Yarmouk University

Department of English Language and Literature

Irbid-Jordan

Ostensive Predicators and Copulas:

Grammaticalization of the Verb “See” in Maghrebi Dialects
as Evidence on the Null Copula Hypothesis

A Thesis submitted in Partial Fulfilment of the Requirements for

the Degree of Master of Arts in Linguistics in the Department of
English Language and Literature

Submitted by
Nadir Mhamdi

Supervised by
Dr. Mohammed Odeh Al-Shorafat

Linguistics Programme
2017

i
ii
 Acknowledgment

My gratitude goes to my supervisor, Dr. Al-Shorafat for his patience and

supervision throughout the thesis writing. It goes without saying that I stand wholly

responsible for any inaccuracies in the thesis.

I would like to express my profound appreciation to the members of the

examining committee for their time and effort dedicated to examine the present

thesis. My gratitude also goes to all of my teachers whose efforts are the beacon that

lit the path of my knowledge-seeking journey. I acknowledge the help and support

of Dr. Lutfi Abu Al-Haija, for I perpetually owe a debt of gratitude to him. I highly

appreciate the help provided by Dr. Hussein Obeidat for his remarks and references

provision prior to the making of this thesis.

Last but not least, I am very grateful for the informants, for they provided me

with raw data and were patient with my queries.

iii
 Dedication

I would love to pay homage to the recently deceased, my friend and brother-

in-law “Nadjib Lalmi”. May this work be in tribute to his tragic passing.

I dedicate the fruit of this work to my mother, my father and my family.

A special “thank you” goes to my uncle Bachir and grandmother Messouda,

and also to my two brothers Naoui and Ahmed keihoul for their support and care.

To my best friend and brother Basem Badarneh whose befriending has been

the highlight of my three-year journey in Jordan.

A brimful of appreciation is expressed to my dear wife who has reached out

for me in times of need and provided solace to me.

I finally dedicate this work to my countries Jordan and Algeria.

iv
 List of Symbols

1. Consonants

Arabic Symbol Description

Consonants
‫ء‬ ʔ Glottal stop
‫ب‬ b Voiced bilabial stop
‫ت‬ T Voiceless dento-alveolar stop
‫ث‬ Ө Voiceless inter-dental fricative
‫ج‬ ʤ Voiced post-alveolar affricate
‫ح‬ ħ Voiceless pharyngeal fricative
‫خ‬ x Voiceless uvular fricative
‫د‬ d Voiced dento-alveolar stop
‫ذ‬ Ð Voiced alveolar fricative
‫ر‬ R Voiced alveo-palatal trill
‫ز‬ z Voiced alveolar fricative
‫س‬ s Voiceless alveolar fricative
‫ش‬ ʃ Voiceless alveo-palatal fricative
‫ص‬ Ş Voiceless alveolar emphatic fricative
‫ض‬ Ḍ Voiced alveolar emphatic stop
‫ط‬ Ţ Voiceless dento-alveolar emphatic stop
‫ظ‬ Ẓ Voiced alveolar emphatic fricative
‫ع‬ ʕ Voiced pharyngeal fricative
‫غ‬ Ɣ Voiced uvular fricative

v
 ‫ف‬ f Voiceless labio-dental fricative
‫ق‬ q Voiceless uvular stop
‫ك‬ k Voiceless velar stop
‫ل‬ l Voiced alveolar lateral
‫م‬ m Voiced bilabial nasal
‫ن‬ n Voiced alveolar nasal
‫هـ‬ H Voiceless glottal fricative
‫و‬ w Voiced labio-velar glide
‫ي‬ Y Voiced palatal glide

2. Vowels

a. Short vowels

‫فتحة‬ a Front, nearly half-open, low unrounded

‫ضمة‬ u Back, nearly close, high rounded
‫كسرة‬ i Front, open, high unrounded

b. Long Vowels

Long vowels are indicated by the double vowels as follows: /aa/,/uu/ and /ɪɪ/.

vi
 List of Abbreviations
ˋ: Intermediate Projection
Acc: Accusative Case
Adj: Adjective
Adv: Adverb
A-P: Articulatory-and-Perceptual
C: Complementiser
CA: Classical Arabic
C-command: Constituent Command
C-I: Conceptual-and-Intentional
COP: Copula
D: Determiner
EF: Edge Feature
EPP: Extended Projection Principle
F: Feminine
FI: Full Interpretation
Foc: Focus
GB: Government Binding Theory
Gen: Gentive Case
GG: Generative Grammar
GL: Grammar of a given language
HMC: Head Movement Constraint
i: instance of occurrence
IC: Idiom Constraint
LF: Logical Form

vii
LI: Lexical Item
M: Masculine
MA: Maghrebi Dialects
MP: Minimalist Program
MSA: Modern Standard Arabic
N: Noun
N: Numeration
NEG: Negation
Nom: Nominative Case
Ø: Null Constituent
Osten: Ostensive Predicator
P: Phrase/Person
PF: Phonetic Form
Pl: Plural
PLD: Primary Linguistic Data
pro: finite clause null subject pronoun
S: Singular
Spec-X: Specifier of a given projection
TAM Marker: Tense, Aspect and Mood Marker
TNS: Tense
Top: Topic
UG: Universal Grammar
V: Verb (VS: verb-subject; SV: subject-verb)
VPISH: Verb-Phrase Internal Subject Hypothesis
XP: X= a given head, P= phrase

viii
 Table of Contents
Acknowledgment .................................................................................................... iii
Dedication ............................................................................................................... iv
List of Symbols .........................................................................................................v
1. Consonants........................................................................................................v
2. Vowels ............................................................................................................ vi
List of Abbreviations ............................................................................................ vii
Table of Contents ................................................................................................... ix
Abstract ................................................................................................................... xi
Chapter One .............................................................................................................1
Introduction ..............................................................................................................1
1.1. Background of the Study ...............................................................................1
1.2. Theoretical Background ................................................................................2
1.2.1. Syntactic Operations in MP ....................................................................7
1.3. Statement of the Problem ..............................................................................8
1.4. Research Question .......................................................................................11
1.5. Purpose of the Study....................................................................................12
1.6. Significance of the Study ............................................................................12
1.7. Sources and Methods...................................................................................13
1.8. Definition of Terms .....................................................................................13
1.8.1. Ostensive Predicator .............................................................................13
1.8.2. Grammaticalization ...............................................................................14
Chapter Two ...........................................................................................................15
Literature Review ..................................................................................................15
2.1. Introduction .................................................................................................15
2.2. Medieval Arabic Theory of Sentence Types ...............................................15
2.3. Arabic Verbless sentences in the Contemporary Theory of Grammar .......17
2.3.1. Verbless sentences are small clauses with no functional projection ....19
ix
 2.3.2. Verbless sentences contain an erased copula that is null in present .....22
2.3.3. Verbless sentences contain a functional tense projection but no copula
27
2.4. Summary......................................................................................................29
Chapter Three ........................................................................................................30
A Little-Known Grammaticalization Path ..........................................................30
3.1. Grammaticalization Theory.........................................................................31
3.2. From Imperatives to Ostensive Predicators ................................................32
3.3. The Verb “See” as a Source of Copulas ......................................................34
3.4. The Case of Maghrebi Dialects ...................................................................35
3.4.1. “raa” as an Ostensive Predicator ...........................................................36
3.4.2. “raa” as a Copula and an Aspect Marker ..............................................37
3.5. Remarks on the Grammaticalization of “raɁaa” .........................................40
Chapter Four ..........................................................................................................43
A New Outlook at Arabic Verbless Sentences.....................................................43
4.1. Functional Categories in Verbless Sentences: A Rationale ........................43
4.2. Conditions for Copula Deletion ..................................................................45
4.3. SVO in Arabic: Canonical Order or Topicalisation ....................................50
4.4. The Null Expletive: External or Internal Merge .........................................70
Chapter Five ...........................................................................................................74
The Structure of Arabic Sentences: A Unitary Characterization .....................74
5.2. The Functional Categories in the Arabic Sentence .....................................74
5.2. The Structure of copular Clauses in Maghrebi Dialects ................................95
5.3. The Negation of “raa” in the Maghrebi Dialects ......................................100
6.Conclusion..........................................................................................................104
References .............................................................................................................106

x
 Abstract

Mhamdi, Nadir. Ostensive Predicators and Copulas: Grammaticalization of the

Verb “See” in Maghrebi Dialects as Evidence on the Null Copula Hypothesis

Master Thesis, Department of English Language, Yarmouk University, 2017.

(Supervisor: Dr. Mohammed Odeh Al-Shorafat)

The current thesis is an attempt to argue for the structure of Arabic clauses. It

builds on the premises of the minimalist program. The primary objective of the study

is to provide a unitary account for Arabic verbless and verbal sentence “SV and VS”.

It is argued throughout the thesis that Arabic verbless sentences contain a verbal

projection whose head is occupied by a verbal affix that lexicalizes in certain

syntactic conditions. The thesis also argues that the subject, be it null or overt,

originates VP-internally. The Null Expletive Hypothesis (Mohammed, 1990; 2000)

is adopted and revisited in order to account for the differences in word order “VS

and SV”. Another major concern in this thesis is the structure of parallel structures

in the Maghrebi Dialects, in which there can be an overt verbal predicate “raa”. It

was argued that the verbal predicate “raa” is the outcome of a grammaticalization

path that has been newly introduced by Creissels (2015).

Key words: Ostensive predicators, grammaticalization, verbless sentences,

Maghrebi Dialects, Null copula Hypothesis, BE-Support Hypothesis.

xi
 Chapter One
Introduction

1.1. Background of the Study

Arabic is a Semitic language belonging to the Afro-Asiatic (Hamito-Semitic)

language family (Aoun et al., 2010). This language family includes languages such

as Syriac, Hebrew and Ethiopian. Arabic is spoken in the Middle East and the

Arabian Peninsula; it spread west to North Africa with the Islamic conquests. It is

estimated that there are more than 400 million native speakers of Arabic.

Classical Arabic (CA) is the result of the standardization of the language of

Quran and the early Arabic literature between the 7th and the 9th centuries. Arabic

has evolved ever since. However, the passing of 14 centuries has not affected CA in

that it is, allegedly, immune to changes as it entertains a holy position in the Islamic

traditions. Modern Standard Arabic (hereinafter referred to as MSA) is a formal

variety of Arabic that is mostly used in education, press and written documents.

Interestingly, there is a number of dialects that emerged from MSA and CA

although the nature of the historical relationship between them is a matter of debate

(see Owens, 2007). The spoken dialects are distinct from each other and from MSA

to a considerable extent. In fact, there are significant differences between MSA and

1
modern Arabic dialects in their phonological repertoires, lexical inventories and

morphosyntactic features.

The Maghrebi Arabic is generally spoken in Morocco, Algeria and Tunisia. it

is also called Western Arabic. The Maghrebi dialects include a set of Arabic and

Berber dialects (Harrat et al., 2016).

1.2. Theoretical Background

Since the beginning of Generative Grammar (GG), a central theoretical issue

is to address Plato’s problem, i.e., how children are capable of acquiring ideal

grammatical competence notwithstanding the penurious input in the process of

acquisition (Chomsky, 1986b). In other words, the Primary Linguistic Data (PLD)

is inefficient compared to the linguistic competence that children achieve.

Built upon a pure cognitive approach, GG postulates that children are innately

(biologically) equipped with a language faculty that enables them to acquire an ideal

competence despite the poor input (PLD) (see Curtiss, 1977; Fodor, 1983; Smith and

Tsimpli, 1995; Jenkins, 2000; Chomsky, 2000, 2001, 2004)).

It is generally believed that the language innate faculty is a set of principles to

build up the rules of Universal Grammar (UG) which is, in turn, a mold upon which

the grammar of any given language (GL) is poured. The Principles of UG have a set

2
of Parameters that are shaped in consonance with the linguistic experience

(exposure to PLD) (see Chomsky, 1981, 1986b; Chomsky and Lasnik, 1993).

Generative Grammar has undergone several refinements since the mid 1950’s;

the latest of which is the Minimalist Program (MP) (Chomsky, 1992). Minimalism,

adopting a firm bottom-up hierarchical derivation of structures, holds the

Government Binding theory (GB) (the finest version of the Principles and

Parameters Theory) as a benchmark as it is a well-structured theoretical framework

that addresses the theoretical concerns in the MP (Hornstein et al., 2004: 17).

Economy is a basic principle in the MP; it subsumes both derivation and

representation of linguistic structures with exclusion of all unnecessary

“superfluous” constituents (Al-Shorafat, 1998). The economy principle applies to

both the structures and the process of deriving these structures.

Economy in the MP applies at two levels: Methodological Economy and

Substantive Economy. Methodological Economy, ensuring parsimony of

explanation, relates to Occam’s Razor, which posits that in explaining a thing, no

more assumptions should be made than are necessary. In other words, “two primitive

relations are worse than one, three theoretical entities are better than four, four

modules are better than five” (Hornstein et al., ibid: 7). The second type of economy

is Substantive Economy. Minimalism is built on locality conditions and

3
wellformedness filters which means that language grammars are economic as such

to certify the maximum structure resources. Substantive Economy materializes as

movement applies only when it must, and that no expressions are superfluous (Full

Interpretation Principle). Examples on the Substantive Economy are: the A-over-A

Condition (Chomsky, 1964), the Minimal Distance Principle (Rosenbaum, 1970),

the Relativized Minimality Principle (Rizzi, 1990) and the Minimal Binding

Requirement (Aoun and Li, 1993).

A basic tenet in Minimalism is that all sentences have form and meaning

properties. Structures interface with systems that give these structures their

Articulatory-and-Perceptual (A-P) properties, and systems that provide Conceptual-

and-Intentional (C-I) characteristics (Al-Shorafat, 1998; Adger and Svenonius

2010). Svenonius (2007) holds that natural languages have features that play a role

in both the syntactic process and phonological representation and/or semantic

interpretation; such features are called Interface Features. The interface of

Grammar with A-P and C-I is a must. If grammatical elements are not interpretable

by this interface, the structure containing these elements is non-readable (ill-

formed). In sum, derivations in MP are minimalized and interpretation is full, i.e. no

elements occur idly in the derivation.

Chomsky (2005: 2) emphasizes that “language is an optimal way to link sound

to meaning”; this stresses the interface of syntactic structures with other cognitive
4
elements. Put clearly, all parameters (Al-Rashed, 2012) or idiosyncratic properties

(Al-Shorafat, 1998; Hornstein et al., 2004) are encoded in the Lexicon. The starting

point of the derivation is selecting items from the lexical array, i.e., numeration.

Numeration is a set of (LI, i) pairs; “LI” is the lexical item and “i” is the instances

of occurrence (Chomsky 1995: 225). For example, the numeration of “I hate her and

you hate her” is represented as follows:

N= {I1, Hate2, Her2, And1, You1}

The computation system “uses these elements to generate derivations and

structural descriptions SD’s” (Al-Shorafat, 1998: 124) by combining them together

through series of syntactic computations. Each lexicon has three features: semantic,

phonological and formal. Formal features include intrinsic or categorical features

such as nouns’ person and gender, and verbs’ case assigning features as well as

optional features added later in the derivation such as nouns’ case and verbs’

agreement features.

Grammar is composed of the syntactic structure which provides information

for two other components: the semantic component and the phonetic form

component which assign the structure a semantic interpretation and a phonological

representation. The computational system arranges items in pairs (π, λ), where π is

a PF object and λ is a LF component; it should be noted that a main principle in the

5
MP is that the pair (π, λ) must have Full Interpretation “FI” (Representational

Economy). If the pair has FI and is legible by the appropriate interface, it converges

at PF and LF (Chomsky 2005: 3) otherwise it crashes.

The minimalist model of grammar posits that the computational system has

access to the lexical items in the numeration through the Select operation. The

syntactic structures are subsequently built through the Move and Merge operations.

The computation, then, splits at a point called spell-out (Al-Shorafat, 1998), the

starting point of overt syntax, resulting in the PF (phonological component) and LF

(covert component).

Retrieved from Hornstein et al. (2004: 67)

6
 1.2.1. Syntactic Operations in MP

There are three major syntactic operations in the minimalist syntax: Merge

(external merge), Move (internal merge) and Agree. These operations apply each in

specific configurations. Merge is the most basic syntactic operation; it is “an

operation that forms larger units out of those already constructed” (Chomsky, 1995:

396). It is always a binary relation; two syntactic objects X and Y are combined to

form a new object {X, Y}. The new object accedes to the properties of either X or

Y signaling either as the head of the pair, and the pair is labeled with the lexical entry

of the head (X-Phrase or Y-Phrase). Not only is merge a binary operation but also it

applies in a recursive fashion. The new formed pair is, in turn, merged with a new

syntactic object once at a time. Merge and move are subject to the Extension

Condition (Chomsky, 1995; Adger, 2002).

Extension Condition: Overt applications of Merge and Move can only target root

syntactic objects (Hornstein et al., 2004: 57).

Move is a case of internal merge (Chomsky, 2001b) where an element that is

formed by external merge moves into a c-commanding position. The Projection

Principle requires that each movement leave a trace marking the original position;

this means that a copy of the moved constituent is left in the place where it originates

(Chomsky, 2001b). Constituents move to satisfy features of a c-commanding head

7
such as EPP feature of the tense head (Chomsky, 1982) or Edge Feature of the

interrogative head (Chomsky, 2005). Once the feature is checked, it is deleted in the

Spec-Head configuration.

Another major syntactic operation in the minimalist syntax is Agree; it is,

arguably, a binary relationship by which the uninterpretable features of a probe are

satisfied by a goal within its C-command domain and vice versa. Features exchange

ensures the deletion of all valued features; a process without which the derivation

crashes. Both Agree and Move are motivated by the need for feature-checking.

However, Move is motivated by a head’s strong unvalued feature that requires a

spec-head configuration while Agree is motivated by weak uninterpretable features

that need not constituent raising. Another difference is that Agree requires both

constituents to be active by means of having uninterpretable features (Chomsky,

1999). The “deactivation” of features necessitates the deletion thereof.

1.3. Statement of the Problem

The medieval Arabic theory of sentence types categorizes sentences by their

first occurring predicative constituent; this classification results in two main

sentence types: the verbal sentence and the nominal sentence. Given that Arabic is

a VSO language, verbal sentences (1) are very common:

8
 1. katab-a lwalad-u ddars-a

wrote.3PSM the-boy-Nom the-lesson-Acc

The boy wrote the lesson

Arabic also allows SVO structures such as

2. Ɂalwalad-u katab-a ddars-a

the-boy-Nom Wrote.3PSM the-lesson-Acc

The boy wrote the lesson

Nominal sentences (Also known as equational sentence (Obeidat & Ferghal,

1994) and verbless sentences (Obeidat & Ferghal, 1994; Aoun et al., 2010)) have no

overt verb:

3. Ɂalwalad-u saʕiid-un

the-boy-Nom happy-Nom

The boy is happy

Verbal sentences are not problematic in that they are comparable in structure

to English sentences. The only issue is determining which among (1) and (2) is the

canonical structure for Arabic verbal sentences, and whether the tense head in (1)

has an [EPP] feature. On the other hand, nominal phrases have no overt copula in

certain contexts such as when occurring in present indicative. However, the past and

future counterparts of (3) are marked with overt verbal projection and tense head.

9
 4. kaan-a Ɂaṭṭifl-u saʕiid-an

was.3PSM the-boy-Nom happy-Acc

The boy was happy

5. sayakuun-u Ɂaṭṭifl-u saʕiid-an

will-be The-boy-Nom happy-Acc

The boy will be happy

The sentences in (4) and (5) are equal in meaning to (3) except for having

different time references. Such structures as (3) are intriguing in that there,

seemingly, is no functional category to hold the tense projection. Nevertheless, The

Full Interpretation principle posits that elements that have a representation at the LF

level must have a syntactic representation as well. In this regard, such sentences are

understood to be in present tense. This prompts the sound assumption that there is a

functional head that holds the tense projection.

The assumption that (3) has a functional tense head and a verbal category

within the projection is economically yielding from a minimalistic standpoint. The

Representational and Derivational Economy principles recommend the use of as few

frameworks as possible to account for as many structures as possible. Such an

assumption means that the binary categorization of structures into verbal and

nominal sentences is rather superficial. Both types can be unitarily characterized

with one framework. However, there is a case asymmetry between (3) and sentences
10
(4) and (5); the complement is assigned nominative case in (3) and accusative case

in (4) and (5). In this regard, any assumption that there is a null copula must account

for case asymmetry.

The Maghrebi Arabic dialects have instances for an overt copula; for example

the MA equivalent of (3) is:

6. ṭṭful raah farɦaan

the-boy is happy

The boy is happy

There is an overt copula in (6). The “raa” can be used to express continuous

time reference:

7. ddraari raahum yalʕbuu

the-boys are.3Pl playing.3Pl

The boys are playing

1.4. Research Question

 What is the structure of Arabic verbless sentences?

 What syntactic operations are responsible for complement case asymmetry in

the presence and absence of copula in (3) and (5)?

 What is the syntactic position of “raa” in Maghrebi Arabic?

11
 1.5. Purpose of the Study

Other languages (English, French…etc.) have equal structures for equational

and non-equational sentences; both contain a tense projection and a verbal copula

with an overt functional head. The absence of an overt copula in Arabic does not

entail its nonexistence. From a minimalist perspective, it is theoretically useful to

present as few frameworks as possible to account for as many structures as possible.

This study aims at proving that Arabic nominal phrases have not only a functional

tense projection but also a verbal copula (verb is a functional category in nominal

phrases) that is nullified in certain contexts. Moreover, this study aims at providing

a framework that is applicable to both Arabic nominal and verbal sentences.

1.6. Significance of the Study

The study addresses an issue that prompted a long-running debate between

linguists (Bakir, 1980; Fehri, 1993; Obeidat & Ferghal, 1994; Benmammoun, 2000

among others). Results from this study are beneficial in characterizing all Arabic

sentences as being lumped up under one analysis. Moreover, the study is based on a

linguistic analysis of verbless sentences in Berber and Meghrebi Arabic dialects. The

view that Arabic nominal sentences is beneficial in that it enables the linguistic

theory to account for all Arabic sentences as having the same basic structure.

12
 1.7. Sources and Methods

There are several issues that will be tackled in this thesis, which requires that

several research methods be adopted in order to fully investigate each issue. Unlike

the society-based disciplines, syntax requires a smaller number informants.

Therefore, this research bases its results on the interviews of a native speaker of

Algerian Arabic and a native speaker of Russian in order to get the target linguistic

structures. Consultation of Standard Arabic experts was also mandatory in the

judgment of the grammaticality, or the lack thereof, of certain structures of MSA.

Subjective judgment of structures was avoided as much as possible. Another means

of collecting data is referring to the related literature about the topic. The articles and

books provided an indispensable source for both materials, i.e., examples, and

theoretical justification.

1.8. Definition of Terms

1.8.1. Ostensive Predicator

Ostensive predicators are grammatical words or expressions that are “combined

with noun phrases to give sentences aiming to draw the attention of the addressee to

the presence of some entity” (Creissels and Taine-Cheikh, 2016: 1). For example,

the French voici, English here is, etc. Ostensive predicators are also termed

“presentative particles” (Petit, 2010).

13
 Ostensive predicators are strongly related to copulas in that the meanings they

entail are close. However, there is a difference between ostensive predicators and

copulas. Ostensive predicators have a deictic component in that their argument must

be present in the communicative context and their clause can be neither in negation

nor in interrogation. For example, the “raa” in Maghrebi Arabic:

8. Raahum lddraari

There-are the-kids

There are the kids

9. * Ma-raahumʃ lddraari

Neg-There-are-Neg the-kids

The kids are not there

1.8.2. Grammaticalization

A process “…leading from lexical to grammatical (functional) categories”

(Heine, 2003: 575). It is a “dynamic, unidirectional historical process whereby

lexical items in the course of time acquire a new status as grammatical and

morphosyntactic forms, and in the process come to code relations that either were

not coded before or were coded differently” (Traugott and König, 1991).

14
 Chapter Two
Literature Review

2.1. Introduction

How human natural languages achieve predication (known in Arabic grammar

as Ɂisnaad ‫إسناد‬or Ɂixbaar ‫ )إخبار‬is central to the Arabic medieval theory of grammar

and even in contemporary linguistics. Predication is generally achieved by verbs or,

in some languages, any other grammatical category (Ibrahim, 2005). Arabic has

sentences with, seemingly, no overt verbal predication.

This chapter briefly introduces the medieval Arabic theory of sentence types.

Moreover, it presents different views about the structure of Arabic nominal

sentences in contemporary linguistics, and provides criticism for these views.

2.2. Medieval Arabic Theory of Sentence Types

The categorization of sentences in the early Arabic theory of grammar was based

on the first occurring predicative constituent. Two types of sentences are recorded:

a. dʒumla fiʕliyya (verbal sentence): starting with a verb.

1. dʒaaʔ-a rrajul-u

came the-man-Nom

The man came

b. jumla ismiyya (nominal sentence): starting with a nominative noun:

15
 2. Ɂaṭṭifl-u saʕiid-un

the-boy-Nom happy-Nom

The boy is happy

This binary categorization of Arabic sentences is based on the principle of

“ʕamal” (government)1 in which the governor “ʕaamil” must always precede the

governee “maʕmuul” (Obeidat & Ferghal, 1994; Peled, 2007). In this view, the verb

is the governor affecting the complement in the verbal sentence whereas the

inception “ʔibtidaaʔ” is the abstract operator in the nominal sentence (Sībawayh Al-

Kitāb, 1: 239). One major issue in this categorization is that some constructions fall

within neither of the aforementioned types. For example, a sentence can start with a

definite propositional phrase (3) or an adverbial phrase (4):

3. fi ddaar-i radjul-aani

In the-house-Gen men-Dl-nom

There are two men in the house

4. hunaak-a radjul-un waaħid-un

there man-nom one-nom

There is one man

1
The Arabic term ʕamal has a number of corresponding terms in western linguistics: government
(Zwartjes, 2007), grammatical effect and operation (Levin, 1979; Versteegh, 1994; Talmon,
1993).

16
 Some later grammarians specified the conditions under which a non-verbal

(specifically participial) predicate is analogous to verbal predication (see Talmon,

1993; Peled, 2007). Ibn Hisham (1360) discusses Arabic sentence types in a way

that is close to that of Sibawayh; he identified a third sentence type to account for

the problems in (3) and (4), that is, the “Jumla ᶁarfiya” “adverbial sentence” (494).

A forth type was introduced by Zamaxshari “Jumla ʃarṭiyya” “conditional Phrase”.

Ibn Hisham, however, rejects this categorization based on the premise that

conditional sentences are a specific case of verbal sentences (see Peled, ibid: 177).

2.3. Arabic Verbless sentences in the Contemporary Theory of Grammar

Languages like English always have an overt spell-out of the verb. Tense in

these languages have [+D] and [+V] features; the former determining the

relationship between the tense and the subject [EPP], and the latter determining its

relationship with the verb (Chomsky, 1995). Sentences in such languages are not

problematic in that the principle of feature-checking can account for the order of

constituents. The [+D] feature causes the subject to raise from the spec-VP to spec-

TP in order to check the [+D] feature; the [+V] causes auxiliaries and light verbs to

move to the T-head position (Benmamoun, 2000).

Arabic allows of structures where there is no phonological spell-out of the

verb/copula. In fact, Arabic verbless sentences have been analyzed in the literature

17
in different ways2; the most paramount of which are Bakir (1980), Fehri (1993),

Mouchaweh (1986, cited in Benmamoun, 2000), Benmamoun (2000 and 2008) and

Jelinek (1981). One analysis posits that verbless sentences contain a verbal copula

that is deleted (Bakir, ibid) or phonologically null (Fehri, 1993), another analysis

regards verbless sentences as small clauses with no functional projection; i.e. matrix

small clauses (Mouchaweh, ibid, cited in Benmamoun, 2000). A more recent

analysis views verbless sentences as having functional tense projection but no

copula. In other words, verbless sentences contain a functional tense projection that

is marked for present tense only but no verbal copula (Jelinek, 1981; Hazout, 1995;

Benmamoun, 2000; Al-balushi, 2012).

The following section presents the analyses of Arabic verbless sentences as

follows:

 Verbless sentences are small clauses with no functional projection

(Mouchaweh, 1986, cited in Benmamoun, 2000).

 Verbless sentences contain an erased copula that is null in present (Fehri,

1993).

2
The works of Bakir (1980), Jelinek (1981), Fehri (1982), Eid (1983), Mouchaweh (1986), Fassi
Fehri (1993) and Bahloul (1994) are recommended for more details on the structure of Arabic
verbless clauses. Also, for cross-linguistic investigation see the works of Berman and Grosu
(1976), Doron (1986), Rapapport (1987) and Shlonsky (1997) on Hebrew.
18
  Verbless sentences contain a functional tense projection but no copula

(Jelinek, 1981; Benmamoun, 2000).

2.3.1. Verbless sentences are small clauses with no functional projection

Mouchaweh (1986: cited in Benmamoun, 2000: 39) argues that Arabic verbless

sentences are matrix small clauses in that they do not contain any functional

projection. The same analysis is applied to verbless sentences in Hebrew by

Rapaport (1987). This type of sentences contains only a lexical projection with no

functional projection above it. This lexical projection contains a subject and a non-

verbal predicate which can be an AdjP (adjectival Phrase) (5), NP (Noun Phrase)

(6), PP (Prepositional Phrase) (7) or, not mentioned in Benmamoun (2000), AdvP

(Adverbial Phrase) (8).

5. al-bayt-u waasiʕ-un

the-house-Nom large-Nom

The house is large

6. Aħmed-u muʔallim-un

Ahmed-Nom teacher-Nom

Ahmed is a teacher

7. Aħmed-u fi ddaar-i

Ahmed-Nom in house-Gen

19
 Ahmed is in the house

8. Aħmed-u hunaa

Ahmed-Nom here

Ahmed is here

In this analysis, sentences of this type (sometimes termed predicational

sentences) can be introduced as in the following diagram:

9.

According to this analysis, there is no functional category to hold the tense

head. This means that there is a predication relationship between constituents

without any functional head. In this regard, verbless sentences are adjunction

structures having a subject in a specifier position adjoined to a maximal projection

of a lexical head (Moro, 1995; Rothstein, 1995). Benmamoun (2000) considers this

analysis to be comparable to structures such as “I saw [John sad]”. In fact, there are

a number of problems that arise with this analysis:

 Verbless sentences can be coordinated with a CP. For example:

10.al-baytu waasiʕun wa laakin hal sayakuunu dʒamiila(n)?

20
 the-house-Nom large-Nom and but whether will-be.3PMS beautiful-Acc

The house is large but will it be beautiful?

According to the Coordination Condition (Radford, 2009: 59):

Co-ordination Condition

Only constituents of the same category can be coordinated.

The first part of (10) is a verbless sentence while the second part (in bold) is

a CP introduced with the interrogative complementiser hal “whether”. It follows

from (10) and the coordination condition that verbless sentences are CP.

 Verbless sentences can contain a temporal adverb, for example:

11.al-dʒaww-u dʒamiil-un al-yawm(a)

the-weather-Nom nice-Nom today

The weather is nice today

According to the anchoring rule (Harper and Charniak, 1986: 5):

If the time period of the event associated with the first verb of a sentence

can overlap the time period associated with an adverb, then the adverb

can modify that event and can potentially modify the other events in the

sentence (based on the overlap rule). The utterance event cannot be

modified using the anchoring rule.

21
 This means that temporal adverbs have to be anchored by the tense of the same

clause (Eisele, 1988; de Saussure, 2007). The fact that (12) and (13) are

ungrammatical is strongly suggestive of the existence of a tense head that anchors

the temporal adverb:

12.*al-dʒaww-u dʒamiil-un Ɂams

the-weather-Nom nice-Nom yesterday

The weather is nice yetserday

13.*al-dʒaww-u dʒamiil-un ɣad-a(n)

the-weather-Nom nice-Nom tomorrow

The weather is nice tomorrow

It is clear now that verbless sentences are not mere lexical projections of the

lexical head but rather fully-fledged structures with a CP-layer and a functional tense

projection.

2.3.2. Verbless sentences contain an erased copula that is null in present

The early works on Arabic verbless sentences all agreed that there is no verb

in this type of sentences and, hence in contemporary terminology, no functional verb

projection (VP)3. Such a claim is challenged by some contemporary linguists such

as Bakir (1980), Fehri (1993) and Bahloul (1993) who all argue that there is a copula

3
See Sibawayh, 796 [1977]; Ibn Jinni, 1010 [1993]; Abo Hian, 1377 [2001] and Ibn Hesham
1590 [1994])
22
in verbless sentences but it is null/ deleted. The latter draws on cross-linguistic

analyses of Hebrew and Russian by Falk (2004) and Babby (1981).

Fehri (1993) argues that the copula is present in the syntactic representation of

seemingly verbless sentences, yet it fails to lexicalize because it carries unmarked

Tense Feature [-past]. On the other hand, the copula has [+past] feature in verbal

sentences, which forces it to lexicalize. This suggests that there is a functional head

for the tense projection. In the same vein, Bakir (1980) suggests that the noun phrase

preceding the copula is the topic of the projection and the subject is a deleted

pronominal constituent occurring after the copula. For a sentence such as (5), he

proposes the following analysis:

14.

The diagram implies that there are two processes of deletion that apply to verbless

structures; the pronominal subject deletion and the copular deletion. One reason for

23
the pronominal subject deletion is that it is “co-referential with the topic-NP” (Bakir,

ibid: 176).

One reason for considering the pre-copular phrase as a topic rather than a subject

comes from wh-questions. According to Bakir (ibid), the fact that the pre-copular

phrase appears on the left of wh-word and not on the right proves that it is a topic:

15.al-djaww-u kayfa yakuun-u fi ʃʃitaaʔ?

the-weather-Nom how is in the-winter-Gen

The weather, how is it in winter?

Fehri’s analysis differs from Bakir’s in two aspects: first, it assumes that there is

a functional projection to hold tense; second, the copula is null rather than deleted.

He proposes “aspectuo-temporal” and modal constraints for copula nullification:

Spell out the copula as kwn when Mood, Aspect, and/or Tenses are

specified, otherwise spell it out as zero (p156).

Benmamoun (2000) attributes this analysis to mere quest for analytical

consistency. However, such a view is theoretically beneficial in that it provides a

unitary characterization for structures with copular and verbal predicates.

The anchoring rule is a strong evidence that there is a functional head to hold the

tense in verbless sentences. Examine the following:

24
 16.Ɂaṭṭifl-u katab-a ddars-a Ɂamsi

the-boy-Nom Wrote.3PSM the-lesson-Acc yesterday

The boy wrote the lesson

17.*Ɂaṭṭifl-u Katab-a ddars-a ɣada(n)

the-boy-Nom Wrote.3PSM the-lesson-Acc yesterday

The boy wrote the lesson tomorrow

The temporal adverbs“Ɂamsi” and “ɣada(n)” in (16) and (17) must have the

same time reference as the verb in the same clause. Examine the following:

18.*al-dʒaww-u dʒamiil-un Ɂams/ ɣada(n)

the-weather-Nom nice-Nom yesterday/tommorw

The weather is nice yesterday/tomorrow

19.al-dʒaww-u dʒamiil-un ɁalɁaan

the-weather-Nom nice-Nom now

The weather is nice now

The fact that (18) is ungrammatical and (19) is grammatical suggests that the

temporal adverbs are anchored by a functional element with an appropriate time

reference. Note that when the past copula “kaana” is added to (18); the structure

becomes grammatical:

20.kaan-a al-dʒaww-u dʒamiil-an Ɂams

was the-weather-Nom nice-acc yesterday

25
 The weather was nice yesterday

The lexicalized past copula and the temporal adverb in (20) are compatible in

time reference, and the clause is thus grammatical.

The analyses of Bakir (1980) and Fehri (1993) fall short with respect to two

main issues:

First, the analysis considering that there is a null copula in verbless sentences

does not account for the case assignment asymmetry in that an overt copula assigns

an accusative case to the predicate while the predicate of a null copula is assigned a

nominative case. Dechaine (1993) points to the fact that it is not clear why a null

copula and an overt one assign two different cases. Al-Liheibi (1999) points that null

and elliptical elements retain their function, and their effect on other constituents

remains clear. Second, Bakir and Fehri’s analyses (1980; 1993 respectively) do not

offer clear specifications under which the deletion of the copula is obligatory or

optional4.

4
for more details on the conceptual argumentation against the deleted copula hypothesis See
Moutawakil (1987) and Benmammoun (2000) and Shlonsky (1997).
26
 2.3.3. Verbless sentences contain a functional tense projection but no copula

This stance of analysis is adopted by Jelinek (1981) and developed by

Benmamoun (2000, 2008). Jelinek (1981: 47) argues that there is an auxiliary phrase

node specified for the present tense feature as follows:

21.

Aoun et al. (2010) agree with Benmamoun (2000) in that verbless sentences

contain “no verbal copulas and no element carrying tense” (p4). What is new about

Benmamoun’s analysis (2000) is that, while the previous analyses discuss whether

or not there is a copula in verbless sentences, his analysis attributes differences to

fundamental differences between present and past tense. In other words, verbs in

present and past tenses are not in the same position. He argues that, even in verbal

sentences, the verb is in a position lower than negation in present tense and higher

than negation in past tense. That the verb is in a high position in past is what forces

the presence of a verbal copula in past tense.

According to Benmamoun (2000), having [+V] feature, the past tense has to

be paired with a verbal element causing the verb to be in a higher position in verbal

sentences and forcing the presence of a copula in verbless sentence. Since, he argues,

27
present tense does not have [+V] feature, it remains in a position lower than negation

in verbal sentences and has no copula in verbless sentences. Aoun et.al (2010: 4)

carry out in the same vein arguing that: “there is no verbal copula and no element

carrying tense” in verbless sentences and “a tense projection may not require the

projection of a VP”.

Benmamoun’s account proposes a theory of categorical features of tense in

that the heads of the tense projection in present, past and future have different

categorical specifications:

 Present [deictic] [+D]

 Past [+D, +V]

 Future [+D, +V]

Jelinek’s and Benmamoun’s analyses (1993; 2000 resectively), however

different, share the contention that there is neither verbal nor copular constituents in

Arabic verbless sentences. However, reasons for the categorical features differences

between tenses are far-fetched and not fully developed in the analysis. Moreover,

for economy purposes, it is more yielding to regard verbal and verbless sentences as

having the same D-structure.

28
 2.4. Summary

Arabic verbless sentences is a central issue in the study of Arabic syntax albeit

no unequivocal framework of analysis has been proposed as of yet. The scrutinizing

of literature reveals a main dispute and a sub-dispute. The main dispute is whether

Arabic verbless sentences are matrix clauses (Mouchaweh, 1986, cited in

Benmamoun, 2000; Rapaport, 1987) or fully-fledged clauses with functional

categories (Bakir, 1980; Jelinek, 1981; Fehri, 1993; Bahloul, 1993; Benmamoun,

2000; Al-balushi, 2012 among others). This dispute is resolved in that there is a good

deal of empirical evidence in favor of analyzing verbless sentences as clauses with

functional categories. The second dispute is whether these functional categories

include a tense head and a verbal projection (Bakir, 1980; Fehri, 1993; Bahloul,

1993) or a tense head specified for present tense but no verbal copula (Benmamoun,

2000).

The only prominent shortcoming of Fehri’s analysis (1993) is that it did not

account for case assignment asymmetries and did not thoroughly explain the

conditions of copula nullification. Given that, any analysis that justifies case-

assignment asymmetries and provides well-demarcated specifications for the

condition of copula deletion is worthy of adoption.

29
 Chapter Three

A Little-Known Grammaticalization Path

“The World Lexicon of Grammaticalization” by Heine and Kuteva (2002)

provides a thorough inventory of grammaticalization paths and the sources and

developments of grammatical forms. The inventory discusses more than 400

processes of grammaticalization from 500 languages. However, it does not discuss

the grammaticalization path, mentioned in Creissels (2015: 11), which leads the verb

“see” to an ostensive predicator and/or a copula as follows:

1. SEE > OSTENSIVE MARKER > COPULA (> TAM MARKER)

Creissels discusses such a path of grammaticalization with regard to Mande

language family5. On the other end of the spectrum, Hengeveld (2011) and Pustet

(2003), among others, discuss the processes of copularization without

acknowledging the verb “see” as a prospective source of copulas.

In light of the above, this chapter discusses the copularization of the verb “see”

in the Maghrebi dialects with regard to the path of grammaticalization presented by

Creissels (2015) within the theory of grammaticalization in Heine and Kuteva

(2002). The chapter attempts to identify the “raa” in the Maghrebi dialects as a

primarily imperative form of the verb “raɁaa” “see” in MSA, which underwent a

5
For details on the internal classification of Mande languages, see Vydrin (2009).
30
process of grammaticalization to become an ostensive predicator, a copula and an

aspect marker. It is noteworthy that the use of the form “raɁaa” in the discussion

does not imply the past time reference but rather the infinitive form.

3.1. Grammaticalization Theory

Grammaticalization involves the change of constituents from the lexical

category to the grammatical category or, at times, from one grammatical form6 to an

even more grammatical one (Heine and Kuteva, 2002; Lehmann, 2002). The

grammaticalization theory is primarily concerned with the origins and development

paths of grammatical forms. The process of grammaticalization involves four main

mechanisms (Heine and Kuteva, 2002):

a. Desemanticization (semantic bleaching): grammaticalized words often

undergo a loss in their semantic content; they sometimes become

semantically void.

b. Extension (context generalization): grammaticalized words are used in

different contexts.

c. Decategorialization: grammaticalized words lose their syntactic freedom

and some of their morphological properties.

6
The term “grammatical forms” or “grams” generally corresponds to what syntacticians refer to
as “functional categories”.
31
 d. Erosion (phonological reduction): grammaticalized words undergo a

loss in their phonological substance and their phonological content

changes.

3.2. From Imperatives to Ostensive Predicators

The path of grammaticalization proposed by Creissels (2015) entails that

ostensive predicators in some languages originate from the verb “see”. Cross-

linguistic studies about ostensive predicators support this claim. Examine the

following example from French:

2. Je vois les enfants

I see the.Pl kids

I see the kids

3. Voici les enfants!

look-here.clitic the.Pl kids

Here are the kids!

4. Voilà la question!

see-there the.Fem question

Here is the question!

An interesting piece of trivia is that the words “voici” and “voilà” are the same

as the imperative of the verb “voir” “to see” with the addition of the adverbial deixis

32
“ici” and “là”. This supposition means that “voici” is a grammaticalized clitic of the

structure “vois ici” which literally means “see here!” while “voilà” is the

grammaticalized form of “vois là!” “See there”.

It is clear that the imperative form of the verb “see” in French is a source of

grammaticalization material. However, having a form of the verb “see”

grammaticalized form for the verb “see” into an ostensive predicator does not entail

the continuity of the grammaticalization path to a copula. Be that as it may, French

has a separate copular constituent “être”. Examine the following example:

5. Voilà! Les enfants sont là.

Osten! The.Pl kids are there.

There we go! The kids are there.

The reason that French does not allow the further grammaticalization of the

verb “see” into a copula is that the language already has its own copular

constructions. The mapping of the form “voir” onto the function “copula” is already

existent in French, so language either dispenses with the old form-function pair and

replaces with the new one or blocks the grammaticalization, i.e. forming a duplicate

form-function pair.

33
 3.3. The Verb “See” as a Source of Copulas

As shown earlier, French permits the grammaticalization of the verb “see” only

into an ostensive predicator and does not have a grammaticalized copula from the

verb “voir”. I argue that whether or not a language permits further

grammaticalization of the verb “see” is determined by language need/economy.

Languages that have copular structures, such as French, need not the

grammaticalization of “see” into a copula. However, languages without an overt

copula, such as the Mande languages and Arabic, may allow such a path.

The Mande languages can have non-copular clauses (Creissels, 2015), and,

according to the claim posed above, are expected to allow the grammaticalization of

the verb “see” to a copula. Creissels (ibid: 7) illustrates two cases of the

grammaticalization of “kaa” “see”:

6. Ku ŋaloŋ ka bɛlɛi mu.

1Pl man see house in

We saw a man in the house.

7. I seŋkau ka!

2Sg money OST

Here is your money!

8. Ŋaloŋ ka bɛlɛi mu.

34
 man COP house in

The man is in the house.

The structure in (6) shows the past form of the verb “see” in the Kpelle variety

of Manding. The verb is grammaticalized into an ostensive predicator in (7) and into

a locational copula in (8)7. Interestingly, the Kpelle variety retains the

grammaticalized forms alongside the original source of grammaticalization

(Creissels, 2015).

3.4. The Case of Maghrebi Dialects

Taine-Cheikh (2013) discusses the “raa” in some Maghrebi dialects as a

copula originating from the verb “raɁaa” “see”. The following discussion builds

upon Taine-Cheick’s analysis and argues that the verb “raɁaa” is also

grammaticalized into an ostensive predicator in accordance with the

grammaticalization path proposed by Creissels (2015).

The discussion assumes “raa” not only as an equative copula but also as a

locational copula and an aspect marker. The final part of the discussion tackles some

major differences between the grammaticalization of the verb “see” in Arabic,

French and the Mande languages.

7
For more details on the copular clause in the Mande languages see Westermann (1930).
35
 3.4.1. “raa” as an Ostensive Predicator

As argued above, it is the imperative form of the verb “see” that undergoes

grammaticalization. The imperative form of the verb “raɁaa” in Standard Arabic is

“ri”. It should be noted that the verb “raɁaa” is almost non-existent in the Maghrebi

dialects, and in most modern Arabic dialects, as it is replaced by “ʃaaf” in these

dialects. Some old speakers in the Algerian dialect still use the Arabic cognate such

as in:

9. Riitu lbaariɦ

Saw-him yesterday

I saw him yesterday

The use of ostensive predicators that are cognate with the verb “raɁaa” is very

common in the Maghrebi dialects. Examine the following example:

10. Raahum ddraari

Osten-them the-kids

There are the kids!

It should be noted that the ostensive predicator “raa” cannot occur without a

clitic pronoun even though it can occur without the noun phrase. Moreover, the

pronoun has to be co-referential with the noun phrase it precedes:

11. *Raa ddraari

36
 Osten the kids

There are the kids.

12. “dʒaw ddraari?” “raahum!”

“came.3Pl the-kids?” “Osten.3Pl”

“Did the kids come?” “There they are!”

3.4.2. “raa” as a Copula and an Aspect Marker

The observation of the Maghrebi dialects reveals that, unlike Standard Arabic,

the locational (13) and equative (14) clauses are expressed with an overt copular

predication:

13. Papa raah f lxidma

dad is.3PMS in the-work

Dad is at work

14. Omar raah mriiɖ

Omar is.3PMS ill.M

Omar is ill

Copulas whose origin is the imperative form of the verb “raɁaa” are very

common in these dialects. Moreover, “raa” can be used with clauses with an overt

verbal predication; it functions as an aspect marker.

37
 Aspect markers are constituents that denote the duration of the action, i.e.

whether the action is ongoing on completed (Radford, 2009: 442-443). To that end,

“raa” can function as a marker of perfect aspect, and, thus, a perfect auxiliary (15),

and a marker of progressive aspect, and, thus, a progressive auxiliary (16).

15. lɁustaad raah dʒa

the-teacher.M has.3PMS come.3PMS

The teacher has come

16. lɁustaad raah dʒay

the-teacher.M is.3PMS coming.3PMS

The teacher is coming

It should be noted that the structure in (15) is understood to have a different

aspectual properties than structures without the aspectual auxiliary “raa”. Examine

the following example:

17. lɁustaad dʒa w raaɦ

the-teacher.M came.3PMS and went.3PMS

The teacher came and went

The action in the structure (15) is perceived to be completed, i.e. it has

perfective aspectual properties whereas the structure (17) is not necessarily a

completed action.

38
 Moreover, the aspectual auxiliary “raa” can mark the continuity of the action

when combined with a verb in present tense (18) or gerundive (19)

18. Mohammed raah yuntriini

Mohammed is.3PMS trains.3PMS

Mohammed is training

19. Mohammed rah dʒaay

Mohammed is.3PMS coming.PMS

Mohammed is coming

Similarly, such structures are different from structures without the auxiliary

“raa”, which express habitual action or facts:

20. Muhammed yuntriini

Mohammed trains.3PMS

Mohammed trains.

The previous examples show that the copularization of the verb “raɁaa” in the

Maghrebi dialects fulfills the function of aspect marking and predicative marking.

In both cases, the grammaticalized constituents are empty functional categories with

little, if any, semantic content.

39
 3.5. Remarks on the Grammaticalization of “raɁaa”

The analysis of grammaticalization in French, the Mande languages and

Meghrebi Arabic dialects reveals some key differences. First, ostensive predicators

and copulas in these languages are cognates with the verb “see”. Interestingly,

French and Mande languages retained the original form of the verb in addition to the

grammaticalized forms. Meghrebi dialects, however, dispensed with the verb

“raɁaa” as a lexical element and replaced it with the verb “ʃaaf”.

Second, the ostensive predicator in Meghrebi Arabic is different from that of

French in that it has more morphological freedom. The ostensive predicators “voici”

and “voilà” in French are non-inflectional in that they stay the same regardless of

the complement. However, “raa” in Maghrebi dialects requires a clitic pronoun that

agrees in number and gender with the noun phrase it precedes.

Another key difference is that the grammaticalization of the verb “see” in

some Maghrebi dialects is accompanied with a considerable phonological loss. For

example:

21. Mohammed (r)ahu mliiɦ?

Mohammed is-he ok

Is mohammed ok?

40
 At this juncture, it should be noted that the grammaticalization of the verb

“see” in Arabic is not peculiar to Maghrebi dialects. In fact, it can be observed that

even Standard Arabic has a grammaticalized form of “see”. For example:

22. Yaa turaa hal sayaɁtii Omar?

O be-seen.1PMS whether will-come.3SM Omar

I wonder, will Omar come?

In addition, some Arabic dialects, such as the Jordanian dialect, have structures

that contain a grammaticalized form of the verb “see”. Examine the following:

23. Taraa ldʒaw naar

See.1SM the-whether fire

Indeed, the weather is hot

The difference between (23) and “raa” in Maghrebi dialects is that “taraa” is

fixed in the presence of an overt subject; it has no inflectional freedom and does not

require a resumptive pronoun. However, if the subject in (23) is pronominal in

nature; it can cliticize onto “taraa”. For example:

24. Taraani Ɂurduni

See-me Jordanian

I am Jordanian

41
 The findings in this chapter confirm the path of grammaticalization in

Creissels (2015). Arabic dialects and even Standard Arabic have grammaticalized

forms of the verb “raɁaa”. The Maghrebi dialects adhere to the path of

grammaticalization:

25. SEEimperative > OSTENSIVE MARKER > COPULA (> TAM MARKER)

It is shown in the discussion that even though Maghrebi dialects follow the

same path as French and the Mande languages, the morphological and syntactic

properties of the grammaticalized constituents are not parallel. “raa” in the Maghrebi

dialects is more morphologically and lexically free than are “voici/voila” in French

and “ku” in the Manding language family.

The discussion of “raa” as a constituent of copular constructions cannot be

completed without the discussion of its syntactic structure in sentences with an overt

verbal predication and sentences without it. The following chapter discusses issues

that arise with the postulation of “raa” as a prospective copula and a marker of

Aspect, and whether the clitic-pronoun that attaches to “raa” is the result of

agreement or subject indexation.

42
 Chapter Four

A New Outlook at Arabic Verbless Sentences

As shown in the previous chapters, some structures in Standard Arabic are

characterized by the absence of an overt copula in positive present tense. It has also

been shown that many an analysis dismisses the prospect of a verb projection in

verbless sentences. This chapter outlines the theoretical benefits of characterizing

Arabic verbless sentences as having a functional tense projection and a verbal

category. Moreover, it shows the complications that arise from positing otherwise.

4.1. Functional Categories in Verbless Sentences: A Rationale

There is a solid theoretical validation to assume that Arabic verbless sentences

contain a verbal projection. This goes along the lines of Radford's assumption that

there is "[…] a simple model of clause structure in which complement clauses are

CP+TP+VP structures" (2009: 342). With this in mind, the theoretical and empirical

evidence refuting the VP-missing structures is more compelling.

The fact that verbless sentences can be coordinated with verbal sentences refutes

the analyses adopted by Benmamoun (2000) and Mouchaweh (1986, cited in

Benmamoun, 2000).

43
 1. Al-bayt-u wasiʕ-un wa sa-yakun-u awsaʕ(a)

The house-Nom large-Nom and will be larger-Acc

The house is large and it will be larger

2. Al-bayt-u wasiʕ-un lakin yaħtaj-u ?ila ttarmeem

The house-Nom large-Nom but needs to reconstruction-Gen

The house is large but it needs reconstruction

The sentences in (1) and (2) show that “Al-baytu wasiʕun” can be coordinated

with a VP and a TP. This leads to the fair assumption, which I shall attempt to argue

for, that verbless sentences contain not only a functional tense projection but also a

verbal predicate.

3. Al-bayt-u wasiʕ-un

The-house.Nom large.Nom

The house is big

Seemingly, (3) is verbless in that it contains no overt copula. However, the past

and future equivalents of the sentence raise interesting observations in that they are

equal in meaning, with different time references “tense interpretation”, yet the

copula is overt and the adjective is assigned accusative case in the past and future.

4. Al-bayt-u kana wasiʕ-an

The-house.Nom was large.Acc

The house was large

44
 5. Al-bayt-u sa-yakun-u wasiʕ-an

The house.Nom will be large.Acc

The house will be large

One problem that arises from (4) and (5) is how to account for sentence (3). First,

the sentences are equal in meaning, with different time references, leading to the

logical assumption that they are equal in structure with difference only in the head

T, yet the assumption that (3) contains a null T “yakun” must account for the

nominative case assigned to its complement. One other issue to account for is why

the T-head is nullified in present tense only.

4.2. Conditions for Copula Deletion

It has long been argued that copula deletion in the nominal sentences is

conditioned by the tense/time references. However, data of Arabic show that the case

is more intricate (Obeidat & Ferghal, 1994).

Mood clearly plays a role in determining the spell-out, or the lack thereof, of the

copula. Obeidat & Ferghal (ibid) argue that copula in Arabic obligatorily surfaces in

the Imperative and Subjunctive moods (6-7). Consider the following:

6. Kun radjul-an

Be man-ACC

Be a man

45
 7. Law kaan-a lbayt-u wasiʕ-an laʃtaraytuh-u

If be-past the-house- large-Acc then-bought-I-it

If the house was large; I would have bought it

Sentence (6) and (7) are in Imperative and Subjunctive moods respectively, and

the spell-out of the copula is obligatory. It can be concluded, thus, that only

Indicative mood permits the deletion of the copula. However, it would be by no

means plausible to assume that all Indicative mood sentences allow or disallow the

overt spell-out of the copula. In this view, Obeidat and Ferghal (ibid) lay out a rule

for copula deletion in Arabic:

Arabic copula obligatorily surfaces in the Imperative and Subjunctive mood.

As for the Indicative mood, the copula obligatorily materializes with past and

future Time References, but it is obligatorily deleted with present Time

References when newness and timelessness are the relevant Aspects, and

optionally when habituality/recurrence is the relevant Aspect (p19).

The main issue is why copula is deleted only in these conditions. This can be

accounted for by assuming that the nullification is discourse-bound8. Let us suppose

that the speech discourse tense is understood to be present and positive in Arabic

due to the fact that tenses are simple in Arabic. In other words, let us suppose that

8
See Gergal (2009) for more details on the distinction between sentence-bound and discourse-
bound ellipses.
46
sentences in standard Arabic are by default positive and present9. This means that

if the sentence is neither positive nor in present, these syntactico-semantic changes

have to be realized phonologically. The realization of past or future is through the

overt spell-out of the copula, and the realization of negatives is through an overt

particle such as “laysa”. By the same token, since the sole function of the T-head is

to determine the tense which is already inferred in (3), language economy dictates

that the penurious T-head be phonologically nullified. One support for this

assumption is that the T-head can be given an overt spell-out with little if any

significant change of meaning.

8. Al-bayt-u yakun-u wasiʕ-an

The house-Nom is large-Acc

The house is large

Note that though sentence (8) is grammatically legitimate; it is deemed by native

speakers of Arabic to be redundant (Obeidat & Ferghal, 1994).

What is noteworthy at this juncture is that constituent-deletion that is dictated by

the economy of representation is not uncommon in languages. For one, Schu¨tze

(2004) argues that non-negated clauses contain a positive counterpart to the NEGP.

It is known that English negation is realized with an overt negation marker (not)

9
Although this is merely hypothetical, we assume that since verbs have a default case (Abdul
Sattar, 2012), verbs can have a default tense and polarity status.
47
being the specifier of the negation phrase (NegP) (Radford, 2009). Radford (2009:

171) argues that positive sentences contain a Polarity phrase to configure it as a

positive clause, yet it is null. What causes the positive polarity phrase to be null and

the negative one to be overt is the discourse configuration. Put delicately, English

sentences are understood to be positive by default, so their positive polarity markers

are null. Configuring a sentence otherwise, i.e. negation, requires the polarity marker

to be spelled out. By the same token, Arabic sentences are by default positive and

present. Changing the default properties requires overt constituents. For this reason

polarity changes require an overt negation particle such as “laysa, lam and lan” “not”

while time-reference changes require the overt spell-out of the copula.

Further yet, the Null Copula Hypothesis lends support from the use of temporal

adverbs “yesterday, today…etc.” as they clearly denote tense which concords with

the T-feature held by the T-head.

9. Al-bayt-u wasiʕ-un al-yawm-(a)

The house-Nom large-Nom today

The house is larger today

10.*Al-bayt-u wasiʕ-un ghad-a(n)

The house-Nom large-Nom tomorrow

The house is large tomorrow

48
 11.*Al-bayt-u wasiʕ-un ?ams

The house-Nom large-Nom yesterday

The house is large yesterday

The ungrammaticality of (10) and (11) implies that the structure in (09) has a

tense element that takes present temporal adverbs only. This further supports the

assumption that there is a functional tense projection in (09). Fehri (1993: 87)

suggests that the copula is present but fails to lexicalize because it carries an

unmarked T-feature [-past]. However, the copula in past holds marked [+past] T-

feature which compels it to be given an overt spell-out. I shall expand on this

assumption by assuming that the modal verb10 "qad" renders the T-feature [-past]

marked and, therefore, forces the null copula to be realized.

12.qad yakun-u Al-bayt-u wasiʕ-an

may be-3PSM the house-Nom large

The house may be large

13.*Al-bayt-u qad wasiʕ-an

The house-Nom may large-Acc

The house may large

10
Aoun et al. (2010); Benmamoun (2000); Shlonsky (1997) and Fehri (1993) argue that "qad" is
a modal verb "modality particle".
49
 Note that the order in (12) can change. The copula “yakuun” is obligatorily

spelled in the presence of the modal verb “qad”, which means that once the tense

head is occupied, the copula surfaces on the phonetic level. This lends support to the

assumption that there is a copula that is null in certain conditions.

The second issue arising with the Null Copula Hypothesis (Fehri, 1993) is the

case assignment asymmetries as the complement of null copula is assigned

nominative case while that of the overt copula is assigned accusative case. Before

tackling the agreement asymmetry issues, there is need to better establish a unitary

framework for analyzing Arabic sentences. Indeed, prior to proposing a framework

that unifies Arabic verbal and nominal sentences, let us first account for the

differences, or the lack thereof, between VSO and VSO sentence.

4.3. SVO in Arabic: Canonical Order or Topicalisation

It has been shown earlier that Standard Arabic alongside modern Arabic dialects

permit structures with SVO order11. Consider the following:

14.ʔal-ʔawlaad-u naamuu (MSA)

The-kids-Nom slept-3PlM

The kids slept

15.Lwlaad rәgduu (Algerian Arabic)

11
“Arabic dialects are more tolerant of SVO structures” (Soltan, 2011: 249).
50
 The-kids slept-3PlM

The kids slept

The issue is whether or not the order in (14) is a canonical order, which

embraces the assumption that T-head in Arabic has EPP feature that resulted in the

movement of the subject "ʔal-ʔawlaadu" to occupy the spec-T position and check

the EPP feature. The alternative analysis is that SVO in Arabic is the result of a

process called Topicalisation, i.e., a process marking the preposed element as the

topic of the sentence (Radford, 2009). The analysis discussed in Bakir (1980) means

that preverbal (pro)nominals are not syntactic subjects per se. However, this analysis

cannot be accepted unless we assume that the T-constituent in Arabic has no EPP

feature12.

There is a number of complications that arise from the analysis of preverbal

subjects as topicalised constituents. First, the assumption in the Topicalisation

analysis is that the canonical and unmarked order of Arabic sentences is VSO and

the SVO is an instance of Left Dislocation. This implies that (14) is originally as

follows:

12
The main assumption throughout this thesis is that the tense head in Standard Arabic has EPP feature.

51
 16.

There are two movement processes in (16), one that leads the verb to move to

the T-head position. This movement is triggered by the strong tense affix occupying

the T position. The second movement causes the Determiner Phrase to move from

the complement-V position to the Spec-Top position13. This movement is triggered

by the Edge Feature [EF] carried by the null Topic head (Radford, 2009: 327).

The first issue that arises from (16) is that the marked order VSO allows for

partial agreement between the preverbal subject and the verb. Such an analysis does

not account for the agreement loss in (17):

17.Naam-a ʔal-ʔawlaad-u

Slept-3PSM the-kids-Nom

The kids slept

13
See Rizzi (1997; 2001; 2004) for more details on the split CP hypothesis.
52
 Another reason to argue against the topicalisation analysis comes from the

following examples14:

18.ʔattufaaħat-u ʔakalahaa lwalad(u)

The-apple ate-it the kid-Nom

The kid ate the apple

19.Zayd-un yaʤib-u ʕalayhi rraħiil-(u)

Zayd-Nom has to-him the-leaving

Zayd has to leave

20.Zayd-un yaʤib-u ʕala lwalad-i llaðii maʕahu rraħiil-(u)

Zayd-Nom has to the-boy-Gen who with-him the-leaving-Nom

The boy who is with Zayd has to leave

The examples (18), (19) and (20) are cases of topicalisation. In these cases,

there is a resumptive pronoun in the extraction domain, even if the extraction cite is

a prepositional phrase (19) or complementiser phrase (20). Structures with no

resumption are ungrammatical. With that in mind, the SV structures have no

resumptive pronouns, which is overlooked in the analysis discussed by Bakir (1980)

overlooks that.

14
The examples (72) and (73) are taken from Soltan (2011: 253).
53
 The analysis of preverbal subjects as instances of topicalisation is problematic

when viewing other types of topicalisation. Apart from subject topicalization, Arabic

has another type of topicalization, which results in OVS word order. This type is

referred to as object-topicalization (Bakir, 1980; Farghal, 1992; Ryding, 2005;

Ford, 2013). Consider the following example:

21. ʔattufaaħat-u ʔakalahaa lwalad-(u)

The-apple ate-it the kid-Nom

The kid ate the apple

The preverbal nominal “ʔattufaaħatu” originates within the VP and raises to

the spec-Top. This means that the spec-Top occupied, and the Edge Feature of the

Topic head is satisfied and deleted. Consequently, no other constituents are triggered

to move to the spec-Top position because the “EF is always deleted when satisfied”

(Chomsky, 2006: 11). Now consider the following example:

22.ʔattufaaħat-u, lwalad-u ʔakalahaa

The-apple ate-it the kid-Nom

The apple, the kid ate it

The example (22) shows two preverbal nominals. It follows from (21) that the

nominal “ʔattufaaħatu” is in spec-Top position, and no other constituents can adjoin

to the spec-Top. This means that the nominal “lwaladu” is in a position lower than

54
the spec-Top. Such structures of OSV are grammatically legitimate although rarely

used by Arabic native speakers for being redundant.

The examples of OSV show that the topicalisation resulted in the preposing

of the object and that the preverbal subject is not the result of topicalisation.

Another opposition to the topicalization analysis (Bakir, 1980) comes from

the following example:

23.Al-yawm-a lħaqq-u sayusmaʕ-(u) al-yawma

The-today-Acc the-truth-Nom will-be-heard the-today

Today, the truth will be heard

Let us agree with the analysis discussed by Bakir (1980) that the preverbal

subjects in (22) and (23) are cases of topicalization. This means that the nominal

“ʔattufaaħatu” in (22) and the adverbial “al-yawma” in (23) are in positions higher

than the spec-Top. However, Radford (2009) posits that there is only one projection

higher than the Topic Phrase, namely the Force Phrase ForceP. This means that

unless we accept that the nominal and the adverbial above are in spec-Force position;

the analysis of Bakir (1980) is unrepresentative of such structures.

The topicalization of adverbials is not documented in the syntax of Arabic, which

means that the adverbial in (23) is unlikely to be in the spec-Top position. This is

supported by the following example:

55
 24.ʔattufaaħat-u, lbaariħa-ta lwalad-u ʔakalahaa

The-apple, last-night-Acc the kid-Nom ate-it

The apple, the kid ate it last night

As posited earlier, the spec-Top position is occupied by the topicalised object

“ʔattufaaħatu”, which means that the adverbial “lbaariħata” in (24) is in a projection

lower than the TopP. In line with the analysis of Radford (2009), we assume that the

adverbial is between the Spec-T and the spec-Top, i.e., in the spec-Foc position.

The totality of this analysis suggests that preverbal subjects are not

tropicalized constituents, contra the analysis presented by Bakir (1980), but rather

preverbal subjects in the spec-T position that have originated VP-internally and

raised to satisfy the EPP feature of the tense head.

We noted earlier that the preverbal subject can be analysed as instances of

topicalisation only if we assume that the adverbial “al-yawma” in (23) and the

nominal “ʔattufaaħatu” in (22) are analysed as being in spec-Force position.

Accordingly, we would expect such constituent to be positioned on the left-edge of

the force head as follows:

56
 25.

26.

The tree diagrams in (25) and (26) show the outcome of assuming that the

preverbal subject is an instance of topicalization. However, such analysis is

problematic in two ways: first, the example in (24) shows that there are two

constituents before the preverbal subject; the adverbial and the nominal. Knowing

that the EF of the force head is deleted once satisfied means that the Force head can

57
attract no specifier once the adverbial adjoins to the spec-Force position. This leaves

no room in the derivation for the nominal “ʔattufaaħatu” within the CP in that there

is no higher head than the ForceP. Second, Arabic allows testing the analysis in (25)

and (26) as it has overt force complementisers in the matrix clause, i.e., “ʔinna” and

the embedded clause, i.e., “ʔanna” (Jalabneh & Abdellatif, 2014). Given the analysis

in the tree diagram above, we expect the adverbial in (23) and the nominal in (22) to

be positioned on the left-edge of “ʔinna” as follows:

27.*Al-yawm-a ʔinna lħaqq-a sayusmaʕ-(u) al-yawma

The-today-Acc the-truth-Nom will-be-heard the-today

Today, the truth will be heard

The structure in (27) is ungrammatical, which suggests that the adverbial in

(23) cannot be in a spec-Force position; instead, I suppose that it is in a spec-Foc

position. This means that the tree diagrams (25) and (26) are not representative of

the structures (22) and (23) respectively.

One might argue that the adverbial can occupy the spec-Force position in

sentences such as the following:

28.Ħasibt-u al-yawm-a ʔanna lħaqq-a sayusmaʕ

Ø-pro thought.1PS the-today-Acc that the-truth-Acc will-be-heard

I thought today that the truth will be heard

58
 The example in (28) shows that the adverbial is positioned on the left-edge of

the embedded-clause-specific complementiser “ʔanna”, which suggests that it is in

the spec-Force position. However, there are several reasons to refute that. For one,

the adverbial in (28) is CP-external (not within the domain of the embedded clause

complementiser). Evidence for that is brought from the semantic-syntactic mapping

of constituents.

Ross (1967) states that relative clauses are islands. Constituents within the

domain of the complementiser are impervious to the grammatical operations.

Moreover, Chomsky (1999), in his Impenetrability Condition, states that the

elements within the complementiser domain are impenetrable to other heads. I shall

expand by assuming that constituents within the C-domain are also impenetrable to

the process "modify". In other words, the tense head within the domain of the CP

cannot be anchored by adverbials that are CP-external. Examine the following

examples:

29.ħasibt-u lbaariħat-a anna lħaqq-a sayusmaʕ ɣadan

pro thought.1PS that the-truth-Acc will-be-heard tomorrow-Acc

I thought yesterday that the truth will be heard tomorrow

The example in (29) shows that the tense anchoring, which is a process of

time reference agreement, is obligatory between elements within the same CP;

otherwise the structure becomes ungrammatical as in the following example:

59
 30.*ħasibt-u ɣadan anna lħaqq-a sayusmaʕ lbaariħat-a

pro thought.1PS that the-truth-Acc will-be-heard the-today-Acc

I thought yesterday that the truth will be heard tomorrow

The structure in (30) is ungrammatical because the tense in the embedded and

matrix clauses is not appropriately anchored by the adverbial within the respective

CP’s. Now examine a similar example to (28):

31.ħasibt-u al-yawm-a/lbaariħat-a ʔanna lħaqq-a sayusmaʕ

Ø-pro thought.1PS today-Acc/last-night-Acc that the-truth-Acc will-be-heard

I thought today that the truth will be heard

The example (31) shows that the sentence is grammatical and that the pre-

complementiser adverbials are not anchored by the verb within the embedded clause.

This amounts to saying that the adverbial in (28) is not in spec-Force but rather

pertinent to the matrix clause. In other words, the adverbial “al-yama/lbaariħata” are

constituents within the domain of the main clause CP.

The entirety of the previous discussion suggests that the preverbal subjects are

not topicalised subjects but rather originate in a position that is lower than the Topic

Phrase and the Focus Phrase. Soltan (2006) has a different view about preverbal

subjects, his analysis is referred to as “Base-Generation Analysis”. He argues that

the VS and SV orders in Arabic are not related to whether or not there is a subject

raising to spec-T but rather are “a consequence of two different base-generated

60
structural representations” (Soltan, 2006: 239). This means that preverbal subjects

are base-generated in the spec-T position rather than raising via movement from the

VP. In order to overcome the agreement complications that arise from his analysis,

he posits that there is a null pro in the spec-v, which has a full agreement with the

tense head in gender, person and number. This means that the preverbal subject plays

no role in subject-verb agreement. On the other hand, the post-verbal subjects are

base-generated in the spec-v and remains in situ, as shown in the labelled bracketing

below (Soltan, 2006: 256):

32.

33.

However, there are several complications that arise from this analysis. First,

his main assumption is that the tense head in SV structures (32) has EPP feature

while it does not in VS structures (33). From a minimalistic point of view, this is

problematic in that the EPP is a parametric subject-related repulsion feature

(Biberauer et al., 2010). Assuming that some tense heads have EPP features while

others do not suggests that there is a parametric variation in the same language.

However, assuming that the same head can have EPP feature in some structures and

not in others requires an account for the patterns of this parametric variation. Another
61
issue with Soltan’s analysis (2006) is the fact that the preverbal subjects are assigned

nominative case, which means that they, contrary to his assumption, take part in the

agreement process. Moreover, Soltan (2006) assumes that there is a full agreement

between the tense head and the null pro in the spec-v position. He also claims that

post-verbal subjects are in the spec-v position. However, this claim is contradictory

in that post-verbal subject-verb agreement in Standard Arabic is partial

(AlShammiry, 2016). Put delicately, Soltan does not explain why the T and the spec-

v have full agreement in SV but partial agreement in VS.

Contra Soltan (2006), I assume that preverbal and post-verbal subjects are not

the result of two different base-generated structural representations but rather the

result of two different movement processes as will be explained in the subsequent

chapter. In fact, I venture to posit that the lexical subject in both VS and SV

structures is base-generated VP-internally, and that the tense head in Arabic has EPP

feature in both structures. This goes in accordance with Radford's supposition that

"every T constituent has EPP feature" (2009: 240), and that "every T constituent

must be extended into a TP projection which has a specifier" (ibid: 455). The

preverbal subject in SV structures raises to the Spec-T position to satisfy the EPP

feature of the tense head.

There are several empirical and theoretical considerations to refute the

analysis of Soltan (2006) about the base-generation of preverbal subjects in the spec-
62
T position. In fact the analysis outlines in this thesis is built on the premise that

subjects in SV and VS originate within the VP. Consider the following example:

34. a. ʔal-walad-u yaʔkul-u

the boy is eating

b. ʔal-bint-u taʔkul-u

the girl is eating

c. ʔal-ʔawlaad-u yaʔkuluu-na

the boys are eating

d. ʔal-banaat-u yaʔkulna

The girls are eating

The previous examples show that the preverbal subject-verb agreement is full

in number, gender and person. Let us test the analysis of Soltan (2006) against the

agreement model of agreement proposed by Chomsky (2000 et seq.)15. Pursuant to

Soltan’s analysis (2006), the structure (34.d) is derived as follows:

15
Chomsky’s probe-goal model of agreement is revisited in Chomsky (2001; 2004; 2007; 2008).
63
 35.

The examples in (34) show that the overt agreement is between the lexical

subject and the verb in the T position. This leads to the logical assumption that the

probe-goal relationship is between these two rather than between the tense head and

the alleged pro. If we are to assume that the agreement is between the preverbal

subject and the verb; we need to assign one a “probe” status and the other a “goal”

status. This leads to one of these following three possibilities:

a. The preverbal subject is the probe which locates the verb as it is suitable

goal and, hence, agrees with it.

b. The verb in the tense position is the probe which locates the preverbal DP

as its goal, given that it is not within its domain of c-command.

64
 c. The verb in the tense position is the probe locating the preverbal DP as it

is suitable goal as it is within its domain of c-command, i.e., the DP

originates within the VP.

The first possibility means that the probe of the agreement process does not

necessarily have to be the head of the phrase. However, this is theoretically loose

from a minimalistic point of view due to the fact that all agreement probes are

typically the heads and the minimal projections of given phrases (Radford, 2009).

This means that the probe of the subject-verb agreement is the head, i.e., the verb.

The second possibility satisfies the condition that the probe is the head. However,

the probe does not c-command the goal. This is against the spirit of minimalism,

which calls for reducing the theoretical apparatus involved in the syntactic derivation

to a minimum. This implies that the second possibility is invalid in that it envelops

a syntactic operation that is not subject to the relation c-command. Therefore, it is

more theoretically yielding to assume that all syntactic operations are subject to the

relation c-command. The third possibility meets the two requirements and is closer

to the syntax of minimalism. First, it acknowledges that the head of the projection is

the typical probe. Second, it conforms to the c-command relation as the goal “the

DP” is c-commanded by the probe “the verb”. Put otherwise, the final possibility is

the closest to the spirit of minimalism in that it presents the agreement relationship

as “a process involving a head and a constituent within its domain of c-command”.

65
 The model of probe-goal agreement brings solid evidence to refute the base-

generation analysis of Soltan (2006). Other pieces of evidence to refute such analysis

come from the idiomatic expressions. Idioms are phrases or sentences with

idiosyncratic “inherent” meaning that does not necessarily equate the sum of its

parts. Radford (2009) argues that the idioms are a unitary constituent. This means

that the constituents that are part of the idiom pertain to the same projection. Now

consider the following example from Jordanian Arabic:

36.ʔakal hawa

pro-ate air

He is in dire straits

The example in (36) shows an idiom that is a unitary constituent as it is the

intermediate projection of the verb phrase, as follows:

37. [V’ [V ʔakal] [DP hawa]

Having idiosyncratic interpretation means that the unitary constituents forming

the idiom are irreplaceable. This is supported by the fact that the tense and the subject

of the previous idiom can change (38.a) whereas the verb (38.b) or the complement

of the verb (38.c) cannot.

38. a. Raħ nuukul hawa

We will be in dire straits

66
 b. ?? ʔaxað hawa 16

pro took air

he has straits

c. ?? ʔakal ɣabara

pro ate dust

he had dust in his mouth

The example (38.a) shows that the tense and the subject of the idiom can

change without affecting the idiosyncratic meaning of the idiom. This is because the

two are not within the V’. However, changing the constituents within the V’ results

in changing the inherent interpretation of the expression as in (38.b) and (38.c). The

reason why the idiosyncratic interpretation of the idiom is changed in (38.b) and

(38.c) is because the changed constituents are V’-internal. For explanatory purposes,

let us use the term “Idiom Constraint” to refer to the non-changeability of the

constituents of the idiom. This means that the examples (38.b) and (38.c) violate the

Idiom Constraint. Now consider the idiom from MSA:

39.ʔixtalaţ-a lħaabil-u binnaabil

pro mixed the-roper-Nom with-the-archer-Gen

All hell broke loose

16
The question mark means that the structure is grammatically legitimate, yet it does not represent the
inherent interpretation of the idiom.

67
 The example in (39) is an idiom in MSA, which forms a unitary constituent,

i.e., all of its elements are within the same projection. According to the Idiom

Constraint outlined above, we would expect idiom-internal constituent to be

unchangeable and, therefore, within the same projection. Now consider the

following examples:

40.ʔal-ħaabil-u sajaxtaliţ-u binnaabil

the-roper-Nom will-mix with-the-archer-Gen

All hell will break loose

41.?? ʔal-ħaabilu sajaltabisu binnaabil

the-roper-Nom will-be-confused with-the-archer-Gen

All hell will break loose

42.?? ʔal-ħaabil-u sajaxtaliţ-u bi-mustaxdim-i nnibl

the-roper-Nom will-mingled with-user-Gen the-bow-Gen

All hell will break loose

43.?? Mustaxdim-u lħabl-i sajaxtaliţ-u binnaabil

user-Nom the-rope-Gen will-mingled with-the-archer-Gen

All hell will break loose

The example in (40) shows that the tense head can be changed in the previous

idiom without changing the idiosyncratic meaning of the idiom. This means that the

Idiom Constraint is not violated suggesting that the tense is idiom-external and, thus,

68
not within the same projection. The example in (41) shows that the inherent

interpretation of the idiom is changed, which means that the Idiom Constraint is

violated when the verb is changed. This implies that the verb is an idiom-internal

constituent. Moreover, the examples in (42) and (43) show that the complement of

the verb and the lexical subjects are idiom-internal in that the changing of either

results in changing the idiosyncratic meaning of the idiom and, thus, violating the

Idiom Constraint.

It follows from the previous examples that the verb, the lexical subject and

the complement of the verb are within the same projection. According to Soltan’s

analysis (2006) the TP is the smallest projection that includes all the three. However,

the example in (40) shows that the tense head is idiom-external. This means that the

analysis of the structure in (40) must have the subject, the verb and the complement

in one projection while excluding the tense head. Such a requirement is not attained

under the analysis of Soltan (2006). This gives support to the assumption made

above that the subject in SV structures originates VP-internally. Such an assumption

means that the previous idiom is a VP with the lexical subject being the specifier of

the verb. This does not violate the Idiom Constraint as the subject, the verb and the

complement are all VP-internal and unchangeable.

On the other end of the spectrum, VS sentences have no overt preverbal

subject. However, unless the EPP feature is checked by the left-dislocation of a DP;
69
the derivation crashes at LF (Svenious, 2001). Mohammad (1990; 2000) proposes a

solution for the EPP feature in VS structures. In his analysis “the Null Expletive

Hypothesis”, the tense head in VS structures has EPP feature that is satisfied by a

null expletive that is inert at the Logical-From level by dint of having no intrinsic

meaning and empty at the Phonetic-Form level by dint of having no overt spell-out.

A crucial issue that is overlooked in the Null Expletive Analysis (Mohammed, 1990;

2000) is where the base-generation of the null expletive is.

4.4. The Null Expletive: External or Internal Merge

The analysis of Mohammed (1990; 2000) suggests that there is a null expletive

pronoun in the spec-T position. He argues that the verb agrees with the “left-most

conjunct” (2000: 136). This goes in line with Chomsky’s assumption that the raising

of expletives is an instance of external merge rather than internal merge (1999;

2001), i.e. movement, where the expletive pronouns are merged directly with the T'

to occupy the spec-T position necessary to satisfy (delete) the EPP feature of the

tense head. However, such analysis is problematic with regard to agreement

structures. Consider the following example:

44. Naama lʔawlaad-u

pro slep-3PSM the-kids-Nom

The kids slept

70
 The example shows that there is a partial agreement between the verb and the

post-verbal subject in VS structures. The Null Expletive Hypothesis gains suppost

from such structures with partial agreement, as the hypothesis suggests that the

partial agreement is the result of the verb agreeing with another constituent, i.e., the

null expletive. This is convincing if we assume that the expletive pronoun is always

singular and gender-variant in conformity with the post-verbal subject, and agrees

with the verb accordingly. However, assuming that there is an agreement operation

between the null expletive and the verb means that there is a probe-goal relation.

It is argued earlier that the probe is typically the head and the minimal

projection in the derivation. This signals the verb “naama” as the probe and the null

expletive as the goal. However, assuming that the null expletive is base-generated in

the spec-T position via external merger with the T’ means that the goal is in a higher

position than the probe and, thus, out of its domain of c-command. This is

problematic from a minimalistic perspective as the syntactic relations are subject to

c-command. Such a complication is overcome if we posit that the null expletive, like

the preverbal subject, originates VP-internally, within the c-command of the probe,

and raises to satisfy the EPP feature of the tense head in the VS structures.

English also provides evidence that preverbal expletives originate VP-

internally. Consider the following example:

45.It is believed that he is going to win

71
 If we adopt the analysis of Chomsky (1999; 2001) and Mohammed (1990; 2000)

and assume that the expletive is merged directly with the T’; we end up with the

following structure:

46.

It is clear that the tense head "is" agrees in number and person with the

expletive pronoun "it". The reason to assume that "is" cannot agree with the subject

of the second CP "he" is due to the Impenetrability Condition, which blocks the

constituents within the c-command domain of a complementiser from being affected

by higher heads. Moreover, the tense head "is" is independent of the pronoun head

in that a structure such as "it is believed that WE are going to win" is also

grammatical.

If we adopted the assumption that the expletive pronouns originate within the

VP; we end up with the following representation:

72
 47.

This analysis is closer to the spirit of minimalism as it has a probe-goal

relation between a head, which is the probe, and another constituent, which is the

goal, within its domain of c-command.

The agreement pattern of Arabic and English provides evidence that the

expletive pronoun, specifically the null expletive in Arabic VS structures, is

generated within the verb phrase and raises to the spec-T position to satisfy the EPP

feature of the tense head in compliance with the Verb-Phrase Internal Subject

Hypothesis17.

17
The assumption that the subject are base-generated VP-internally has been accepted since the
mid 1980’s. It is discussed by several authors including Speas (1986) and Sportiche (1988) among
others. The term is probably coined by Koopman and Sportiche (1991).
73
 Chapter Five

The Structure of Arabic Sentences: A Unitary Characterization

It is the core of minimalism that the theoretical considerations are reduced to

a minimum when accounting for sentence types and the structure thereof. Chapter

four argues that accounting for nominal, SV and VS sentences as having three

different structures is invalid and against the spirit of minimalism. The chapter,

instead, posits that these sentences can be accounted for using the same framework.

This chapter introduces the structure of Arabic verbless sentences with particular

reference to the SV sentences and their analogous VS counterparts.

5.2. The Functional Categories in the Arabic Sentence

The main claim in this thesis is that all Arabic sentences are fully-fledged

constructions that are CP+TP+VP. What is different about these structures is the

phonological, syntactic and lexical properties of the constituents of their functional

categories.

The structure of an Arabic sentence is analogous to that of the English. Consider

the following structure:

1. ʔal-walad-u yalʕab-u bil-kura-ti

the-kid-Nom plays-3PSM with-the-ball-Gen

74
 The kid plays with the ball

The structure of this sentence can be represented in the following tree

diagram:

2.

The noun kura is merged with the determiner al to form the determiner phrase

DP al-kura, which, in turn is merged with the preposition bi forming a prepositional

phrase PP bil-kurati. Note that the final ti is the result of linking the short vowel i,

which is a genitive case marker, and the final t (Shamsan & Tayyib, 2015). The PP

is merged with the verbal head yalʕabu to form the intermediate projection of the

verb phrase V’.

In accordance with the Verb Phrase Internal Subject Hypothesis (VPISH),

subjects originate within the verb phrase (McCloskey, 1997). Therefore, the subject

ʔal-walad is merged as the specifier of the verb forming the verb phrase. The verb

phrase is, then, merged with a tense head that is affixal in nature forming the

75
intermediate tense projection T’. The tense head has EPP feature that require it to be

extended to a maximal projection.

It is believed that the syntactic operations apply as soon in the derivation as

possible in conformity with the Earliness Principle (Pesetsky, 198918). Thus, the

strong tense affix acquires a host by triggering the movement of the verb yalʕabu

from V to T19. another movement is spec-V to spec-T by which the DP ʔal-walad

raises to the specifier of the tense head. This movement applies to satisfy the EPP

feature of the tense head. The maximal projection of the tense head is merged with

a null declarative complementiser forming the complementiser phrase CP.

In order to better understand the structure of these sentences, we shall opt for

the more updated framework, namely the split-projection analysis20. Such analysis

helps understand the structural differences that result in semantic differences. Let us

first consider the difference between the structure in (1) and the following structure:

3. yalʕab-u ʔal-walad-u bil-kurat-i

plays-3PSM the-kid-Nom with-the-ball-Gen

18
The Earliness Principle (Pesetsky, 1989), at bottom, is an alternative to Chomsky’s Economy of
Derivation Principle (1989). Pesetsky (1989: 3) posits that instead of choosing the derivation with
the smallest number of steps, we should opt for the one where grammatical filters are satisfied as
early in the derivation as possible.
19 The strength of affixes and movement is discussed in Chomsky’s work (1975). It is referred to
as Chomsky’s Strength Metaphor (Radford, 2004).
20
The split-VP analysis was initiated by Chomsky, and was developed by Larson (1990), Chomsky
(1995) and Harley (2002) (See: Radford (2004).
76
 The kid plays with the ball

There are subtle differences between sentences (1) and (3), the most

paramount of which is the constituents order. Sentence (1) is SV while (3) is (VS).

Indeed, such a difference is by no means haphazard nor arbitrary. There is a

difference in meaning between the two structures. The contention is that semantic

differences are bound to correspond to structural differences.

Consider the derivation of (1): The prepositional phrase PP bil-kurati is

merged as the complement of the lexical verb yalʕabu forming the intermediate

projection of the verb phrase V’ structure bellow:

4.

In accordance with the VPISH, the subject originates VP-internally.

Therefore, the V’ is merged with the DP ʔal-waladu to form the VP. This VP is, in

turn, merged as the complement of a null light verb which is a strong affix that

requires a host of a verbal nature. This merge results in the vP as shown below:

77
 5.

It should be noted that the syntactic operations apply as soon as merge is

complete. The diagrams, however, dismiss that for mere explanatory purposes.

Subsequently, the vP in (5) is merged with a tense head, which is a strong affix with

EPP feature.

6.

The tense affix acquires a host by triggering the movement of the lexical verb

and the verbal affix occupying the light verb position to adjoin to the T position. In

78
addition, the EPP feature is satisfied and, thereby, deleted by triggering the

movement of a (pro)nominal constituent. The Attract Closest Condition (Richards,

1997) locates the DP ʔal-waladu as an appropriate candidate for movement being

the closest nominal constituent. This movement results in the extension of the

intermediate projection T’ to the maximal projection of the tense phrase as shown in

diagram (7).

7.

The diagram shows that there are three movement operations that apply in the

derivation so far. The first movement is the raising of the lexical verb yalʕabu from

V-to-v. This movement is triggered by the light verb, which is a strong verbal affix

in the v position. The second and third movements apply simultaneously. Movement

(2.a) is triggered by the strong tense affix, which causes the lexical verb and the light

verb affix to raise to the T position. Affix hopping does not apply in Arabic inasmuch

79
as Arabic tense affix is strong as Arabic has a rich morphology21. In chorus with

(2.a), another movement applies, namely (2.b), where the DP ʔal-waladu undergoes

Spec-V to Spec-T movement. This movement is triggered by the EEP feature of the

tense head. It should be noted that the placement of the null affixes before the verb

in the tree diagram above is for explanatory purposes and does not denote that these

affixes are prefixes. For all intents and purposes, Arabic morphology is a root-pattern

system that allows prefixes, infixes and suffixes (Watson, 2002).

The TP is merged with a null declarative complementiser to form the CP

shown below:

8.

21
See Radford (2009: 143-164) for more details on the relationship between the tense affix relative
strength and the richness of agreement morphology.
80
 As movement operations involve a copy-deletion process, the remnant of the

deleted copies are not visible at the PF level (Chomsky, 1995).

9. [CP [C Ø] [TP [DP ʔal-waladu ] [T' [T Ø+Ø+yalʕabu] [vP [v Ø+yalʕabu] [VP [DP

ʔal-waladu [V’ [V yalʕabu] [PP bil-kurati]]]

At this juncture, there is a need to account for the structural differences between

the structure in (1), as illustrated in (8), and its VS counterpart structure in (3). As

shown in the previous chapter, VS sentences contain a null expletive in the spec-T

position as posited by Mohammad (1990; 2000). His analysis was revisited as we

argued that the null expletive originates within the verb phrase. Let us consider the

derivation of the VS counterpart of (1):

Let us suppose that the derivation of (3) runs parallel to (1). However, unlike SV

structure, VS structures contain an intermediate projection of the light verb, and that

this projection is extended to its maximal projection via a merger operation with a

null expletive pronoun as follows:

10.

81
 Subsequently, the vP in (10) is merged with a strong tense affix with EPP

feature. The former feature triggers the raising of the lexical verb and the light verb

to the T position, and the latter feature requires the T’ to be extended to TP via

merger with a constituent of a (pro)nominal nature. The Attract Closest Condition

dictates that the null expletive be moved to the left-edge of the tense head by virtue

of being the closest candidate for movement.

11.

As soon as the null expletive moves to spec-T position the EPP feature is

satisfied and deleted, i.e., no other movement operations are triggered by this feature.

This justifies why the DP ʔal-waladu remains in situ. The derivation carries on

parallel to that illustrated in (8) resulting in the following structure:

82
 12.[CP [C Ø] [TP [PRN Ø] [T' [T Ø+Ø+yalʕabu] [vP [PRN Ø] [v' [v Ø+yalʕabu] [VP [DP

ʔal-waladu] [V’ [V yalʕabu] [PP bil-kurati]]]

At this juncture, there is a need to account for the verbless sentences in MSA

such as:

13.Ɂana muɁallim-un

I teacher-nom

I am a teacher

The framework’s validity is achieved if it is applied to this type of sentences.

However, some requirements have to be met as the framework has to account for:

R1: The customary null spell-out of the copula.

R2: The obligatory spell-out of the copula when negated.

R3: The obligatory spell-out of the copula in the presence of a modal verb.

R4: The accusative case assigned to its complement in case of obligatory spell-

out.

Let us consider the derivation of (13). The structure in (13) runs parallel in

derivation to the structures in (1) and (3). The nominal muɁallim is merged with a

null determiner to form the DP as follows:

83
 14.

The DP is merged as the complement of a verbal head. The assumption is that

this head is a verbal affix. This merger operation results in the intermediate

projection of the verb phrase as illustrated below:

15.

The intermediate projection of the verb phrase is extended to its maximal

projection via merger with the lexical subject in line with the VPISH. This forms the

VP below:

16.

84
 This VP is merged as the complement of the null light verb, which is a strong

affix that triggers the V-to-v movement as follows:

17.

The vP is, subsequently, merged with a null tense head with tense and EPP

features to form T’. The claim that the null tense head has tense feature is motivated

by Chomsky’s discussion that all head must have interpretable features at the

semantic interface. Radford (2004) expands on this by claiming that all heads,

whether null or overt, have interpretable features, and that “a seemingly null T

constituent contains an abstract affix carrying an interpretable tense feature” (p.

143). This means that the null tense head in (18) has a tense feature.

In this regard, the verbal affixes in v and V position require an appropriate

host. The tense feature of the tense head qualifies it as an appropriate host for the

verbal affixes, which signals the movement of the verbal affix in V to T. however,

movement is governed by the Head Movement Constraint. Therefore, the verbal

85
affix in V raises to v then it raises to T with null light verb as shown below in (1.a).

On the other hand, the EPP feature requires the tense head to have a nominal on the

left-edge. This triggers the closest (pro)nominal to move to the spec-T position as

shown below in movement (1.b). The two movements apply at the same time

resulting in the TP in (18).

18.

The structure in (18) is merged with a null declarative complementiser to from

the full CP in the fashion shown earlier. What is noteworthy at this point is that the

tense head is a null complete head rather than an affix, which means that there is an

appropriate host for the verbal affixes.

The structure of the seemingly verbless sentence is, thus, represented as

follows:
86
 19.[CP [C Ø] [TP [PRN Ɂana] [T' [T tense+Ø+Ø] [vP [v Ø+ Ø] [VP [PRN Ɂana] [V’ [V Ø]

[DP muɁallimun]

The derivation after the copy-deletion process is handed over to the phonetic

form, which gives it the spell-out shown in (13). This framework somewhat explains

the syntactic intricacies behind the null spell-out of the copula in certain conditions

(R1). In other words, the framework assumes that there is a null tense head in the

verbless sentences in positive present tense; the adjunction of the verbal affixes to

this null head results in the null spell-out of the constituent “[T tense+Ø+Ø]”. Now

consider the following example:

20.lan akuuna muɁallim-an

not be teacher-Acc

I will not be a teacher

21.*lan muɁallim-un

not be teacher-Acc

I will not be a teacher

The structure in (21) is ungrammatical in that the copula has to be overtly

spelled out. In order to understand the syntactic reasons behind the obligatory spell-

out, we need to break down the derivation of the structure in (20). The structure in

(20) is rather different from the previous structures. The structures discussed so far

87
are either verbless, SV or VS with an overt lexical subject. However, the structure in

(20) is, seemingly, subjectless.

We argued earlier that the VS and SV structures differ in whether or not there

is a null expletive in the spec-T position. Therefore, determining the structure of (20)

necessitates arguing whether it is SV or VS.

There is a solid empirical evidence to assume that the null subjects in Arabic

are preverbal, and that all subjectless clauses are SV. For one, subject-verb

agreement asymmetry has implications about the word order of the sentence.

Consider the following examples:

22.ʔarriʤaal-u ʤaaʔuu

The-men-Nom came-3PMP

The men came

23. ʤaaʔa rriʤaal-u

came-3PMS the-men-Nom

The men came

24.*ʤaaʔuu rriʤaal-u

came-3PMP the-men-Nom

The men came

The example in (22) is SV and subject-verb agreement is full whereas the

example (23) is VS, and agreement is partial. The ungrammaticality of the example

88
in (24) shows that full agreement in Arabic occurs only if the lexical subject is

preverbal22. Let us consider the following examples:

25.lan yakuunuu muɁallim-iina

not be-3PMP teachers-Acc

They will not be teachers

26.*lan yakuuna muɁallim-iina

not be-3PMP teacherS-Acc

They will not be teachers

The example in (25) shows an instance of full agreement, and the non-

elliptical example (26) shows that the full agreement is obligatory in the, seemingly,

subjectless sentences. We noted earlier that full agreement in MSA occurs only in

the context of preverbal subjects. This implies that the null subject in (20) is

preverbal, and the structure is SV. In light of that, the structure in (20) is derived as

follows:

The noun phrase muɁallim is merged with a null determiner to form the DP.

Subsequently, the DP is merged as the complement of a verbal affix forming V’. The

maximal projection of the verb phrase VP is formed via the merger of the null first

22
Such structures are grammatical in older varieties of Standard Arabic and in many modern
Arabic dialects (See Wright,1898; Russel, 1984;). It is referred to as the language of /akaluuni
lbaraaɣiiθ/ “the flees devoured me” by Sybawayh (Owens, 2007).
89
person singular pronoun, referred to as “Ɂana” for explanatory purposes, and the V’

as shown in (16). The VP is merged with a null light verb, which is a strong affix

that triggers the movement of the null verbal affix to adjoin to it in the same way as

in (17) forming the vP.

Subsequently, the vP is merged as the complement of the negation head “lan”

to form the negation phrase NEGP illustrated below:

27.

The negation phrase is merged as the complement of the tense head, which is

a tense affix with EPP feature, to form the intermediate projection T’. The EPP

feature triggers the null pro in spec-V to move to Spec-T. The tense head has a tense

feature [TNS] that requires it to attach to a verbal host.

There is a number of empirical pieces of evidence to posit that the negation

head is located between the VP and the TP even though negation appears before the

verb in examples such as the following:

90
 28.Lan yaʃrab-a lħaliib-(a)

pro NEG drink-Acc-3PSM the-milk

He will not drink milk

29.Lam yaktub(i) ddars-(a)

pro NEG write-Jus-3PSM the-lesson

He did not write the lesson

30.ʔalqalamu Laa yaktub(u)

The-pen NEG write-Nom-3PMS

The examples above show three instances of preverbal negation. This may

suggest that the negation head is above the tense head. However, the sentences in

(28), (29) and (30) above have three different time references “future, past and

present respectively”. What is noteworthy about the three sentences is the verb is in

the same form even though the tenses are different. This suggests that the verb is not

inflected tense-wise, and that the time reference is not held by the verb but by another

constituent as the verbs are non-finite.

In line with Aoun et al. (2010), the negation head “lan” is tensed negation

particle, which means that it can adjoin to the tense head. This can be argued by the

fact that the tense head “sawfa” and the negation head “lan” are mutually exclusive.

Consider the following example:

91
 31.*Sawfa lan yaʃrab-a lħaliib-(a)

pro NEG drink-Acc-3PSM the-milk

He will not drink the milk

The structure (31) is ungrammatical because the tense head “sawfa” is

superfluous in the derivation. Syntactically, the tense head is already occupied by

the negation head, so there is no room for extra tense heads.23

In this regard, the negation head raises to attach to the tense affix “Neg-to-T

movement”. Once the tensed negation head adjoins to the tense affix there is no need

for other constituent movement.

32.

The raising of the negation head to the tense head satisfies its tense feature

and provides a host for the tense affix in T, which eliminates the need of the verbal

23
Aoun et al. (2010) argue that there are other tensed negation heads.
92
affixes to move to adjoin to it. The verbal affixes remain stranded, but the derivation

cannot be handed over to the PF component with floating affixes, i.e., stranded,

otherwise it crashes. The stranded affixes are spelled as a form of the verb to be in a

process that is similar to the DO-support in English, as illustrated below:

33.

Although not documented in the literature, the hypothesis of BE-support saves

the derivation from crashing. We can account for the accusative case assigned to its

complement by assuming that the copula resulting from the BE-support has the same

lexico-syntactic properties of the verb “yakuun”. It is known that the overt “yakuun”

assigns accusative case to its nominal or adjectival complement.

93
 The BE-support hypothesis in the previous analysis gains support from other

structures such as the structures that contain a modal verb “qad” or any other head

that can occupy the tense head such as “sa” or “sawfa”. Consider the following

examples:

34. Ɂana qad akuunu muɁallim-an

I may be teacher-nom

I may be a teacher

35.*Ɂana qad muɁallim-un

I may teacher-nom

I may be a teacher

36. Ɂana sawfa/sa akuunu muɁalli-man

I may be teacher-nom

I may be a teacher

37.*Ɂana sawfa/sa muɁallim-un

I may teacher-nom

I may be a teacher

The obligatory spell-out of the copula can be accounted for if we adopt the

BE-support hypothesis. In the previous examples, the tense head is occupied by non-

94
inflectional constituents24 obviating the need for verb movement to T. Given that the

verbal affixes cannot attach to either of them, they remain stranded, which

necessitates the BE-suport (R3).

5.2. The Structure of copular Clauses in Maghrebi Dialects

As shown in previous chapters, MSA “verbless”25 sentences have their

counterparts in Maghrebi Dialects. The following sections discuss the structure of

the “raa” clauses with particular reference to the framework proposed earlier. It

should be noted that case in modern dialect is not apparent on word-final (Aoun et

al., 2008). Consider the following sentence:

38.Lbeet raah wasәʕ

The house is-3Psg large

The house is large

In order to understand the structural and derivational intricacies of the

structure in (38), we need to understand the syntactic properties of the main verb

“raa”. We have noted in earlier chapters that the verb “raa” in Maghrebi dialects is

24
“qad”, “sa” and “sawfa” are non-inflectional as they show no patterns of case, gender, number
or person agreement. Such agreement morphemes cannot attach to them, for we cannot say “ʔana
sawfa” is the first person singular counterpart “hum sawfuuna” or “hiya sawfat”.
25
The quotation marks denotes that the label “verbless” does not denote the absence of a verbal
projection but rather is a conventional term.
95
a grammaticalized form of the imperative form of the verb “raʔaa” “see”, which

spells out as “raa”.

A major syntactic difference between the copula and the verb “see” is that the

latter is transitive26 while the copula grammaticalized from it is intransitive. This

syntactic difference results in the verb “raʔaa” requiring an object whereas its

grammaticalized form does not. Consider the following examples27:

39. Ahmed, raa lbeet!

Ahemed, look-Impr the-house

Ahmed, look at the house!

40. Lbeet raah ya Ahmed

the-house look-it O Ahmed

Look at the house, Ahmed!

We noticed that the verb “raa” in imperative is transitive and has an object.

The example (40) shows that there is a resumptive pronoun cliticised onto the main

verb as a result of object-topicalisation. Although such structures are extinct in the

Maghrebi dialects and the uses of the verb “raʔaa” is no longer common, we would

expect the verb to agree with the subject in the following fashion:

26
The verb “see” is intransitive when expressing the ability to see or understand.
27
Such structures ceased to exist in Maghrebi dialects. Therefore, the examples are reconstructive
rather than expressive of the current linguistic practices.
96
 41.a. Layla, ree lbeet!

Layla, look at the house

b. ʃʃabeeba, ruu lbeet!

Guys, look at the house.

The agreement pattern in the previous example is reconstructed by analogy to

other verbs with the same root and syntactic properties, such as the verb “daa”

“take”, which conjugates as “dee” “duu” in feminine and plural respectively. The

point of this discussion is to note that the verb “raʔaa”, being the source of

copularisation, has subject-verb agreement markers. However, the copula “raa” in

the Maghrebi dialects is morphologically non-inflectional as the root “raa” is

retained regardless of the subject’s number and gender.

We hypothesize that the clitic pronoun is a devised clitic pronoun to achieve

subject-verb agreement. In other words, the agreement pattern apparent on other

verbs in these dialects cannot be applied to the grammaticalized imperative of

“raʔaa”. As a result, the language makes use of the pronoun apparent in the original

structures, such as in (40), to mark subject-verb agreement.

With the assumption that the clitic pronoun attached to the end of the verb

resolves the agreement issue, we would expect the derivation of (38) to be as follows:

97
 The adjective wasəʕ is merged as the complement of the verb raa to form the

intermediate projection of the verb phrase V’. The latter is extended to its maximal

projection via merger with the lexical subject lbeet as illustrated bellow:

42.

Subsequently, the VP is merged as the complement of a null light verb which

is a strong affix that requires a verbal host. This triggers the movement of the verb

raa to adjoin to it to form vP as illustrated below:

43.

This vP is merged as the complement of the tense head, which is a strong tense

affix with EPP feature forming the T’. The tense feature requires the T head to

acquire a host by triggering the v-to-T movement. Moreover, the EPP feature

98
requires the T head to have a (pro)nominal specifier on the left-edge of it, which

triggers the spec-V to spec-T movement, resulting in the TP illustrated bellow:

44.

Subsequently, the TP in (39) is merged with a null complementiser to form

the CP. At a certain stage of the derivation, the agreement operation applies resulting

in the [Ø+Ø+raa] becoming “raah”, as the “h” is a subject-verb agreement marker.

99
 45.

5.3. The Negation of “raa” in the Maghrebi Dialects

Another issue that arises in light of this discussion relates to negation. It

should be noted that the structure in (31) is generally negated as follows:

46. lbeet ma-raahʃ wasәʕ

The house NEG-is-NEG large

The house is not large

100
 The discussion of negation in this section builds on the Complex Head

Hypothesis of Benmamoun (2000), which posits that the negation head is a complex

head that consists of a proclitic and enclitic “maa+…+ʃ”, which is spelled as “miʃ”

in the absence of any intervening constituent. Such analysis opposes the view set by

many, positing that “miʃ” is a different negation particle28.

The structure in (46) is derived in a similar fashion to (38) as illustrated

below:

47.

Subsequently, the vP is merged with a negation head consisting of a proclitic

and enclitic, as follows:

28
Al-Moumani (2011) among others analyze “miʃ” and “maa+..+ʃ” differently.
101
 48.

The negation phrase is merged with an affixal tense head with EPP feature.

The tense feature of the tense head requires it to have an appropriate host. This

triggers the movement of the verb to raise to the T position. However, the head

movement is governed by a condition called Head Movement Constraint (HMC),

which posits that “Head movement is only possible between a given head and the

head of its complement” (Radford, 2009: 176). Put otherwise, HMC requires that

head movement be applied in a step-at-a-time way, and that movement is permitted

only between a given head and the head directly above it within the same structure.

The HMC leads to the v-to-NEG-to-T movement. In line with Benmamoun’s

analysis (2000), the negation clitics adjoin to the verb resulting in the adjunction

[maa+raa+ʃ]. The movement proceeds to the tense position.

Triggered by the EPP feature of the tense head, another movement takes

place leading the closest DP to move to the spec-T position. This Attract Closest

102
Condition locates the DP “lbeet” as the closest nominal candidate. The derivation

continues as the TP is merged with a null complementiser forming the CP illustrated

bellow:

49.

As noted earlier, the clitic pronoun “h” is the result of subject verb agreement.

103
6. Conclusion

The current study represents an attempt to provide a unitary characterisation

for Arabic clauses. The main assumption in the thesis is that Arabic verbless

sentences are fully-fledged clause with CP, TP and VP layers. The thesis accounts

for Arabic SV, VS and verbless clauses.

With regard to Arabic VS structures, it has been argued that the preverbal

subject are in the spec-T position but are base-generated VP-internally in accordance

with the VPISH. The raising of the subjects is triggered by the EPP feature of the

tense head. It has also been argued that all tense heads in Arabic have EPP feature.

With regard to VS clauses, it has been argued that the reason the lexical

subjects remain in situ is because there is a null expletive in the spec-T position in

order to satisfy the EPP feature of the tense head. This accorded with the Null

Expletive Hypothesis (Mohammed, 1990; 2000). However, the thesis argued that

the expletives are not base-generated in the spec-T but rather raise to that position

from the spec-v. This movement is triggered by the EPP feature of the tense head.

It has also been argued that Arabic “verbless” clauses contain a verbal

category and a tense head. It was argued that the verbal category is occupied by a

verbal affix that raises to the tense head to attach to a null present tense head. It was

argued that this adjunction remains null. However, the tense head can be occupied

104
by constituents other than the aforementioned null tense head. These tense include

the modal verb “qad”, the future tense heads “sa” and “sawfa” or the tensed negation

head “yakuun”. This causes the verbal affix and the null light verb to remain

stranded, which causes the derivation to crash. The derivation is saved from crashing

by a process called “BE-suppot”. We hypothesized that the stranded affixes

lexicalize in the form of a dummy verb “yakuun” in a process that is analogous to

the English “DO-support”.

Another concern of the study is to account for Arabic verbless phrases’

counterpart in the Maghrebi Dialects. It was argued that the “raa” appearing in the

Maghrebi dialects locative and equational clauses is a grammaticalized form of the

verb “see”. The thesis accounts for the structure of such clauses with particular

reference to the standard Arabic clauses.

105
References

Abdul Sattar, H. (2012). Fundamentals in Classical Arabic. Sacred Learning:

Chicago
Abu Hian, A. (1377- AD), Ettatheel w Attkmeel, H. Hendawi (ed.), 2001 Impression,
Damascus, Dar Al-Kalm.

Adger, D. (2002). Core Syntax: A Minimalist Approach. Oxford University Press.

Adger, D. & Svenonius, P. (2010). Features in Minimalist Syntax. In The Handbook

of Linguistic Minimalism, C. Boeckx (ed.), pp. 27-51. Oxford: Blackwells.

Al-Balushi, R. (2012). Why Verbless Sentences in Standard Arabic are Verbless.

Canadian Journal of Linguistics/Revue Canadienne de Linguistique, 57(01), pp. 1-
30.

Al-Liheibi, F. M. (1999). Aspects of Sentence Analysis in the Arabic Linguistic

Tradition, with Particular Reference to Ellipsis (Doctoral dissertation, University of
Durham).
Al-Momani, I. M. (2011). The syntax of sentential negation in Jordanian Arabic.
Theory and Practice in Language Studies, 1(5), 482-496.

Al-Rashed, M. A. (2012). Argument Structure in Arabic: Lexicon or Syntax?

(Doctoral dissertation, Arizona State University).

AlShammiry, K. M. (2016). Subject Verb Agreement in Modern Standard Arabic

(MSA) and Saudi Dialect of Arabic (SA): A New Minimalist Account. International
Journal, 4(2), 83-92.

Al-Shorafat, M. O. (1998). The Minimalist Program and the Structure of Arabic

Clauses in an Agr-Based Model. Papers and Studies in Contrastive Linguistics, 34,
pp. 123-139.

Aoun, J., & Li, Y. H. A. (1993). Syntax of Scope. (Vol. 21). Mit Press.

Aoun, J. E., Benmamoun, E., & Choueiri, L. (2010). The Syntax of Arabic.
Cambridge University Press.

Babby, L. H. (1980). Existential Sentences and Negation in Russian. Karoma Publ.

106
Bahloul, M., & Harbert, W. (1993). Agreement Asymmetries in Arabic. In
Proceedings of the eleventh West Coast conference on formal linguistics (Vol. 15),
p. 31.

Baker, M., &Vinokurova, N. (2010). On Tense and Copular Verbs in Nonverbal

Predications in Sakha. Rutgers Working Papers in Linguistics, Vol. 3, pp. 31-63.
New Brunswick: Rutgers University Press.

Bakir, M. J. (1980). Aspects of Clause Structure in Arabic: a Study in Word Order

Variation in Literary Arabic. Indiana University Linguistics Club.

Benmamoun, E. (2000). The Feature Structre of Functional Categories: A

Comparative Study of Arabic dialect. Oxford & New York, Oxford University Press.

Benmamoun, E. (2008). Clause Structure and the Syntax of Verbless Sentences.

Current Studies in Linguistics Series, 45, 105.

Biberauer, T., Holmberg, A., Roberts, I., & Sheehan, M. (2010). Parametric
Variation Null Subjects in Minimalist Theory. Cambridge: Cambridge University
Press.

Chomsky, N. (1964). Current Issues in Linguistic Theory. In The Structure of

Language, J. A. Fodor & J. J. Katz, (eds.), 50-118. Englewood Cliffs NJ: Prentice
Hall.

Chomsky, N. (1981). Lectures on Government and Binding. Dordrecht: Foris.

Chomsky, N. (1982). Some Concepts and Consequences of the Theory of
Government and Binding. Cambridge, MA: MIT Press.

Chomsky, N. (1986a). Barriers. Cambridge, MA: MIT Press.

Chomsky, N. (1986b). Knowledge of Language: Its Nature, Origin and Use. New
York, NY: Praeger.

Chomsky, N. (1989). Some Notes on Economy of Derivation and

Representation. MIT Working Papers in Linguistics, 10, 43-74.

Chomsky, N. (1993). ‘A Minimalist Program for Linguistic Theory’. In The View

from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger, K. Hale &

107
S. J. Keyser (eds.) (1995), 167-217 . Cambridge, MA: MIT Press. [Reprinted in
Chomsky,1-52.]

Chomsky, N. (1995). The Minimalist Program. MIT Press.

Chomsky, N. (1999). Derivation by Phase. In MIT Occasional Papers in Linguistics.

MIT.
Chomsky, N. (2000). Minimalist Inquiries: The Framework. In Step by Step: Essays
on Minimalist Syntax in Honor of Howard Lasnik, R. Martin, D. Michaels, & J.
Uriagereka (eds.), 89-15. Cambridge, MA: MIT Press.

Chomsky, N. (2001a). Derivation by Phase. In Ken Hale: A Life in Language, M.

Kenstowicz (ed.), 1-52.: MIT Press.

Chomsky, N. (2001b). Beyond Explanatory Adequacy. MIT Occasional Papers in

Linguistics 20. Cambridge, MA: MIT Press

Chomsky, N. (2004). Beyond Explanatory Adequacy. In Structures and Beyond, A.

Belletti (ed.), 104-131. Oxford- New York: Oxford University Press.
Chomsky, N. (2005). On Phases. Unpublished Manuscript. MIT

Chomsky, N. and H. Lasnik. 1993. The Theory of Principles and Parameters. In

Syntax: An International Handbook of Contemporary Research, J. Jacobs, A. von
Stechow, W. Sternefeld, & T. Vennemann (eds.), 506-569. Berlin: Walter de
Gruyter. [Reprinted in Chomsky (1995), 13-127.].
Creissels, D. (2015). Copulas Originating from ‘See/Look Verbs in Mande
Languages. In Symposium “Areal Patterns of Grammaticalization and Cross-
linguistic Variation in Grammaticalization Scenarios” (pp. 12-14). Mainz
University

Creissels, D., & Taine-Cheikh, C. (2016). From ‘See’ Verbs to Copulas: a Little-
Known Grammaticalization Path. Retrieved from:
http://lacito.vjf.cnrs.fr/documents/pdf_membres/Cheikh_Biblio_type.pdf.

Curtiss, S. (1977). Genie: A Psycholinguistic Study of a Modern-Day ‘Wild Child’.

New York, NY: Academic Press.
Déchaine, R. M. A. (1993). Predicates across Categories: Towards a Category-
Neutral Syntax. Doctoral Dissertation. Massachusetts.

108
De Saussure, L. (2007). Recent Advances in the Syntax and Semantics of Tense,
Aspect and Modality (Vol. 185). Walter de Gruyter.

Eisele, J. C. (1988). The Syntax and Semantics of Tense, Aspect, and Time Reference
in Cairene Arabic. UMI.

Falk, Y. N. (2004). The Hebrew Present-tense Copula as a Mixed Category. In

Proceedings of the Lexical Functional Grammar “04 Conference”, B. Miriam&
K.H. Tracy. New Zealand: Canterbury University.
Farghal, M. (1992). The Arabic Topic-Comment Structure. Journal of King Saud
University, 4(1), 47-62.

Fehri, F. A. (1993) Issues in the Structure of Arabic Clauses and Words. Kluwer,
Dordrecht.

Ford, D. (2013). The Influence of Word Order on Modern Standard Arabic

Information Structure (PDF file).www.gial.edu/.../Ford-Info-Structure-and-Word-
Order-MSA.pdf. (Retrieved in June)

Fodor, J. A. 1983. The Modularity of Mind, Cambridge, MA: MIT Press.

Gergel, R. (2009). Modality and ellipsis: Diachronic and synchronic evidence (Vol.
202). Walter de Gruyter.

Harley, H. (2002). Possession and the double object construction. Linguistic

variation yearbook, 2(1), 31-70.
Harper, M. P., & Charniak, E. (1986). Time and Tense in English. In Proceedings
of the 24th Annual Meeting on Association for Computational Linguistics, 3-9.
Morristown NJ: Association for Computational Linguistics.

Harrat, S., Meftouh, K., Abbas, M., Hidouci, K. W., & Smaili, K. (2016). An
Algerian Dialect: Study and Resources. International Journal of Advanced
Computer Science and Applications-IJACSA, 7(3).Chicago

Hazout, I. (1995). Action Nominalizations and the Lexicalist Hypothesis. Natural

Language & Linguistic Theory, 13(3), 355-404.

Heine, B., Kuteva, T. (2002). World Lexicon of Grammaticalization. Cambridge:

Cambridge University Press.

109
Hengeveld, K. (2011). The Grammaticalization of Tense and Aspect. The Oxford
Handbook of Grammaticalization, 580-594.

Hornstein, N., Nunes, J., & Grohmann, K. K. (2004). Understanding Minimalism.

Cambridge: Cambridge University Press.

IbnHesham, J. (1590- AD). Awdaho Al-Masalik vol 1.(1994) Impression, Beirut,

Dar al-fi kr.

Ibn Jinni, A. (1002-AD). SirrSinaat Al-Iarab. H. Hendawi (ed.), (1993) Impression,

Damascus: Dar Al-Qalm.

Ibrahim, A. (2005). Light Verbs in Standard and Egyptian Arabic. In Perspectives

on Arabic Linguistics XVII-XVIII: Papers from the Seventeenth and Eighteenth
Annual Symposia on Arabic Linguistics, 117-131. Amsterdam: John Benjamins.
Jackendoff, R. (1972). Semantic Interpretation in Generative Grammar. Cambridge,
MA: MIT Press.
Jalabneh, A., & Abdellatif, K. (2014). Embedded Complementizer Phrase
Subjunctive and Features Checking: A Contrastive Study between English, Arabic
and German. Damascus University Journal, 30( 3+4), 109-135.

Jelinek, E. (1981). On Defining Categories: Aux and Predicate in Colloquial Arabic.

Ph.D Thesis, University of Arizona: Tuscon.
Jenkins, L. (2000). Biolinguistics: Exploring the Biology of Language. Cambridge:
Cambridge University Press.
Koopman, H. and Sportiche, D. (1991). The Position of Subjects. Lingua, 85, 211-
258.

Larson, R. K. (1990). Promise and the Theory of Control. Linguistic Inquiry, 22(1),
103-139.

Lasnik, H. & M. Saito. (1984). On the Nature of Proper Government. Linguistic

Inquiry, 15,235-289.

Lasnik, H. and M. Saito. (1992). Move α: Conditions on Its Applications and

Outputs. Cambridge, MA: MIT Press.
Lehmann, C. (2002). Thoughts on Grammaticalization (1st ed.). Erfurt: Seminar für
Sprachwiss der Univ.

110
Levin, A. (1995). The fundamental principles of the Arab grammarians' theory
of'amal. Jerusalem Studies in Arabic and Islam, 19, 214-232.

McCloskey, J. (1997). Subjecthood and subject positions. In Elements of

grammar (pp. 197-235). Springer Netherlands.

Mohammad, M.A. (1990). The Problem of Subject-Verb Agreement in Arabic:

Towards a Solution. In Perspectives in Arabic Linguistics I, Eid, M. (ed.), 95-125.
Amsterdam: Benjamins.
Mohammad, M. A. (2000). Word Order, Agreement, and Pronominalization in
Standard and Palestinian Arabic (Vol. 181). John Benjamins Publishing.

Moro, A. (1995). Small Clauses with Predicative Nominals. In Small Clauses. A.

Cardinalettiog & M.T. Guasti (eds.), 109-132. San Diego: Academic Press.
Mouchaweh, L. (1986). De la Syntaxe des Petites Prespositions. Doctoral thesis,
Universite de Paris VIE, Paris.

Obeidat, H., &Farghal, M. (1994). On the Status of the Equational Sentences in the
Grammar of Arabic. Abhath Al-Yarmouk, 12(2), 9-35.

Owens, J. (2007). A Linguistic History of Arabic. Oxford: Oxford University Press.

Peled, Y. (2007). Problems in the Medieval Arabic Theory of Sentence Types. In

Approaches to Arabic Linguistics, E. Ditters & H. Motzki (eds.), 149-188. Leiden:
Brill.

Pesetsky, D. (1989). Language-Particular Processes and the Earliness Principle.

MS, MIT.

Petit, D. (2010). On Presentative Particles in the Baltic Languages. In Particles and

Connectives in Baltic, N. Nau & N. Ostrowski (eds.), 151-170. Vilnius, Vilniaus
Universitetas: Acta Salensia 2.
Pustet, R. (2003). Copulas: Universals in the Categorization of the Lexicon. OUP
Oxford.
Radford, A. (2004). Minimalist Syntax: Exploring the Structure of English.
Cambridge University Press.

Radford, A. (2009). Analyzing English Sentences: a Minimalist Approach. New

York: Cambridge University Press.
111
Rapapport, T. (1987). Copula, Nominal, and Small Clauses. Ph.D Thesis, MIT,
Cambridge.

Richards, N. (1997) What Moves Where When in Which Language? PhD

dissertation, MIT. (A published version appeared as Richards 2001.)

Rizzi, L. (1990). Relativized Minimality. The MIT Press.

Rizzi, L. (1997). The Fine Structure of the Left Periphery. In Elements of Grammar,
L. Haegeman (ed.), 281 337. Kluwer, Dordrecht.
Rizzi, L. (2000). Remarks on Early Null Subjects. The Acquisition of Syntax: Studies
in Comparative Developmental Linguistics, 269-292.

Rizzi, L. (2001). On the Position “Int(errogative)” in the Left Periphery of the

Clause. Current Studies in Italian Syntax: Essays Offered to Lorenzo Renzi, G.
Cinque &G. Salvi (eds.), 287-296. Amsterdam: Elsevier.

Rizzi, L. (2004). Locality and left periphery. Structures and beyond. The
cartography of syntactic structures, 3, 223-251.
Rosenbaum, P. S. (1965). A Principle Governing Deletion in English Sentential
Complementation. Yorktown Heights NY: International Business Machines Corp.
Ross, J. R. (1967). Constraints on Variables in Syntax. PhD dissertation, MIT.

Ross, D. (2012). A Proposal for a Dichotomy in the Core of Arabic Syntax. Illinois
Symposium on Semitic Linguistics (ISSL). University of Illinois at Urbana-
Champaign.
Rothstein, S. (1995). Small Clauses and Copular Constructions. In Small Clauses,
A. Cardenaletti & T. Guasti (eds.), 27-48. San Diego, CA: Academic Press
Russell, R. A. (1984). Historical Aspects of Subject-Verb Agreement in Arabic. In
ESCOL 84: Proceedings of the First Eastern States Conference on Linguistics.
Columbus, OH: Ohio State University.
Shamsan, M. A. H. A., &Attayib, A. M. (2015). Inflectional Morphology in Arabic
and English: A Contrastive Study. International Journal of English Linguistics, 5(2),
139.
Shlonsky, U. (1997). Clause Structure and Word Order in Hebrew and Arabic: An
Essay in Comparative Semitic Syntax. Oxford University Press.

112
Shlonsky, U. (2000). Subject Positions and Copular Constructions. In Interface
Strategies, H. Benis & M. Everaert (eds.), 325-347. Amsterdam: Royal Netherlands
Academy of Arts and Sciences.

Schütze, C. T. (2004). Synchronic and diachronic microvariation in English

do. Lingua, 114(4), 495-516.
Sibawayh, A. (741). Al-Kitaab. (1977) Impression, Cairo, Dar Al-Qalam Press.

Smith, N. V. & Tsimpli, I. M. (1995). The Mind of a Savant: Language-Learning

and Modularity. Oxford: Blackwell.

Soltan, U. (2006). Standard Arabic Subject-Verb Agreement Asymmetry Revisited

in an Agree-based Minimalist Syntax. Agreement Systems, 239-265.

Soltan, U. (2011). On Issues of Arabic Syntax: An Essay in Syntactic

Argumentation. Brill's Journal of Afroasiatic Languages and Linguistics, 3(1), 236-
280.
Speas, P. (1986). Adjunction and Projections in Syntax. PhD dissertation, MIT,
Cambridge, MA.

Sportiche, D. (1988). A theory of floating quantifiers and its corollaries for

constituent structure. Linguistic inquiry, 19(3), 425-449.

Svenonius, P. (2001). Subjects, Expletives, and the EPP. Oxford: Oxford University
Press.
Taine-Cheikh, C. (2013). Grammaticalized Uses of the Verb ṛa(a) in Arabic: a
Maghrebian specificity. African Arabic: Approaches to Dialectology.
doi:10.1515/9783110292343.121

Talmon, R. (1993). Two Early "Non-Sībawaihian" Views of ‘amal’ in Kernel-

Sentences. Zeitschrift für arabische Linguistik, (25), 278-288.

Traugott, E. C., & König, E. (1991). The Semantics-Pragmatics of

Grammaticalization Revisited. Approaches to Grammaticalization, 1, 189-218.

Versteegh, K. (1994). The notion of ‘underlying levels’ in the Arabic grammatical

tradition. Historiographia linguistica, 21(3), 271-296.
Vydrin, V. (2009). On the problem of the Proto-Mande homeland. Вопросы
языкового родства–Journal of Language Relationship, 1, 107-142.

113
Watson, J. C. (2002). The Phonology and Morphology of Arabic. Oxford: Oxford
University Press on Demand.

Westermann, D., & Melzian, H. J. (1930). The Kpelle language in Liberia:

grammatical outline, colloquial sentences and vocabulary. Berlin: Dietrich Reimer /
Ernst Vohsen.
Wright, J. (1898). 1905. The English dialect dictionary. 6 vols. London: Henry
Frowde.

114
 ‫مستخلص‬

‫امحمدي‪ ،‬نذير‪ .‬تعابير اإللتفات وأفعال الربط‪ :‬تقعيد الفعل "رأى" في اللهجات المغاربية وإثبات‬

‫فرضية فعل الربط المحذوف‪ .‬رسالة ماجستير بجامعة اليرموك‪( 7102 .‬المشرف‪ :‬أ‪.‬د‪ .‬محمد‬

‫عوده الشرفات)‬

‫تمثل هذه الدراسة محاولة لفهم البناء النحوي للجمل في اللغة العربية‪ .‬وقد تبنت الدراسة النظرية التقليصية‬

‫(‪ )Minimalist Theory‬كأساس للتحليل البنيوي النحوي للتراكيب اللغوية‪ .‬وهدفت الدراسة إلى تقديم تحليل‬

‫نحوي موحد للجمل الفعلية واإلسمية في اللغة العربية‪ .‬وقد توصلت الدراسة إلى تقديم تحليل يتضمن فعل ربط‬

‫محذوف في الجمل اإلسمية (الالفعلية) مع مراعاة الشروط اللغوية التي تقتضي الظهور اللفظي لفعل الربط‪ .‬كما‬

‫تطرقت الدراسة إلى اإلختالف بين الفاعلية واإلبتداء لسأسماء التي تبب الفعل‪ .‬باإلاافة إلى لل ‪ ،‬تطرقت‬

‫الدراسة إلى عملية تقعيد الفعل رأى في اللهجات المغاربية حيث أن المعنى الوظيفي والداللي للفعل تغير إلى‬

‫كونه عبارة لفت إنتباه وفعل ربط لا وظائف متعددة‪.‬‬

‫الكلمات المفتاحية‪ :‬عبارات لفت اإلنتباه‪ ،‬أفعال الربط‪ ،‬التقعيد‪ ،‬الجمل الالفعلية‪ ،‬اللهجات المغاربية‪ ،‬فراية فعل‬

‫الربط المحذوف‪ ،‬فراية الضمير اإلبتدائي‪/‬الحشوي المحذوف‪ ،‬فراية فعل الربط الداعم‪.‬‬

‫‪115‬‬