such as questions and conditional sentences.
NPIs can be informally outlined in (4):

(4)

Negative Polarity Items:

They are expressions or words that occur within the scope of negation
or negative-like contexts, and they cannot stand alone to express
negation.

Positive polarity items (also referred to as affirmative PIs) (Giannakidou, 2011),

on the other hand, are considered to be a major branch of polarity items beside the
negative polarity items. PPIs have features that make them work differently from the
negative polarity items since they only occur in affirmative contexts. Therefore, they do
not scope under negation and negative-like contexts (Giannakidou, 2011). After the work
of Baker (1970) who differentiated PPIs from NPIs, there are many studies that discuss
the PPIs as a separate class of PIs such as Szabolcsi (2004). The word already and would
rather, for example, provide the best examples of positive polarity items in English, as in
(5) and (6):
(5) He already told you the story.
*He did not already tell you the story.

(Baker, 1970)

(6) I would rather go to New Orleans.

*I would not rather go to New Orleans.

(Baker, 1970)

The examples in (5) and (6) display that there are PPIs, as in already and would
rather, are not permitted to occur within the scope of negation, yet some PPIs , as
mentioned by Szabolcsi (2004), share same features with NPIs such as some /somethingtype PPIs. PPIs can be informally defined as in (7):
(7) Positive Polarity Items (PPIs):
They are lexical items that do not scope under negation.

Having discussed PIs above and how they are divided into NPIs and PPIs, now let
us look at the other part of this paper that concerns NC. NC has received much attention
in literature. The studies include Baker (1970), Giannakidou (2000), Haspelmath, (2005),
Labov (1972), Penka (2011). Moreover, it is cross-linguistically heterogeneous in terms
of meaning and the interpretation of negation.
Before set out to explore this phenomenon, its definitions should be reviewed.
Even though the NC concept can be defined differently depending on the syntax and the
semantics of the target languages, there is an agreement among linguists upon some
general feature of this term. NC is traditionally defined as in (8):
(8) Negative Concord (NC):
Two negative constituents, i.e. a negative particle and an N-word contribute
negation once.
According to this definition, a sizable number of languages display NC, as in
most European languages such as Italian (Haspelmath, 2005; Penka, 2011). The examples
(9-10) provide examples of this phenomenon in European languages:
(9)

Maria non ha visto nessuno.

(Italian)

Maria neg has seen N-word

Maria hasnt seen anybody.
(10) a. Nikto

ne

(Penka, 2011)

priel

(Russian)

Nobody Neg came

Nobody came.
b. Nikto

be

N-person neg

vide Nikogo
saw N-person

10

(Haspelmath, 2005)

No one saw anyone.

(Penka, 2011)

In (9), the NC phenomenon is illustrated in Italian. The N-word nessuno cooccurs with the negative particle non to form NC. Although nessuno and non seem to be
double negation, they, in reality, do not. They should express negation once, and the
meaning is still negative. In Russian, as in (10a-b), the examples reveal that the number
of N-words can be more than one in a sentence like (10b). Nikto and Nikogo are N-words
that co-occur with the main negative particle be to form negation once. This reflects that
more than one N-word may participate in NC, and the meaning is still negative.
Standard English is traditionally classified as a non-negative concord language. It
works in a different fashion from other European languages with respect to the treatment
of the logical interpretation of N-words and negative particles (Penka, 2011). Consider
the following examples in (11a-b):
(11) a. I do not see nobody.
The intended meaning: I saw somebody
b. Nobody did not come.
The intended meaning: Everybody came

(Penka, 2011)

In English, as illustrated, the semantic behavior of N-words and the negative

particles act differently from NC languages because the N-words and the negative
particles cancel each other to express a double negation which causes an affirmative
sense, so the languages that show the pattern of double negation as in English are called
double negation languages (DN). Noteworthy, even though English is not a negative
concord language, some non-standard varieties of English show the ability to express NC

11

like some English vernaculars and Old English (Penka, 2011). N-words can be informally
outlined in (12):
(12) N-word
Lexical elements that carry the negative sense on their own and can stand alone
without the need to be licensed.
1.6 Conclusion
This chapter concerns briefly the core elements of this study. Section (1.1) deals
with the introduction to the thesis and the target lexical items of the study, which are
arranged in table 1. Then, in (1.2), I touched upon the language of the study and the
historical evolution of Arabic language in three historical phases, classical Arabic, middle
Arabic and MSA. The research questions were introduced in section (1.4). Finally, in
(1.5), the definitions of key terms were provided to give the reader general sketches of
the theoretical background of the study before heading to the next chapter which talks
about the general framework of the study.

12

Chapter 2
THE THEORETICAL FRAMEWORK OF THE STUDY
2.1 Generative Syntax
NPIs and NC phenomena need to be investigated within the syntactic and
semantic framework. Hence, this study will draw on some syntactic and semantic
theories, approaches and hypotheses that may help in the analysis of this study.
From the early days of formal generative syntax, linguists have tried to develop
and examine theories that govern language structure and meaning in order to generalize
these theories and discover feasible explanations for linguistically different phenomena.
By tracing back the discipline of generative syntax, many scholars have contributed
effectively and invented some theories and approaches that provide the possible correct
structures for a particular language. In generative syntax, it is highly influenced by the
central tenets of the American Linguist, Noam Chomsky (1965) who defines generative
syntax as:
Generative syntax must be a system of rules that can iterate to generate an
indefinitely large number of structures. This system of rules can be analyzed into
the three major components of a generative grammar: the syntactic, phonological
and semantic components (p. 16).
In his early work, Syntactic Structures, Chomsky (1957) introduced and redefined
some syntactic ideas starting from the language itself, which has been seen cognitively in
the works of Chomsky. In addition, the grammar of a language can be defined as sets of
finite elements of a sentence to produce generatively infinite sentences, and this is
assumed to be applicable to any human language (Chomsky, 1957, p. 2).

13

He also developed what is known as Transformational Grammar, which involves

drawing the relationship between the syntactic elements of a sentence or between
sentences in general. Within this framework, natural languages remarkably have two
levels of representation, surface and deep structure. The semantic meaning has to do
closely with the deep structure which goes beyond the uttered phrases or sentences. For
example, the use of active and passive voice in English clarifies these concepts in
examples (13a-b):
(13)

a. Charlie broke the window.

b. The window was broken by Charlie.

(Yule, 2010, p. 97)

In examples (13a-b), according to Transformational Grammar, each sentence has

two levels of representations, surface structure and deep structure. The former indicates
to the main syntactic lexical categories that each sentence has. Both of sentences consist
of NP and VP, but what makes them different is that each sentence has a different focus.
Even though both sentences have almost the same lexical items, in (13a), the main focus
is on Charlie while in (13b) the focus is on the window rather than Charlie (Yule, 2010,
p. 98).
The salient feature of Transformational Grammar is that human beings are
equipped with an innate linguistic system which enables a native speaker of a language to
speak subconsciously. This is called in Chomskys terms "I-language," which stands for
internalized language. Radford (2009) states that the notion of I-language triggered
Chomsky to invent The Universal Grammar Theory (UG) which seeks to answer the
following question: what are the defining the unique features of the grammars of human
I-language? (p. 13)

14

Language is a model of idealized competence which establishes particular

relationships between the meaning and sounds, and human languages grammar
computationally create syntactic structure which interfaces with other components of
mind, that is, the syntactic structure interacts with the Phonetic Form (PF) and the
semantic components in a complex internalized process (Chomsky, 2006; Radford,
2009). More concretely, Radford (2009) stated that the lexicon (words and phrases) is the
starting point to generate grammar. The words combine with each other to form the
syntactic structure. Then, the semantic component converts the resulted syntactic
structure into the semantic representation, which forms the logical meaning of the
syntactic structure. The third component, the phonetic form (PF) links the previous two
components with each other. The PF explains how the native speaker articulates the
sounds, which are the final language production. In The MP, Chomsky emphasizes
several principles that are supposed to be universally applicable to analyze the syntactic
structure of any language.
2.2 Syntactic Structure and some principles
The syntactic structure is a level of representation that is discussed by Chomsky
(1995). The syntactic structure of any language basically consists of small sets of lexical
items combine with each other to form a constituent, and the lexical structure will be built
-up by series of movements and merger operations (Chomsky, 1995). Consider the
following example in (14) below:
(14) The president may blame himself.

(Radford, 2009, p. 63)

In (14), the reflexive pronoun himself merges with the verb blame to form the VP.
The resulting VP, in turn, merges with may to form the T-bar may blame himself. The T-

15

bar constituent may blame himself immediately merges with the DP, the president, to
form the TP. More importantly, in (14), it reveals that there are principles in The MP that
must not be violated, i.e. the headedness and binarity principles. The former means that
every nonterminal node in a syntactic structure is a projection of the head while the latter
indicates to each nonterminal node is binary-branching (Radford, 2009, p. 43).
The notions of binarity and headedness lead us to know more about X-bar theory
that visually clarifies the various relations between the constituents. X-bar theory is a
syntactic theory for visual representation of the hierarchy of syntactic structures based on
the bar notation (Radford, 2009). Originally, X-bar theory was developed by Chomsky to
capture the deep meaning of the syntactic structures. Therefore, it aims to divide the
syntactic constituents into categories projected by their heads. In the early version of Xbar theory, Chomsky labeled the trees based on three projections: Minimal (head),
Intermediate (X) and maximal projection (XP). By using such categories, tracing the
mutual relationships among the constituents will be easier to navigate. The projections of
X-bar theory are conventionally presented in the following tree in (15):
(15)

XPX
P

X X

The specifier

The head

The complement

By applying this categorization on (14), the X category holds two essential

elements, the head which is called the minimal projection and its complement. The head
and its complement form X that represents the intermediate projection of the whole

16

constituent. Then, the X merges with the specifier of XP to form the maximal projection.
A sentence such as (14) can be drawn according to X-bar theory, as in (16):
(16) The president may blame himself.

(Radford, 2009, p. 63)

TP

DP

V
P

The president

VP

may

P
m

PRN

blam
a

himself
himself
imself

The tree in (16) tells us that there are several principles from The MP applied in
hi

this tree such as the c-command and binding conditions.

C-command (stands for
b
m

mself

constituent command) denotes how two

constituents are related to each other in the
lame
ay

same tree (Chomsky, 1986). In the tree in (16), the reflexive pronoun
himself is chi
mself is a sister of the T, so it ccommanded by the DP the president because the DP node

commands all the constituents underneath it. Furthermore, c-command is crucial thing to
facilitate the binding condition for the grammaticality of the sentence.
T
Even though Chomsky
argues that the categorical information in bar notations is
P
redundant, and should
be removed Tfrom the trees, I will still use the earlier version of X-

bar theory (category-based theory).

2.3 The (Non)Veridicality Approach.
Polarity items (PIs) tends to be more semantic-oriented rather than a syntactic
phenomenon since they can be analyzed purely within the semantic framework. The
(Non)veridicality Approach by Giannakidou (1998, 2002), treats PIs in relation to the
(non)veridicality property which is related closely to the truth condition, whereby NPIs

17

are licensed semantically under nonveridical and antiveridical contexts. Giannakidou

(2002) formalized a definition for the notion of veridicality and (non)veridicality in terms
of the truth as in (17):
(17) (Non)veridicality for propositional operators:
a. A propositional operator F is veridical iff Fp entails p: Fp p;
Otherwise F is nonveridical.
b. A nonveridical operator F is antiveridical iff Fp entails not p:
Fp p.
From the definition (17), there are three veridicality conditions that should be
addressed in detail. The propositional operator F will be one of three conditions:
veridical, nonveridical or antiveridical operator. Veridical operators such as positive
operators do not license NPIs while the nonveridical and antiveridical operators are able
to do so (Giannakidou, 2002). Furthermore, nonveridical environments may include
modal verbs, intentional operators and questions. There are only two antiveridical
contexts: negation and without context (Giannakidou, 2002).
What constitutes as PIs is also addressed in Giannakidou (2002). She redefines
the notion of PIs (NPIs and PPIs) as in (18):
(18) A linguistic expression is a polarity item iff:
a. The distribution of is limited by sensitivity to some semantic
Property of the context of appearance; and
b. is (non)veridicality, or a sub-property thereof: {veridicality,
nonveridicality, antiveridicality, modality, intensionality,
extensionality,episodicity, downward entailingness}.
18

(Giannakidou, 2002, p. 5)
Figuratively, PIs are described as expressions that have a semantic deficiency
that prevents them from working properly in all contexts. Therefore, the deficient items
need to be cured by licensing them under certain environments (Giannakidou, 1998, p.
17).
According to Giannakidou (2002), NPIs are attracted by several environments and
go beyond negation. In the following section, I summarize the licensing environments as
they appear in Giannakidous works (1998, 2002):
NPIs appear in questions. Because questions are classified under the nonveridical
operators that semantically allow NPIs to occur, they are considered to be a good
environment for licensing NPIs (Giannakidou, 1998, 2002), as shown in (19-20) below:
(19) a. Heb

je

ook maar iets gezien?

have-2sg you anything

(Dutch)

seen?

Did you see anything?

b. Idhes

tipota?

(Greek)

saw.perf.2sg API?

(20)

Did you see anything?

(Giannakidou, 2002)

Have you seen any student?

(English)

Another licensing environment is future tense. NPIs are licit in the scope of future
expressions (Giannakidou, 1998, 2002). The future tense has a nonverdical meaning
which allows NPI to occur, but the present and past progressive tend to be veridical.
Hence, they predictably exclude NPIs (Giannakidou, 1998, 2002). This is found in many
languages like English, as in (21):

19

(21) The children will leave as soon as they discover anything

(Giannakidou, 2002)
In addition, habitual adverbs license NPIs, such as the English adverbs as soon as
and usually, but not with always adverbs (Giannakidou, 1998, 2002). In Greek, the
adverb sinithos license the Free Choice Item (FCI) opjodhipote, as illustrated in (22):
(22)

Sinithos dhiavaze opjodhipote vivlio me megali prosoxi.

Usually read-3sg

FCI

book with great attention

S/He usually read any book very carefully.

(Giannakidou, 2002, p. 10)
The generic sentences generally represent an ideal environment to license NPIs
(Giannakidou, 1998, 2002). Most of our sentences are generic in the sense that they
express a well-known relationship between the sentence arguments, for example, any lion
can hunt a gazelle, so NPIs such as any can be used in generic sentences, as in (23):
(23) Any lion can easily hunt a gazelle.
Modal verbs and imperatives allow NPIs to occur within their scope because they
are nonveridical environments (Giannakidou, 1998, 2002), as shown in (24) and (25)
respectively:
(24) Anyone may go to Phoenix.
(25) Pick any piece of paper.
The last two environments, that are reported by Giannakidou (1998, 2002), are the
directive intensional verbs, as in (26) and the conditional sentences, as in (27):
(26) John would like to invite any student.
(27) If you have any problem, you should call 911.

20

(Giannakidou, 2002)

To sum up, Giannakidou (1998, 2002) proposes a semantic approach to capture the
phenomenon of NPIs in relation to the semantic dependency which is seen as the source
of sensitivity that limited the distribution of PIs.
2.4 Negative Concord (NC)
The realization of negation varies from language to another. As mentioned earlier
in chapter 1, negative concord is various cross-linguistically and differs from DN.
Haspelmath (2005) indicates that the vast majority of worlds languages show at least a
pattern of negative concord. It is useful to recall the definition of NC which is stated in
chapter 1. NC is expressed when two negative constituents, a negative particle and an Nword, contribute negation only once. In contrast to NC, in DN, each negative constituent
expresses negation separately, so the DN results in an affirmative sense. Let us draw a
comparison between two examples of contrasting languages, Standard English and
Spanish, in terms of the realization of negation as exemplified in (28):
(28)

a. Nobody has seen nothing.

Intended meaning = everybody has seen something.
b. *(No) vino
neg came

nadie
n-person

Nobody came.

(Penka, 2011)

It is clear from the examples in (28) that DN languages such as Standard English
produce an affirmative sense. In Spanish, on the other hand, the negative particle and the
N-word co-occur with each other to express the negative sense once, as shown in (28b)
(Penka, 2011).

21

Languages can be classified into two different types of NC: strict and non-strict
NC languages (Giannakidou, 2000). Strict NC languages require the negative particle to
obligatorily co-occur with N-words, while non-strict NC languages do not require the
negative particle to accompany the N-word (Giannakidou, 2000; Penka, 2011). Many
Romance languages such as Spanish reflect the patterns of non-strict NC, as in (29):
(29 )

a. *(No) vino
neg came

nadie

(Spanish)

n-person

Nobody came.
b. Nadie

(*no) vino.

n-person neg

came

Nobody came.

(Penka, 2011)

In Spanish, in preverbal position, the negative particle no has to precede the Nword nadie, as in (29a). Otherwise, the sentence is ungrammatical. In contrast to (29a),
the example in (29b) is initiated by the N-word Nadie. In this case, it is not grammatical
to combine the N-word and negative particle if the N-word Nadie is in preverbal position
(Penka, 2012).
The strict NC languages behave in the opposite way of non-strict NC languages.
In the Slavic languages, for example, the negative particle and N-words compulsorily cooccur with each other to express a grammatical negative sentence. This is clearly seen in
Russian as in (30):
(30)

Ja nikogo

*(ne) vizu.

I n-person neg

(Russian)

see

I do not see anyone.

(Brown, 1997)

22

In conclusion, the NC phenomenon is common across languages. The main

assumption of NC phenomenon is that the negation is expressed by two elements, the
negative particle and N-words leading to express negation once. Some languages, in
contrast, do not display the NC .Instead, they show DN. In these languages, the negative
constituents cancel each other to produce an affirmative sense.
2.5 Conclusion
In this chapter, I laid out the details of the general framework of this study by
reviewing the theories and approaches that may give us plausible explanations for the
distribution of NPIs and NC phenomena in MSA. I adopted the Generative Grammar by
Chomsky (1995), X-bar theory and The (Non)veridicality Approach by Giannakidou
(1998, 2002) as a theoretical framework of the study. As for the NC, I utilized the
proposal that is suggested by Giannakidou (2000). In this proposal, languages can be
divided into two types: strict and non-strict NC languages based on the presence of the
negative particle in the sentence.

23

Chapter 3
NEGATION AND NC IN MSA
3.1 Introduction
The syntax of negation is a cornerstone for studying the structure of any language,
and every language has its own ways and rules to form and express negation. In Arabic,
there is much variety of negative particles that can be used to express negation. This
variety in negation becomes more complex especially when we talk about the Arabic
dialects that express negation in more different and complex ways than MSA. However,
in the following, I will review the general sketches of negation in MSA in terms of the
main negative particles and some controversial issues that consistently appear in the
literature.
3.2 The Position of NegP in MSA
Many researchers have established their analysis of negation in Arabic based on
the traditional hypothesis of NegP by Pollock (1989). According to his hypothesis,
negation is projected by its head, the negative particle. Regarding situating the NegP in
MSA, scholars, who studied negation in Arabic, divided into two camps. The first group
tends to be with the traditional explanation of NegP. In this view, the NegP is situated in
internal position, i.e. a position between TP and VP, as in English, as illustrated by the
tree in (31) (Aoun et al., 2010; Benmamoun, 2000; Ouhalla, 2002). The other group
adopts the view that indicates the NegP is projected by the head, negative particle, above
TP position or so-called external position (Fassi Fehri, 1993). In (32) and (33), they
display the internal and external positions in MSA respectively:

24

(31)

The NegP between TP and VP

TP

V
T
P
NegP
Neg
VP

(32)

.
a-tullab-u

laa

ya-drus-un

the-students-nom Neg 3m-study-indic

The students do not study.
(33)

(Benmamoun, 2000)

.
ma aad-un faal-a
Neg one-nom

did-3sm

haadaa
this

no one did this.

(Fassi Fehri, 1993, p. 30)

3.3 The Negative Particles in MSA

In MSA, there are five distinct negative particles. Morphologically, they are laa,
lan, laysa, maa and lam. In the syntax of negation in MSA, there are many studies that
have provided certain explanations to capture the distribution of negative particles in
Arabic. Aoun et al.(2010), Benmamoun (2000), Fassi Fehri (1993) and Ouhalla (2002)
divided these particles into two major groups based on the inflection for tense and
agreement. The first group is represented by laa and its variants: lan, lam and laysa.The
second group, on the other hand, contains maa. The first group, in turn, can be divided
into two sub-groups according to the inflection. Lan and lam inflect for the future and

25

past tense respectively while laysa inflects for agreement, and in the first group, the
negative particles are derived from laa which tends to be neutral in pertaining to the
inflection because it inflects for neither tense nor agreement (Aoun et al., 2010;
Benmamoun, 2000; Fassi Fehri, 1993; Ouhalla, 2002). Moreover, Benmamamoun (2000)
suggests that there are only two negative particles in Arabic, laa and maa and the others
are just inflected variants of laa.
3.3.1 Laa and its variants
As mentioned above, there are two tensed counterparts of the negative particle
laa. They are lam and lan. They inflect for tense but not for agreement. Laa occurs in
present tense interpretation, as in (32). lan occurs within the future tense interpretation
while lam, on the hand, has a past tense interpretation as shown in (34) and (35)
respectively (Aoun et al., 2010; Benmamoun, 2000; Fassi Fehri,1993; Ouhalla, 2002).
The negative laa that occurs in present tense is seen as the default form of
negation in MSA because the negative particles, lan and lam are derived from laa
(Benmamoun, 2000). In addition, in few cases, laa can negate the future tense provided
that the sentence that is being negated must be in the imperfective aspect (present tense)
and has a word or a phrase that indicates the future or the probability, as in (36) below:
(34)

.
t-tullab-uu

Ian

ya-hab-uu

the-students-nom Neg.fut 3m-go-indic

The students will not go.'
(35)

.
a-tullab-u

lam

ya-drus-uu

26

the- students-nom neg-pst 3m-study-indic

The students do not study.
(36)

(Benmamoun, 2000)

qad laa

ya-hadhar-uu

May Neg 3pm-attend-indic

l-awalad-u
the-boys-nom

a-drass-a

adan

the-class-acc tomorrow

The boys may not attend the class tomorrow.

Even though, the sentence in (36) is in the present tense, it has a future
interpretation because the adverb of time adan and the modal particle qad make it
semantically acceptable to negate a future tense sentence by using laa which basically
negates the present tense (for more details about NegP and modality see Fassi Fehri,
(1993)). Such adverbs or expressions that co-occur with laa are traditionally called in
Arabic Qarinah which are words that give a hint to the interpreted tense. Also, the
purpose behind negating a future tense with laa is to express weak negation while
negating a sentence with lan and lam is semantically stronger and more emphasized.
Additionally, Laa is used to negate imperative sentences (Fassi Fehri, 1993), as in
(37):
(37)

Laa t-ahab
Neg 2sm-go
do not go.
Lastly, laa is used as a particle for denial in the discourse especially in the
conversation with yes/no questions (Aoun et al., 2010), as in (38):

27

(38)

.
Laa, lam

a-ktub

No, neg-pst 1s-write

No, I didnt write

(Aoun et al., 2010, p. 112)

Benmamoun (2000) argues that lam and lan are in complementary distribution
with the tensed verbs. When the negative particles, lam or lan inflect for tense, the main
verbs will not be able to do so because in MSA the negation can hold the tense, so in this
case, there is redundant information in the verb and negation as well. Consider the
following example in (39)
(39)

*
*A-tullaab-u

lam

ahab-uu

The-students-nom neg.past went-3pm

The intended meaning: the students didnt go.
(Benmamoun, 2000, p. 96)
The sentence in (39) is ungrammatical because it has the tensed negative particle,
lam which inflects for the past and the main verb which is in the past tense too, so in this
case, both are inflected to the same tense, which is not grammatical in MSA. In a similar
fashion, lan cannot co-occur with the future tense either. All in all, the presence of lan
and lan result in elimination of tense inflection from the main verb in the sentence that is
being negated (Aoun et al., 2010; Benmamoun, 2000).
Benmamoun (2000) suggests that in MSA, the NegP is located in a position
between TP and VP, and laa and its variants are heads of the NegP. Hence, the verb must
28

move throught NegP to the tense in order to check the V feature. In his proposal, the
NegP will block the movement of the verb to the tense, so to prevent that, a merger
between the verb and the Neg-head must be done, and then the resulted complex head
(Neg-head and V) moves to the tense, as presented in the tree in (40):

TP

(40)
T
T
Negn +V

3.3.2 The negative particle Maa

The negative particle maa appears mostly in two environments: the past and the
present tense, as in (41) and (42) respectively (Aoun et al., 2010; Benmamoun, 2000;
Fassi Fehri, 1993). In (41), the negative particle maa negate the past tense. In this case, it
does not have any effects on the verb or the subject. Regarding negating the present tense
with maa, it may have two aspectual interpretations (Fassi Fehri, 1993). In (42), for
instance, it has two semantic interpretations: it means either a habitual action (she doesnt
read) or negating an action at the time of utterance (she is not reading).
(41)

.
maa qar-a

a-saifat-a

Neg read.past-3sm the-newspaper-acc

He has not read the newspaper.
(42)

.
maa ta-qr-u

l-bent-u
29

Neg 3sf-read-indic the-girl-nom

The girl does not read or The girl is not reading
Benmamoun (2000) controversially suggests that the position of maa is slightly
different from laa and its tensed variants because maa is situated in the Spec of NegP. In
this context, the subject must move to the Spec of TP, but maa blocks the subjects
movement to the TP, so to solve this problem, when the subject originated from the Spec
of VP to the Spec of TP, it must merge with the negative particle maa in the spec of
NegP. Then, the whole complex spec moves to the Spec of TP (Benmamoun, 2000).
Ouhalla (1993) runs a counter argument against Benmamouns explanation of
maa. The former suggests that maa has two roles that run parallel, as a negative particle
and focus element. Thus, it does not occupy the Spec-NegP. Rather, it is located in a
position above the TP, i.e. the head of the focusP (as cited in Aoun et al., 2010).
3.3.3

Laysa
Although laysa has received attention in terms of whether it is a negative particle

or an auxiliary verb, there is consensus on some shared environments and features that
make it peculiar among other negative particles in MSA. As another variant of the
negative particle laa, Benmamoun (2000) states that in contrast to lam and lan, Laysa
does not inflect for tense but for agreement, and it has a morphological ability to host
suffixes, and it also appears in two aspectual contexts: verbless sentences and present
tense contexts as in (43) and (44) respectively:
(43)

.
Lays-a

a-ttalib-u

fi al-madrasa-ti

Neg-3sm the-student-nom in the-school-gen

30

The student is not at school.

(44)

.
Lays-at Huda ta-aamal-u

fi al-mustshafa

Neg-3sf Huda 3sf.works-indic

in the-hospital

Huda does not work in the hospital.

Traditionally, Laysa is classified by Arab grammarians as a modal verb (an
incomplete verb) which semantically holds the essence of negation and past tense at the
same time. This explains why it does not appear in the past contexts as negative particle.
Following this assumption, Alkhawalda (2012) argues that Laysa semantically and
syntactically behaves like an auxiliary verb rather than a negative particle. Rarely, laysa
appears in verbal sentences immediately before the verb without separation of subject
provided that the tense of the sentence must be in the imperfective aspect. In this case, it
works syntactically like laa. It does not assign any case and has no effects on the verbal
sentence either, as in (45):
(45)

Laysa ya-lam-u al-walad-u
Neg 3sm-knows-indic the-boy- nom
The boy does not know.

In summary, in this section, I touched upon the main negative particles in Arabic.
There are five negative particles. They are laa, lan, lam, laysa and maa. These negative
particles have been divided into two groups based on the inflection for tense and
agreement. The first group encompasses laa and its variants, lan, lam and laysa. The
second group includes only maa. Lan and lam are tensed variants of laa while laysa
31

inflects for agreement. Maa, on the other hand, inflects for neither tense nor agreement.
Moreover, the syntactic behavior of Laysa is strange somehow. It shows a reluctant
behavior between a negative particle and incomplete verb that holds the sense of negation
and past tense.
3.4 NC in Arabic dialects
The NC phenomenon is wide-spread in Arabic dialects and MSA. Both show the
availability of this linguistic phenomenon with slight differences in the realization of the
semantic interpretation of it. In this section, I will review the studies that are conducted
mainly on Arabic dialects. Then, I will move to the anylsis of the NC in MSA.
Several studies investigate the NC phenomenon in Arabic dialects. Hoyt (2010) in
his dissertation, Negative Concord in Levantine Arabic, investigates a set of N-words that
contribute to NC in Levantine Arabic. They are badan never, blmaa never, not
once, the negative minimizers hawa and qal nothing and the negative scalar particle
wala not one. Hoyt (2010) classified them into three groups based on the semantic
function that they serve. badan and blmaa are classified under N-words group. hawa
and qal are negative minimizers. Wala is a negative scalar focus particle.
N-words in Levantine Arabic are used to serve two functions. First, to express
negation in a sentence fragment, as in (46), and fail to express negation only when they
are used in full clauses because they need to be licensed (Hoyt, 2010), as in (47).
(46)

Q:

mta btabbna?
When indic-tell-2p
When will you tell us?
A: badan
32

Never
(47)

ma:fi

(Hoyt, 2010)

aiy mskila badan.

not-exist which problem never

There isnt a problem ever.

(Hoyt, 2010)

Hoyt (2010) argues that the distribution of never-words in Levantine Arabic

follows the patterns of strict NC that addressed in Giannakidou (2000).
Unlike the N-words in Levantine Arabic, the negative minimizers hawa and qal
which are idiomatic meanings expressing a lack of minimum quantity and the negative
scalar focus particle wala express non-strict NC because they do not need to be licensed
by the negative particle (Hoyt, 2010).
Hoyt (2010) compared Moroccan Arabic N-words htta not even, w:lu
nothing and wala with those that appear in Levantine Arabic. He found that N-words in
Moroccan are similar to N-words in Levantine Arabic in the sense that both express
negation in sentence fragments, and fail to express negation when they are licensed by a
negative particle. These findings are reinforced by Alqassas (2012) who conducted
another study to investigate the Levantine negation. He found that the analysis of the Nword wala-ada no one shows the patterns of non-strict NC, as in (48a-b), so this result
is not in conflict with the analysis of Hoyt (2010) who mentioned that the wala show the
patterns of non-strict NC.
(48)

a. ma-afni-

wala-ada

neg-saw.3sm-me-neg no-one
No one saw me.

33

( Levantine Arabic)

b. wala-ada afni
no-one saw.me
No one saw me.

(Alqassas, 2012)

Surprisingly, Jordanian Arabic exhibits all kinds of NC patterns, i.e. strict and non
strict NC patterns (Alsarayreh, 2012). On one hand, with the N-word and yet phrases, it
shows the strict NC type which requires the presence of the negative particle with the Nword and not yet phrases (Alsarayreh, 2012), as illustrated in (49) and (50) respectively:
(49)

Maryam *(ma)-btokil
Mary

tuff bilmarrah

neg-eat-3sf apples

(Jordanian Arabic )

N-word

Mary does not eat apples at all.

(50)

(Alsarayreh, 2012)
laaddlan

Maryam *(ma)-tara

l-ktab

Mary

the-book N-word

neg-bought.3sf

Mary has not bough the book yet.

(Alsarayreh, 2012)

One the other hand, wala shows the non-strict patterns of NC in the sense that in
preverbal position does not need to be accompanied by the negative particle, as in (51a)
while the presence of negative particle is a must when it is in post-verbal position
(Alsarayreh, 2012), as in (51b) below:
(51)

a. wala

waad

NCI-DET one

(*ma) da.

Jordanian Arabic

neg came.3S

No one come.
b. *(ma) da

waad.

wala

(Neg) came.3S NCI

one.

No one come.

(Alsarayreh, 2012)

34

Jordanian Arabic oddly exhibits both types of NC, strict and non-strict. Therefore,
it does not fit fully into one category of NC patterns (Alsarayreh, 2012). However, the
contrast and differences of explanations among the Arabic dialects is generally attributed
to the semantic realization of these words dialectically.
3.5 N-words in MSA
There are few studies that investigate N-words and NC in MSA and Classical
Arabic, with the exception of Hoyt (n.d.) and Lucas (2009). However, there is a
considerable degree of overlap between N-words and NPIs in their lexical-semantic
properties, so it is important to find a way to distinguish N-words from NPIs. Therefore, I
will follow the criteria that have been proposed by Giannakidou (2000) and Penka
(2011). In (52), I synthesize the main characteristics of N-words that have been suggested
consistently in the literature as follows:
(52)

N-words:
An expression can be an N-word if
a. It is associated with negative meaning.
b. It commonly accompanies a negative particle to express negation once.
c. It can stand alone to express negation in a sentence fragment.
d. It can occur in unlimited numbers.

Based on the criteria that are proposed in (52), table 2 shows the N-words in MSA.

35

Table2
N-words in MSA
The Arabic Script

The lexical item

Parts of speech

Translation

mutlaqan

Adv

Never

abadan

Adv

Never

albattah

Adv

Never

bad

Adv

Not Yet

In MSA, there are four adverbial N-words: mutlaqan never, abadan never, albattah
at all and the not yet-expression bad not yet.
These N-words have almost the same syntactic and semantic behavior in terms of
expressing negation. They exhibit negation in full clause when they co-occur with a
negative particle (Hoyt, n.d.), as illustrated by the examples in (53-55):
(53)

Ali-un *(lam)
Ali-nom

ya-ra

al-talib-a

neg.pst 3sm-see

the-student-acc

mutlaqan
N-word

Ali has not seen the student at all.

(54)

Hind-un
Hind-nom

*(Laa) tu-ib-u

al-aywanat-i

Neg 3sf-like-indic the-animals-acc N-word

Hind does not like animals at all.

(55)

albattah

36

Al-awlad-u

*(lan)

the-boys-nom

neg.fut

ya-lab-uu fi al-adiqat-

abadan.

3pm-play-indic in the-park-gen N-word

The boys will never play in the park at all

Moreover, they can express negation in sentence fragments, so they appear as an
elliptical expression in order to avoid the unwanted repetition, and the whole sentence is
unpronounced in PF except the N-word, as in (56). This ellipsis is frequently found in
MSA when a yes/no question is negated by a sentence fragment, which is understood
from the context that the sentence has a full negative clause in the underlying structure,
and the speaker deliberately omitted some parts of the sentence for economic purposes.
This is in perfect accord with the definition above in (52) because mutlaqan, abadan,
albattah, and bad can express negation in fragmentary answers.
(56)

Q:
Hal kalt-a a-tufaat-a?
Q eat-2sm the-apple-acc
Have you eaten the apple?
A1:
Mutlaqn = lam akal a-tufaat-a mutlaqan
Never.
A2:
abadan = lam akal a-tufaat-a abadan
Never.

Although Lucas (2009) argues that the N-word abadan tends only to be used in
the context of negation and with reference to the future, the adverbial N-words in MSA,
37

abadan, mutlaqn and albattah can be correctly used with the reference to the past, the
present and the future tense because they do not have an aspectual interpretation, as in
(53-55) respectively. In contrast to the adverbial N-words, the not yet-expression bad
tends to be only used in a sentence with the reference to the past, as in (57), but not with
(58) because it has the tensed negative particle lan which inflects for the future tense.
(57)

.
Mohammad *(lam)

ya-m

bad

Mohammad neg.pst 3sm-walk

not-yet

Mohammad has not walked yet

(58)

.*
*Mohammad lan

ya-m

bad

Mohammad neg.fut 3sm-walk not yet.

The intended meaning = Mohammad will never walk
Even though the N-words in MSA are felicitously licensed by negation, they may
occur at the beginning (clause-initial), in the middle (clause-medial), or at the end
(clause-final). I assume that the adverbial N-words in MSA are originally located
somewhere in the VP, and they can move optionally in clause-medial or clause-initial
position if they are used for non-standard purposes, as in intellectual speech and poetry.
Hence, a sentence like (53) would possibly have three alternative ways to be expressed
correctly as suggested in (59a-c):
(59)

a.
Ali-un *(lam)
Ali-nom

[Final]

ya-ra

neg.pst 3sm-see
38

al-talib-a

mutlaqan

the-student-acc

N-word

Ali has not seen the student at all.

TP
T
Ali
NegP
Neg
VP

lam

Ali
V

ADV
mutlaqan

DP

yara

a-ltaliba

b.
Ali-un

[medial]

mutlaqan *(lam)

Ali-nom N-word

ya-ra

neg.pst

al-talib-a

3sm-see

the- student-acc

Ali has not seen the student at all.

TP
T
Ali
NegP
Neg
mutlaqan
VP

lam

Ali
V

ADV
DP

yara

a-ltaliba

39

mutlaqan

c.
mutlaqan Ali-un
N-word

[initial]

*(lam)

Ali-nom

neg.pst

ya-ra

al-talib-a

3sm-see

the-student-acc

Ali has not seen the student at all.

CP
TP
mutlaqan
Ali

T
NegP
Neg

mutlaqan

VP

lam

Ali
V
yara

ADV
DP

mutlaqan

al-talib-a

Moreover, the default position of the N-words in MSA is exactly like those in
(53-55), to be in the rightmost of the sentence.
The iteration of adverbial N-words can be seen in few sentences especially in
spoken MSA. Semantically, iteration is used to convey the exaggeration of negation to
make it stronger. For example, the adverbial N-word battatan albattah, which is a variant
of the adverbial N-word albattah, expresses the repetition of N-words in MSA, as in (60).
(60)

*(Lam)

yu-saafir

Zayd-un

batatan albattah

neg.pst 3sm-travel Zayd-nom N-word N-word

Zayd never traveled at all

40

This is supported by some variety of Arabic as in Levantine Arabic which can

express repeated N-words, as in the negative scalar wala which can be repeated many
times to express high degree of emphasis of negation (Hoyt, 2010), as illustrated by (61):
(61)

ana ma:rft
I

wala

wala

wa:had

(Levantine Arabic)

not-knew.1s N-word N-word one

I didnt recognize even one of them

(Hoyt, 2010)

3.6 NC in MSA
Giannakidou (2000) argues that NC languages can be classified into two types
depending on the presence of the negative particle that works as a licensor, so a NC
language can be either strict or non-strict one. In the former, a negative particle
obligatorily co-occurs with N-words while in non-strict NC languages, N-words are not
always accompanied by a negative particle. By applying this categorization on MSA, it
arguably shows the patterns of strict NC language. The examples in (53-55) and (56-60)
display that the negative particle (lam, laa and lan) must accompany the N-words,
mutlaqan, albattah, abadan and bad, no matter where the position of N-words is in the
sentence, as in (65a-c).
The presence of the negative particles is necessary in all negative sentences. In
contrast, N-words are extra and can be eliminated without any syntactic effects on the
sentence. In other words, the grammaticality of a negative sentence has to do essentially
with the presence of the negative particle while the presence of the adverbial N-words has
no effect on the grammaticality of the sentence. The only effect that they have is on the
degree of negation. It will be the strongest when the N-words co-occur with negative
particle, as in (62a-b):
41

(62)

a.
Lam

ya-ba

a-ttajiru-u

al-youm-a

Neg.pst 3sm-sell the-retailer-nom the-today-acc

The retailer did not sell anything today.
b .
*(Lam) ya-ba a-ttajiru-u

al-youm-a

albattah

Neg 3sm-sell the-retailer-nom the-today-acc N-word

The retailer did not sell anything at all today.
In the sentences in (62a-b), they have almost the same constituents, but in (62b),
the negation is reinforced by the N-word albattah that works as an intensifier of the
negation. It can be omitted if the speaker does want to convey a stronger negation.
The strict NC patterns in MSA seem to be similar with the patterns of some Slavic
languages such as Polish and Russian in terms of the distribution of N-words. In this type
of NC languages, the strict one, the co-occurrence of N-words and negative particle never
yields a reading with the double negation interpretation (Penka, 2011).
In NC languages, the N-words failed to express negation since the negation is
essentially expressed by the negative particle, so the N-words are not semantically
active (Giannakidou, 2000). In strict NC in MSA, the N-words are slightly different
from their counterparts in strict NC languages. In Arabic, N-words never lead to double
negation, nor fail to express negation. Their semantic role is to express a high degree of
negation, so they intensify the semantic meaning of the negative particle when they cooccur with it.

42

In brief, MSA appears to be a strict NC language and requires that N-words must
be licensed by a negative particle. However, I will leave this chapter with an open
question regarding the lexical semantics of N-words. In case of N-words in Arabic,
further research needs to be conducted to determine whether they are inherently negative
or not.
3.7 Conclusion
This chapter is mainly about the sentential negation in MSA, and its relation to
NC phenomenon. In section (3.2), I reviewed the suggested positions of NegP in MSA.
Then, I follow Aoun et al (2010), Benmamoun (2000) and Ouhalla (2002) who agrees
that the NegP in MSA is situated in a position between TP and VP. Then, I discussed the
negative particles in MSA in section (3.3). There are five negative particles in MSA.
They are laa, lan, lam, laysa and maa.They show mixed picture in terms of inflecting for
agreement, tense. In section (3.4), I synthesized the studies that are about NC in Arabic
dialects. Most dialects show that they are examples of non-strict NC languages while in
section (3.6), MSA is arguably classified as a strict NC language. Following Giannakidou
(2000), I assume that the presence of a negative particle along with N-words gives us a
good evidence to classify MSA as a strict NC language. In addition, the role of N-words
in MSA is to strengthen negation.

43

Chapter 4
NPIS IN MSA
4.1 Introduction
Having discussed the N-words that contribute to NC in MSA, now let us turn to
the other lexical items which are shown in table 3. In this chapter, I will examine these
lexical items to see whether they are NPIs, PPIs, indefinite nouns or none of these in
MSA and their cognates in some Arabic dialects.
Table 3
The indefinite nouns and the determiner NPI in MSA
The Arabic Script

The lexical item

Parts of speech

Translation

aad

A person

ay

A thing

ay

Any

Classically, NPIs are licensed under negation while PPIs do not scope under
negation. This can be found clearly in the works of Baker (1970) and Klima (1964).
Then, many influential studies revealed that NPIs are allowed in negative-like
environments (Giannakidou, 1998, 2002). In Arabic, not enough attention has been paid
to NPIs, with the exception of Alsarayreh (2012), Benmamoun (1997), Erwin (1969) and
Ouhalla (2002).

44

4.2 aad and ay in Arabic dialects

The First two lexical items in table 2 are aad a person, one, someone and
anyone and ay a thing, anything and something. Arabic dialects have many
descendants of these lexical items as NPIs. These counterparts can be found in Jordanian
Arabic, Levantine Arabic, Moroccan Arabic, Iraqi Arabic and Syrian Arabic (Alsarayreh,
2012; Alqassas, 2012; Erwin, 1969; Lucas2009).
In Jordanian Arabic, Alsarayreh (2012) states that there are four types of NPIs:
nominal NPIs, determiner NPIs, Adverbial NPIs and idiomatic NPIs. The nominal type
includes ada anyone, someone and ii a thing. The nominal NPI ada is derived from
the word aad in MSA and Classical Arabic while the nominal NPI ii is derived from
the word ay in MSA and Classical Arabic (Alsarayreh, 2012). They appear as NPIs in
Jordanian Arabic, as in (63) and (64) respectively:
(63)

Maryam *(ma)-afat

ada.

(Jordanian

Arabic)
Mary

neg-saw.3sf one

Mary did not see anyone.

(64)

(Alsarayreh, 2012)

Maryam *(ma)-akalat ii.

Mary

neg-ate.3sf thing

Mary i no ea any hing.

(Alsarayreh, 2012)

According to Erwin (1969), in Iraqi Arabic, the quantifier kullI works as a NPI.
It consists of two parts: the quantifier kull which means every and I a thing. The
latter is derived from the word ay in MSA and Classical Arabic. The quantifier NPI is
licensed under negation in Iraqi Arabic, as in (65):
45

(65)

la wallah

kulli

ma-ind-I

(Iraqi

Arabic)
Neg by God anything neg-have-1s
I dont have anything

(Erwin, 1969)

Lacus (2009) indicates that in Moroccan Arabic, the word say a thing occurs in
the scope of negation, and it is in contemporary distribution with the negative particle s,
as in (66):
(66)

ey

ana ma anas
I

neg sleeping a thing

Im not sleeping (at all)

(Caubet, 1993)

Also, the word si in Syrian Arabic works as a NPI and appears in negative-like contexts
such as questions, as in (67):
am tod

(67)
Q

nn-i kazzab

intend-2ms a thing that-I liar

Are you implying that Im a liar?

(Cowell, 1964)

4.3 The Indefinite Noun ay

In MSA, the lexical item ay which means a thing, anything or something is
classified as an indefinite noun. Also, it is used semantically to indicate inanimate objects
without a particular reference in mind. Consider the following examples in (68) and (69):

46

(68)

.
Ali-un

lam

yu-dr

ay-an

Ali-nom neg.pst 3sm-bring a thing-acc

Ali did bring anything.
(69)

.
Laa a-taqd-uu

bi anna Ali adr-a

Neg 1s-believe -indic that

ay-an

Ali brought-3sm a thing-acc

I do not think that Ali has brought anything.

In negative contexts, the indefinite noun ay appears in two negative
environments, clause-mate negation and superordinate negation, as in (68) and (69)
respectively. In both sentences, it occurs under the c-command of the negative particle
whether it is in the same clause or in the embeddedclause. If it precedes the negative
particle, i.e. out of the scope of negation, the sentence will be judged ungrammatical, as
in (70):
(70)

*
*Ali-un

ay-an

lam

ya-dr

Ali-nom a thing-acc neg.past 3sm-bring

the intended meaning: Ali did bring anything.
Furthermore, it appears in negative-like contexts, that is, in yes/no questions,
imperatives and conditional sentences, as illustrated in (71), (72) and (73) respectively.

47

(71)

Q: Hal kalt-a
Q

ay-an ?

(Yes/no question)

ate-2sm a thing-acc

Have you eaten anything?

(72)

.
xu

ay-an.

( Imperative)

2sm-take-imp a thing
Take anything.
(73)

.
Law anna-ka ax-ta
If that-you

ay-an,

(Conditionals)
rabt-a fi al-mosabaqa-ti

took-2sm a thing-acc won-2sm in the- competition-gen

If you had taken anything, you would have won in the

competition.
Although the indefinite noun ay shows a high tendency to occur under the scope
of negation and negative-like environments, in rare cases, it occurs in affirmative
contexts, as in (74)
(74)

A-tfl-u

ya-ml-uu

fi yad-h

ay-an

the -child-nom 3sm-hold -imp in hand-his a thing-acc

The child may hold something in his hand.

48

4.3 The Indefinite Noun aad

Unlike the indefinite noun ay, aad is used to indicate the animate beings
especially for the singular, so it can be translated variably as a person, one, someone
or anyone based on the host environment in which it appears, as exemplified in (75) and
(76):
(75)

.
Laa yajib-u

ann ya-zoor-a

al-mareed-a

aad-un.

Neg should-indic that 3sm-visit-indic the-patient-acc a person-nom

The patient should not be visited by anyone.
(76)

.
qbal-a aad-un

min

al-nas-i

Came-3sm a person-nom of the-people-gen

Someone came.
It follows the analysis of the indefinite noun ay in the sense that the indefinite

aad appears in three environments: negative, negative-like and affirmative

environments. In (75), it occurs under the scope of negation while in (76), it occurs in an
affirmative sentence.
In negative environments, it must occur under the c-command of negative particle
as in (77).
(77)

A.
Omar lan

ila aad-n.

ya-hab-u

Omar neg.fut

3sm-go-indic

49

to a person-gen

Omar will not go to anyone.

b. *
* ila aad

Omar lan

To a person

ya-hab-u.

Omar neg.fut 3sm-go-indic

The intended meaning = Omar will not go to anyone.

In addition, it appears in other environments, as in questions, imperatives,
modality and conditionals, as illustrated in (78-81) respectively:

(78)

(Questions)

Q: Hal jaa-a aad-un ?

Q came-3s a person-nom
Did anyone come?
.

(79)

A-dr

(imperatives)

ma-aka aad-an

ila al-jamiat-.

Bring-imp with-2sm a person-acc to the-university-gen.

Bring anyone with you to the university.
.

(80)

Qad ya-dxlu-u

(Modality)
aad-un

al-manzil-a.

May 3s-enter-indic a person-nom the-house-acc

Anyone may enter the house
.

(81)

Law anna aad-an ahad-a

If

al-lis-a,

(conditionals)
la-abla-a

al-urtat-a

that a person-acc saw-indic the-thief-acc will-called-3s the-police -acc

If anyone had seen the thief, s/he will call police.

50

In MSA, the indefinite nouns ay and aad show a mixed picture since they
work differently from their counterparts in Arabic dialects. Their descendants in Arabic
dialects appear as NPIs while in MSA, they show the ability to occur under affirmative
contexts.
4.4 The Determiner NPI ay
The determiner NPI ay can be translated as any+NP. It must be followed by an
NP because it cannot stand alone in the sentence as in (82a), so it needs to be followed by
an NP provided that the NP must be headed by an indefinite noun in order to use ay
felicitously, as in (82b). Otherwise, it will not be grammatical as in (82c):
(82)

a. .*
*Zayd

lam

ya-tari ay-a

Zayd-nom neg.pst 3sm-buy any-acc

The intended meaning = Zayad did not buy any+NP
b.
Zayd

*(lam)

ya-tari ay-a sayarat-in

Zayd-nom neg.pst 3sm-buy any-acc car-gen

Zayad did not buy any car.
c. (*)
Zayd-nom lam ya-tari
Zayd

ay-a

(*a)-sayarat-in

neg.pst 3sm-buy any-acc the-car-gen

Zayad did not buy any car.

51

The determiner NPI ay can be optionally used with indefinite nouns which can
stand alone without ay. Moreover, if the determiner NPI ay premodified the indefinite
nouns, the semantic meaning will be stronger, i.e. it gives emphasis to the NP that
follows ay. The examples in (83) illustrate the relationship between the NP and the
determiner NPI ay.
(83)

A.
Zayd-nom lam
Zayd

ya-tari sayarat-an

neg.pst 3sm-buy a car-acc

Zayad did not buy any car.

b.
Zayd

lam

ya-tari ay-a sayarat-in

Zayd-nom neg.pst 3sm-buy any-acc a car-gen

Zayad did not buy any car at all
Even though the sentence in (83a) is fine in term of the logical meaning,
obviously, the sentence (83b) is semantically stronger because the NP is premodified by

ay which gives it more emphasis.

Regarding the licensing environments, the determiner NPI ay noticeably occurs
in negative and negative-like contexts (antivedical and nonveridical environments in
Giannakidous (1998, 2002) terms), but never occurs in affirmative environments
because it is semantically unacceptable as in (84)
(84)

*Ya-hab-u

Ali

ila ay-I

mal-in.
52

3sm-go-indic Ali-nom to any-gen store-gen

The intended meaning: Ali goes to a store
It can optionally appear in some nonveridical environments, that is, in questions,
future sentences, habitual actions, generic sentences, modality, imperatives and
conditional sentences. The following sentences (85-91) illustrate the occurrence of ay
with these environments respectively:

(85)

Q: Hal raayt-a ay-a

Q

talb-in ?

(Questions)

see-2sm any-acc student-gen

have you seen any student

.

(86)

Sawfa a-hab-u

ila ay-

makan-in.

(future tense)

Will 1s-go-indic to any-gen place-gen

I will go to any place.
.

(87)

adat-an

ta-qra-u

ay-a

kitab-in

(habitual actions)

Usually-acc 3sf-read-indic any-acc book-gen

she usually reads any book.

(88)
ay-u

assad-in ya-qtl-uu

(generic sentences)
ay-a

azal-in

Any-nom lion-gen 3sm-kill-indic any-acc gazelle-gen

Any lion kills any gazelle.

53

(89)

Qad au-adr-u

ila ay-

dawlat-in.

(Modality)

May 1s-travel-indic to any-gen country-gen

I may travel to any country.

(90)

ay-a

xu

(Imperative)
wardat-in

pick-imp any-acc a flower-gen

pick any flower.
. ,

(91)

Law ahab-a ila al-jamiat-,

If

wajad-a

(conditional sentences)
ay-a taxasus-in.

went-3sm to the-university, found-3sm any-acc major-gen

If he had gone to the university, he would have found any major.

Because it optionally premodifies the indefinite nouns, the impact of the
determiner NPI ay on the indefinite noun appears in form of case assigning. Therefore,
the indefinite noun, that follows the determiner NPI ay, must take a genitive case
regardless of its position in the sentence.
To sum up, the NPI ay covers all environments that are suggested by
Giannakidous (1998, 2002). It occurs in antivedical and nonveridical environments, but
it is disallowed in veridical environments.
4.5 Conclusion
This chapter is about the NPIs in MSA. The data show that in MSA, there is one
pure NPI which is the determiner NPI ay which is allowed only in two environments: in

54

antivedical and nonveridical ones. On the other hand, ay and aad tend to be
indefinite nouns since they show the ability to occur in affirmative environments. At this
point, I found the indefinite nouns ay and aad ambiguous and hard to judge whether
they are pure indefinite nouns, PPIs or NPIs. It would be beneficial to address this
concern for further research.

55

Chapter 5
CONCLUSION
5.1 Conclusion of Thesis
The primary goal of this thesis is to investigate the distribution of certain lexical
items that are shown in table 1, which have a good association with two linguistic
phenomena, NPIs and NC in MSA. This set can be divided into three groups. The first
group encompasses the N-words, mutlaqan, abadan, albattah, and the not yet-expression,
bad. In this group, the N-words participate in the NC phenomena in MSA. The second
group consists of the two indefinite nouns: ay and aad. The indefinite nouns appear
in three environments: antivedical, nonveridical and verdical contexts. The third group
includes the determiner NPI ay which is licensed optionally by antivedical and
nonveridical environments, so it is a pure NPI.
Chapter 1 was an introductory chapter, which includes some basic information
about this thesis. The main theme of chapter 1 is summarized in three points: research
questions, the targeted language and the definitions of key terms.
Moving towards chapter 2, I laid out the details of the general framework and
methodology of this study by reviewing the theories and approaches that may give
plausible explanations for the distribution of NPIs and NC phenomena in MSA.
Generally, the thesis draws on the generative syntax and The (Non)veridicality
Approach by Giannakidou (1998, 2002). In regards to the NC, I utilized the proposal that
is suggested by Giannakidou (2000).
Chapter 3 was about the negation and NC in MSA. N-words in MSA are
addressed in this chapter. In MSA, there are four N-words. They are mutlaqan, Abadan,
56

albattah and the not yet-expression, bad. These are classified as N-words because they
express negation in a sentence fragment and never lead to a double negation
interpretation when they co-occur with the negative particles. Furthermore, the
interaction between these N-words and the negative particles in MSA shows that MSA is
an example of NC languages. Specifically, I argue that MSA displays the strict NC
patterns that are suggested by Giannakidou (2000) in the sense that a negative particle
obligatorily co-occurs with N-words. Moreover, the role of N-words when they co-occur
with the negative particles is to increase the degree of negation and to put more emphasis
on it.
In chapter 4, I analyzed the indefinite nouns aad and ay and the NPI ay in
MSA. The determiner NPI ay is captured by The (Non)veridicality Approach by
Giannakidou (1998, 2002). It is only licensed by antiveridical and nonveridical
environments, and it is disallowed to occur in veridical environments.
Even though the indefinite nouns aad and ay show high tendency to occur in
negative and negative-like environments, they appear in affirmative contexts. In Arabic
dialects, their descendants appear as NPIs, as in Jordanian Arabic.
5.2 Further Questions
Throughout writing this thesis, several questions for further research have arisen
concerning negation and some related issues. First, since there is an ongoing debate
between those who want to situate the NegP in a position between the TP and VP, and
those who want to place it in a higher position, i.e. above the TP, it would be helpful to
question where the NegP should be situated in Arabic. Second, as we have seen in
chapter 3, the N-words can occur in clause-initial, the middle or clause-final postion in
57

favor of the preposition and postposition feature in the syntax of Arabic, so this linguistic
phenomenon is a possible area for further research. Third, I may pursue further research
on the lexical semantics properties of N-words in MSA to examine whether they are
inherently negative or not.

58

REFERENCES
Abdelali, A. (2004). Localization in Modern Standard Arabic. Journal of the American
Society for Information Science and Technology, 55(1), 23-28. doi:
10.1002/asi.10340
Alqassas, A. (2012). The morpho-syntax and pragmatics of levantine Arabic negation: A
synchronic and diachronic analysis. (Doctoral dissertation). Retrieved from
ProQuest Dissertations and Theses. (1015168824).
AlKhawalda, M. I. (2012). Syntactic analysis of the Arabic word laysa: An Auxiliary
Verb or a Negative Particle. International Journal of Linguistics, 4(3), pp-469.
doi: 10.5296/ijl.v4i3.2373
Alsarayreh, A. A. (2012). The Licensing of Negative Sensitive Items in Jordanian Arabic.
(Doctoral dissertation). Retrieved from ProQuest Dissertations and Theses.
(1018059701).
Aoun, J., Choueiri, L., & Benmamoun, E. (2010). The syntax of Arabic: Cambridge
University Press.
Baker, C. L. (1970). Double negatives. Linguistic Inquiry, 1(2), 169-186.
Benmamoun, E. (1997). Licensing of negative polarity items in Moroccan Arabic.
Natural Language and Linguistic Theory, 15(2), 263-287. doi:
10.1023/A:1005727101758
Benmamoun, E. (2000). The feature structure of functional categories: a comparative
study of Arabic dialects. New York/Oxford, UK: Oxford University Press.
Brandtler, J. (2012). The Evaluability Hypothesis: The Syntax, Semantics and Pragmatics
of Polarity Item Licensing. Amsterdam: John Benjamins.
Brown, D. (1997), The syntax of negation in Russian: A Minimalist Approach, Stanford:
CSLI Publications.
Chejne, A. G. (1969). The Arabic language: Its role in history. University of Minnesota
Press.
Chomsky, N. (1957). Syntactic structures. The Hague: Mouton.
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge: MIT Press.
Chomsky, N. (1986). Barriers. Cambridge, Mass: MIT Press.
Chomsky, N. (1995). The minimalist program. Cambridge, Mass: The MIT Press.
59

Chomsky, N. (2006). Language and mind. Cambridge University Press.

Cowell, M. W. (1964). A reference grammar of syrian arabic: Based on the dialect of
damascus. Washington: Georgetown University Press.
Dominique, C. (1993). Larabe marocain. Paris: Peeters.
Erwin, W. (1969). A basic course in Iraqi Arabic. Washington: Georgetown University
Press.
Fassi Fehri, A. (1993). Issues in the structure of arabic clauses and words. Dordrecht;
Boston: Kluwer Academic.
Giannakidou, A. (1998). Polarity sensitivity as (non) veridical dependency. Amsterdam;
Philadelphia: J. Benjamins.
Giannakidou, A. (2000). Negative ... concord? Natural Language and Linguistic Theory,
18(3), 457-523. doi: 10.1023/A:100647731570
Giannakidou, A. (2002). Licensing and sensitivity in polarity items: from downward
entailment to nonveridicality. CLS, 38, 29-53.
Giannakidou, A. (2011). Positive polarity items and negative polarity items: variation,
licensing, and compositionality. In C. Maienborn, K. von Heusinger, and P.
Portner (Eds.), Semantics: An International Handbook of Natural Language
Meaning. Berlin: Mouton de Gruyter,1660-1712.
Haspelmath, M. (1997), Indefinite Pronouns, Oxford, Oxford University Press.
Haspelmath, M. (2005), Negative indefinite pronouns an predicate negation, in M.
Haspelmath, M. S. Dryer, D. Gil & B. Comrie (eds), The World Atlas of
Language Structures, Oxford, Oxford University Press, 466469.
Holes, C. (1994). Modern arabic :Structures, functions, and varieties. London; New
York: Longman.
Hoyt, F. N. (n.d.). Negative polarity in Arabic. Unpublished paper. Retrieved from
http://fmhoyt.colliertech.org/Hoyt_EALL_Polarity.pdf
Hoyt, F. N. (2010). Negative Concord in Levantine Arabic. (Doctoral dissertation).
Retrieved from http://fmhoyt.colliertech.org/Hoyt_Dissertation.pdf.
Klima, E. S. (1964). "Negation in English" in J. A. Fodor & J. J. Katz. The structure of
language, Prentice Hall: Englewood Cliffs, N. J., PP. 246-323.

60

Laka, M. I. (1990). Negation in syntax: On the nature of functional categories and

projections. (Doctoral dissertation). Retrieved from ProQuest Dissertations and
Theses. (303918582).
Lucas, C. (2010). The development of negation in Arabic and Afro-Asiatic. (Doctoral
dissertation). University of Cambridge.
Ouhalla, J. (2002), The structure an logical form of negative sentences in Arabic, in J.
Ouhalla and U. Shlonsky (eds.), Themes in Arabic and Hebrew Syntax, Dordrecht,
Kluwer Academic Publishers, 299-320.
Penka, D. (2011). Negative indefinites. Oxford; New York: Oxford University Press.
Pollock, J. (1989). Verb movement, universal grammar, and the structure of IP. Linguistic
Inquiry, 20(3), 365-424. Retrieved from http://www.jstor.org/stable/4178634
Radford, A. (2009). An Introduction to English Sentence Structure. Cambridge, UK:
Cambridge University Press.
Ryding, K. C. (2005). A reference grammar of modern standard Arabic. Cambridge:
Cambridge University Press.
Szabolcsi, A. (2004). Positive polarity-negative polarity. Natural Language and
Linguistic Theory, 22(2), 409-452. doi: 10.1023/B:NALA.0000015791
The general Secretariat of the League of Arab. (n. d). The political and geographical
distribution of MSA in Southwest Asia and Africa [Geographic map]. Retrieved
from http://www.lasportal.org
Watson, J. C. (2002). The phonology and morphology of Arabic. Oxford: Oxford
University Press.
Yule, G. (2010). The study of language. Cambridge University Press

61