199 74 3MB
English Pages 167 [168] Year 1989
Pragmatic Functions in a Functional Grammar of Arabic
Functional Grammar Series This series comprises monographs and collections written in the framework of Functional Grammar. The aim is to seek explanations for a wide variety of linguistic phenomena, both language-specific and cross-linguistic, in terms of the conditions under which and the purposes for which language is used. Editors:
A. Machtelt Bolkestein Simon C. Dik Casper de Groot J. Lachlan Mackenzie
General address: Institute for General Linguistics Functional Grammar Spuistraat 210 NL-1012 VT Amsterdam The Netherlands Other books in this series: 1. A.M. Bolkestein, C. de Groot and J.L. Mackenzie (eds.) Syntax and Pragmatics in Functional Grammar 2. A. M. Bolkestein, C. de Groot and J . L Mackenzie (eds.) Predicates and Terms in Funtiona! Grammar 3. Michael Hannay English Existentials in Functional Grammar 4. Josine A. Lalleman Dutch Language Proficiency of Turkish Children born in the Netherlands 5. Jan Nuyts and Georges de Schutter (eds.) Getting one's Words into Line 6. Johan van der Auwera and Louis Goossens (eds.) Ins and Outs of the Predication 7. Judith Junger Predicate Formation in the Verbal System of Modern Hebrew
Other studies on Functional Grammar include S.C. Dik, Functional Grammar (1978), T. Hoekstra et al. (eds.). Perspectives on Functional Grammar (1981), S.C. Dik (ed.). Advances in Functional Grammar (1983). All published by FORIS PUBLICATIONS.
Pragmatic Functions in a Functional Grammar of Arabic
ψ
1989 PORIS PUBLICATIONS Dordrecht-Holland/Providence RI - U.S.A.
Published
by:
Poris Publications Holland P.O. Box 509 3300 A M Dordrecht, The Netherlands Distributor
for the U.S.A.
and
Canada:
Foris Publications USA, Inc. P.O. Box 5904 Providence RI 02903 U.S.A. Distributor
for
Japan:
Sanseido Book Store, Ltd. 1-1, Kanda-jimbocho-cho Chiyoda-ku Tokyo 101, Japan
CIP-DATA
Moutaouakil, Ahmed Pragmatic functions in functional grammar of Arabic / Ahmed Moutaouakil ; [transi, from the French by J.L. Mackenzie]. - Dordrecht [etc.] : Foris. - (Functional grammar series ; 8) Transi, of: Les fonctions pragmatiques en Arabe (manuscript). ISBN 90-6765-271-7 SISO arab 837.1 UDC 809.27-5 Subject heading: Arabic ; functional grammar
ISBN 90 6765 271 7 ® 1989 Foris Publications - Dordrecht No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission from the copyright owner. Printed in The Netherlands by ICG Printing, Dordrecht.
Contents
Foreword
VH
List of Abbreviations and Symbols
IX
Introduction Focus 1.1 Definition of the Focus Function 1.2 Focus Assignment 1.3 From Functional Structure to Constituent Structure: The Expression Rules 1.4 Conclusions
17 17 33
Chapter 2:
Topic 2.0 Introduction and Overview of the Data 2.1 Definition of Topic 2.2 Topic Assignment 2.3 The Nature of the Topic Term 2.4 Positioning of the Topic 2.5 Topic and Anaphora
69 69 71 72 74 76 96
Chapter 3:
Theme 3.1 Properties of Theme in Arabic 3.2 Theme versus other Functions 3.3 Conclusions
101 102 113
Tail 4.1 Definition of the Tail Function 4.2 Tail: 'External' or 'Internal' Function? 4.3 Tail and Reference 4.4 Constituency of Tail 4.5 Tail and Anaphora 4.6 Conclusions
123 124 126 129 131 133 136
Chapter 1:
Chapter 4:
51 66
121
VI
Chapter 5:
Vocative 5.1 Definition of Vocative 5.2 Assignment of Vocative 5.3 Constituency of Vocative Terms 5.4 Conclusions
139 140 141 143 151
Conclusion
153
References
155
Foreword
This book represents a critical application of Functional Grammar, as set out by Dik (1978), to the analysis of pragmatic functions in Modern Standard Arabic. Taking as his data a large number of sentential constructions, Prof. Moutaouakil puts forward a coherent set of proposals for the assignment of pragmatic functions to particular constituents of those constructions. This work will provide fundamental insights for readers interested in the syntax of Modern Standard Arabic and will also form a valuable basis for the study of textual relations in Arabic discourse. Prof. Moutaouakil indeed repeatedly stresses the relation between the pragmatic functions he recognizes and the discourse setting of the utterances to which they apply. The book derives added interest from the attempt to integrate, in a manner that will be mutually enriching, the views of the Arabic grammatical tradition and the claims of Functional Grammar. The transcription employed is 'morphemic' rather than 'phonetic'; for example, the preposition / / h a s always been transcribed with length-mark, even though the vowel is short before a consonant. Each data sentence has been provided with a partial morphemic gloss, which has been made no more detailed than necessary, and a translation into English. The gloss indicates at least the case of the nominal constituents (nom, acc or gen), and that of the clitic pronouns (A = accusative; G = genitive; no indication = nominative). All the data, also the ungrammatical and unacceptable sentences, have been translated; the translations are designed to give not only the meaning but also an idea of the structure and use of the data sentences. We wish to thank Drs. W. Raven, of the Department of Arabic, Free University of Amsterdam, for checking the transcriptions.
The editors, FGS
List of abbreviations and symbols
Semantic functions = Agent Ag = Goal Go = Recipient Ree = Zero 0 = Location Loe — Beneficiary Ben = Instrument Instr = Positioner Po Syntactic functions Subj = Subject Obj = Object Pragmatic Top Foe N-Foc C-Foc Voc Voc-H
functions = Topic = Focus = New Focus = Contrastive Focus = Vocative = Vocative of hailing
Categories V N A CO Q
= = = = =
Verb Noun Adjective Clause operator Question marker
Term operators d = definite i = indefinite DEM = demonstrative
χ Positions
s О V χ сор Ρ0 PI Ρ2 РЗ Ρ4
= = = = = = = = = =
Subject position Object position Verb position non-designated position copula position Top/Foc position CO position Theme position Tail position Vocative position
Cases
nom NOM
= nominative
gen GEN G
= genitive
acc ACC A
= accusative
Number
= singular = dual = plural
s dl pi Persan
(P)l (P)2 (P)3
= first person = second person = third person
General Xl, X2, ..
yi,
У2, ..
ψ
ω SFH HCD TWH CON
x„= y„= = = = = = =
argument variables satellite variables arbitrary predicate arbitrary term operator Semantic Function Hierarchy Hierarchy of Case Determination Topic-worthiness Hierarchy class of contexts
XI
A S
= addressee = speaker
Theories
FG TGG
= Functional Grammar = Transformational Generative Grammar
Introduction
The purpose of this study is to describe the properties of constituents in Modern Standard Arabic which bear one of the following pragmatic functions: Focus, Topic, Theme, Tail and Vocative. Modern Standard Arabic, essentially a written language, is the vehicle of culture in the Arab world: it is the language of contemporary Arabic literature, of the press and of education. Although Modern Standard Arabic is markedly different from 'Classical' Arabic, it continues to display the same fundamental structure, for largely cultural reasons. Indeed, the constructions to be discussed in this study will, for the greater part, be constructions common to both 'états de langue'. Those that no longer belong to Modern Standard Arabic or which survive only with a marked status will on each occasion be indicated as such. Although the specific object of this study is one 'état' of one language, it will also seek to account for the 'universal' aspects of the relevant constructions. In other words, this book is designed to capture the properties of such constructions in Natural Language generally. The constructions involving the pragmatic functions listed above have been described in various theoretical frameworks. The five functions, and indeed others, have been investigated in the framework of traditional Arabic linguistic thought, notably in the works on 'balaga' (approx. 'rhetoric'). The Arabic rhetoricians studied the interaction between the 'maqâl' (utterance) and the 'maqäm' (discourse environment) by considering the pragmatic functions of the constituents of utterances in different situational contexts and the structural (i.e. syntactic) properties that stemmed from these pragmatic functions. Thus, the constituent with the pragmatic function of 't-tahgig' (equivalent to 'Constrastive Focus') was seen generally to occupy initial position, as in the following sentence: (1)
qasldatan ?allaftu (la kitäban) poemacc wrote-Is not bookacc 'It was a poem I wrote (not a book)'
In contemporary linguistic theorizing, it has become standard practice to distinguish between a 'formal' and a 'functional' (or 'pragmatic') paradigm. To the former belong those linguistic theories (such as T G G ) in which
2 Pragmatic Functions in Functional Grammar of Arabic natural language is considered to be an abstract system describable independently of its communicative function, while the latter paradigm covers those theories in which natural language is conceived of as a structure whose properties are at least partially determined by the conditions under which it is used. Exponents of the functional paradigm include 'pragmantax', one development of generative semantics; the 'functionalism' of the Harvard school; and the European functionalist theories, such as the 'systemics' of the London school, 'Functional Sentence Perspective', and 'Functional Grammar' as proposed in recent years by Simon C. Dik and his co-workers. In the linguistic theories flowing from the formal paradigm, pragmatic functions are either ignored, as one might expect, or are considered to be semantic or syntactic notions and are thus given a purely formal treatment.' In the functional paradigm, however, pragmatic functions are viewed as being associated with constituents in accordance with given situational conditions. Of the pragmatic theories currently available. Functional Grammar is theoretically the most satisfactory (see below) and is organized in such a way as to provide not only the best account of pragmatic functions but also the clearest statement of their interactions with semantic and syntactic functions and with formal expression. This study of pragmatic functions in Arabic will therefore be cast in the framework of Functional Grammar (henceforth FG). Nevertheless, since the fíve functions that we intend to investigate have been the object of interesting descriptions in the Arabic grammatical tradition (henceforth the Tradition), I will often have occasion to borrow certain specific analyses, and where necessary even certain concepts from the traditional Arabic linguists. In Moutaouakil (1982), I set out in broad outline a methodology for re-reading these linguists, integrating the theoretical system they employed into contemporary linguistic thinking and exploiting their work in the description of not only Arabic and related languages but also natural language as a whole. I was able to show that the theory underlying traditional Arabic linguistic thinking, despite the apparent disparity of the disciplines ('grammar', 'rhetoric', 'lexicology', 'exegesis', ...), is a pragmatic theory worthy of consideration in the form of a mutual exchange of analyses and possibly also concepts with contemporary pragmatic theories, including FG. The analyses that I shall have occasion to borrow from the Arabic grammatical tradition will enrich FG without affecting the methodological principles of the theory or the organization of the grammar. On the other hand, the comparison of FG and traditional Arabic linguistic thought will permit a re-examination and re-evaluation of certain analyses that have remained unquestioned. Thus, a number of analyses of relevant constructions will be challenged, either partially or in their entirety.
Introduction
3
The fundamental methodological principles of FG^ are as follows: a) The basic function of language is that of an instrument of communication; b) The purpose of linguistics is the description of the speaker's 'communicative competence'; c) Linguistic description should achieve pragmatic, psychological and typological adequacy; It follows from principle a) that FG seeks to describe natural languages from a functional viewpoint, i.e. in such a way that their structural properties are at least partially explicable - both synchronically and diachronically - in terms of the various communicative tasks that language is called upon to carry out in verbal interaction. Principle b) entails a redefinition of the classical dichotomy between competence and performance. The speaker's competence (his knowledge of his language) is a 'communicative competence' in that it is seen as knowledge not only of the linguistic rules but also of the pragmatic rules allowing the speaker to 'perform' in given situation-types with a view to specific communicative goals. Principle c) requires of FG that it should aim for three types of adequacy. In order to meet the requirement of 'pragmatic adequacy', FG provides for an autonomous level of representation for pragmatic functions (Theme, Topic, Focus, ...) alongside levels for syntactic functions (Subject, Object) and semantic functions (Agent, Goal, Instrument, ...). 'Psychological adequacy' is achieved by conforming as closely as possible to psychological models, both of production and of comprehension. As a result, rules adjudged to lack 'psychological reality' are banned^ thus, grammars written in an FG framework will lack any transformational rules'*. 'Typological adequacy', finally, is a requirement placed upon the grammatical description of an individual language to accord as closely as possible with what is known of the universal properties of natural language. The functions relating to the three levels (semantic, syntactic and pragmatic) are in FG considered to be 'primitive' notions in the sense that they are not derived from pre-existent configurational structures'. They play a fundamental part in the organization of the grammar. In contrast to generative grammars of a 'configurational' type - but in parallel to Relational Grammar or Lexical-Functional Theory - the rules 'building' constituent structure operate on the basis of information provided by functional structure and not vice versa. The organization of a Functional Grammar is built around three structures: predicational structure, functional structure and constituent structure·^ these three structures are constructed by three systems of rules: the
4
Pragmatic Functions in Functional Grammar of Arabic
rules in the fund, the function assignment rules and the expression rules. The f u n d subsumes two subsystems which work together to create the predicational structure represented as a 'predicate frame': these are the lexicon and the predicate- and term-formation rules. On the hypothesis that the lexical items of every natural language are either basic (i.e. learnt as such by speakers before they can be used) or derived (i.e. formed by 'synchronically productive' rule from basic lexical items)', the lexicon contains the basic predicate frames and the basic terms, while the formation rules create the derived predicate frames and the derived terms (cf. Figure 1 below). A predicate frame is represented as a structure comprising a predicate (an expression designating a property or relation) and its argument(s). Each predicate frame specifies ö) the form of the predicate; (ii) its syntactic category (V(erb), N(oun) or A(djective)); (iii) the argument positions associated with the predicate, indicated by variables (xi, хг, ..., Xn); (iv) the semantic functions associated with each argument position (Agent, Goal, Instrument,...); (v) the selection restrictions imposed by the predicate on these positions. For instance, the predicate ?a'tâ 'gave' (see note 9) is specified in predicate frame (2) as a verbal predicate (subscript V) taking three argument positions with the semantic functions Agent, Goal and Recipient and the selection restrictions 'animate', 'non-human' and 'animate' respectively: (2)
?a''täv (хь animate(xi))Ag (xi: non-human(x2))Go (хз: animate(x3))Rec
The predicate farih 'merry' is specified as an adjectival predicate (subscript A) taking one argument position with the selection restriction 'animate' and the semantic function Zero (0), which generally indicates the role of the primary participant in a predicate frame designating a State: (3)
farih A (xi: animate(x,))0
Predicates frames are held to designate a 'state of affairs' in which the participants play various roles (reflected in the semantic functions). The state of affairs may be an Action, Process, Position or State®. The participants, depending on how essential they are for the definition of the state of affairs, are represented either as arguments or satellites. Thus, positions (xi), (X2) and (хз) in predicate frame (4) are argument positions, since they are necessary for the definition of the Action designated by the predicate îa'tâ, whereas positions (yi) and (уг) are satellite positions, since they serve merely to specify the circumstances of the Action, namely the time and the place:
Introduction (4)
5
îa'tâv (xi: animate(xi))Ag (хг: поп-Ьитап(х2))оо (хз: атта1е(хз))кес (yi)Time (y2)Loc
As mentioned above, there are also derived predicate frames, created by predicate formation rules.' I will adopt the hypothesis that the basic predicates are those formed on one of the patterns fa Ы, fa V, fa Ы or falal, plus what the Arabic grammatical tradition refers to as 'jämid' (i.e. lexical forms that are neither derived nor can be a source for the derivation of other forms). All other predicates will be regarded as derived, either directly (e.g. the patterns ?afal,fäbl and ftabl or indirectly (e.g. tafâbl and tafani, which are derived from fäbl and fa "al predicates which are themselves derived). Like predicates, terms may be either basic or derived. Basic terms are those expressions which can only be used as terms, such as personal pronouns, interrogative pronouns, proper names, etc. Accordingly, they are listed as such in the lexicon. The majority of terms are derived, i.e. formed by means of term formation rules which create the following structure: (5)
(œxi: v3i(xi): - ! - > + >
Ree + > + >
Ben Instr + > + > + > + >
Loc Time + > + + > +
The SFH implies that a term with the semantic function Agent is the primary candidate for the syntactic function Subject and that a term with the semantic function Goal is the primary candidate for the syntactic function Object. It also implies that the assignment of Subject and Object becomes increasingly difficult as one progresses down the hierarchy. In other words, sentences in which Subject is assigned to a Goal term are more marked than sentences in which it is assigned to an Agent term; sentences in which Subject is assigned to a Recipient term are more marked than those in which it is assigned to a Goal term, and so on. The same applies to sentences in which Object function is assigned to terms other than the G o a l ' \ In FG, the following pragmatic functions are recognized: Focus, Topic, Theme and Tail. The first two are said to be 'internal' in that they are associated with terms of the predication proper; the functions Theme and Tail are said to be 'external', being associated with constituents situated outside the predication proper. The facts of the grammar of Arabic (and indeed of several other natural languages) suggest that the number of pragmatic functions allowed in F G is insufficient. In particular, it will be necessary to distinguish two types of Focus, New Focus and Constrastive Focus, as illustrated in (8) and (9) respectively:
Introduction
7
(8)
""udtu min s-safari laylan returned-Is from the-journey nightacc Ί returned from the journey at night'
(9)
Hindan kallamtu (la HâHdan) Hindacc s p o k e - I s
n o t Hälidacc
'It was Hind I spoke to (not Hâlid)' In chapter 1, I shall provide several arguments supporting the distinction between the two types of Focus". It will also be necessary to add to the four pragmatic functions allowed in FG a fifth function, Vocative, as illustrated in (10-11): (10)
?ayyuhâ l-musäfirüna, sta'iddü О the-travellersnom prepare-2pl 'Travellers, be ready'
(И)
Zaydu, lä tatagayyab katïran Zaydnom not be-absent muchLacc 'Zayd, don't be absent so often'
In Chapter 5,1 shall argue against confusing the 'speech-act' of 'addressing' and 'Vocative' as a pragmatic function. Like Theme and Tail, Vocative will be regarded as an 'external' function. Pragmatic functions are assigned in accordance with specific contextual conditions. Unlike semantic and syntactic functions, pragmatic function assignment is determined by the ('linguistic' and situational) context in which the utterance is produced. For each of the five pragmatic functions to be examined, I shall specify the contextual constraint applying to the assignment of all three types of function (semantic, syntactic and pragmatic) and specific constraints on the assignment of each of the pragmatic functions in turn. The general constraint, which, with Bresnan (1980), I shall call 'the bi-uniqueness constraint', will be formulated as follows: (12)
Bi-uniqueness constraint The terms in a predication are labelled for semantic, syntactic and pragmatic functions in such a way that: (i) no term may have more than one function at each of the three functional levels; (ii) no function may be assigned to more than one term.
The second clause of this constraint, however, does not apply to pragmatic function assignment, in that it is possible to assign one and the same
8 Pragmatic Functions in Functional Grammar of Arabic pragmatic function to more than one element in a sentence. For example, in (13), the function New Focus is assigned to two constituents (Fahâhu and l-yawma)·, in (14), Topic function is also assigned to two constituents (Zaydun and Halidan): (13)
zara Zaydun (Top) ?ahähu (New Focus) l-yawma (New Focus) visited Zaydnom brotheracc-3sG todayacc 'Zayd visited his brother today'
(14)
wahaba Zaydun (Top) Halidan (Top) mälahu (New Focus) donated Zaydnom Hâlidacc moneyacc-3sG 'Zayd donated his money to Hâlid'
As for the constraints on the assignment of individual pragmatic functions, I shall investigate, for each of the five functions, the constraints on the number of constituents that may carry that function, which terms are most likely to carry the function, which terms cannot carry it, etc. As an example of the assignment of syntactic and pragmatic functions, consider again predicational structure (6). Syntactic function assignment may attach Subj and Obj to terms (xi) and (хз) respectively, yielding structure (15): (15)
?a'tâv (dxi: 'Amr(xi))AgSubj (dx2: majallat(x2))Go (dxs: Zayd(x3))RecObj (dyi: bariljat(yi))rime (dy2:maktabat(y2))Loc
Pragmatic function assignment, under appropriate contextual conditions, will attach to (xi) and (хг) Topic and New Focus respectively, giving structure (16), which will be expressed as (17): (16)
?a'tâv (dxi: 'Amr(xi))AgSubjTop (dx2: majallat(x2))GoN-Foc (dxs: Zayd(x3))RecObj (dyi: bâriljat(yi))Time (dy2: maktabat(y2))Loc
(17)
?a°ta "Amrun Zaydan 1-majallata 1-bäriljata fî gave "Атгпот Zayda« the-magazineacc yesterday in 1-maktabati the-librarygcn '"Amr yesterday gave Zayd the magazine in the library'
Constituent structure is constructed by expression rules operating on the information contained in functional structure. These rules determine (i) the case-form of constituents, given their functional properties in the predication; (ii) the form of the predicate (voice, auxiliaries, agreement, ...); (iii) constituent order; (iv) stress and intonation.
Introduction
9
Cases are assigned to constituents in accordance with their semantic, syntactic or pragmatic function". The interaction of the three function types may be formulated as follows: (i)
When a constituent only bears a semantic function, it takes the case conferred upon it by that semantic function; (ii) When a constituent is assigned a syntactic function (Subject or Object), it takes the case conferred upon it by that syntactic function, which thus 'masks'" its semantic function; (iii) When a constituent is assigned a pragmatic function, there are two possibilities: it may either be a term inside the predication proper (i.e. Topic or Focus) or a term outside the predication proper (i.e. Theme, Tail or Vocative). In the former case, it will have a semantic function and possibly also a syntactic function and then takes the case conferred upon it by its semantic or, if applicable, syntactic function. In the latter case, it carries neither a semantic nor a syntactic function, and consequently takes the case conferred upon it by its pragmatic function. The interaction of functions in co-determining cases may be formulated as the following hierarchy: (18)
Hierarchy of Case Determination (HCD) syntactic functions > semantic functions > pragmatic functions
Note that a distinction must be made between abstract case and surface morphological realization. The case assignment rules assign an abstract case (i.e. Nominative, Accusative, Genitive, etc.), which, in inflecting languages like Arabic, will be realized as case markeгs'^ Let us illustrate case assignment with relation to functional structure (16). The output of the assignment of cases will be as follows: (19)
Pa'^tâ (dxi:
'Amr(xi))AgSubjTop
(dx2: majallat(x2))
\ NOM (dxs: Zayd(x3))RccObj
/
ACC
GoN-Foc
\ ACC
(dyi: barihat(yi))Time (dyi: maktabat(y2))Loc
/
ACC
/
GEN
Constituent ordering is viewed in FG as being co-determined by (i) syntactic functions, (ii) pragmatic functions, (iii) the categorial complexity of constituents. As to whether semantic functions also co-determine the
10 Pragmatic Functions in Functional Grammar of Arabic order of constituents, I shall adopt the hypothesis put forward in the framework of the Arabic grammatical tradition that the hierarchical relationship among the semantic roles of constituents is reflected in the order in which these are linearized in the sentence. The hierarchy in question may be formulated as follows: Ag > Ree > Go > Time > Loc > Manner > Cause ... According to the Tradition, the 'natural' (i.e. 'canonical') order of constituents is that which conforms to this hierarchy of semantic roles. It remains to be established, of course, how this hierarchy of semantic functions interacts with the other factors in determining the order of constituents - if the hypothesis put forward by the Tradition is borne out by future research. In order to account for constituent order in Arabic, I propose the following pattern: (20)
(P4), P2, PI P0 V S N/A О X, P3, (P4)
This pattern distinguishes two types of position: 'external' positions (P2, P3 and P4) reserved for constituents which carry 'external' pragmatic functions and which are consequently not part of the predication proper; and 'internal' positions for arguments and satellites belonging to the predication. Positions P2 and P3 are occupied by Theme and Tail respectively: thus, in (21), the constituent C^Amrun), a Theme, is in P2 and, in (22), the constituent (''ilmuhu), a Tail, is in P3: (21)
''Amrun, ?abühu musâfirun "Атгпот fathernom-3sG travelling„om 'As for "Amr, his father is travelling'
(22)
baharanï Zaydun, 'ilmuhu surprised-Is A Zaydnom knowledgenom-3sG 'Zayd surprised me, his knowledge' 'Zayd's knowledge surprised me'
The third external position, P4, is reserved for a constituent with the pragmatic function Vocative: P4 may come either before P2 or after P3. Thus, Theme, Tail and Vocative are ordered as follows with respect to the Predication: (23)
(Vocative), Theme, Predication, Tail, (Vocative)
Introduction
11
In pattern (20), positions PI, P0, S, O, X and N / A are intended for, respectively: a clause operator (CO); a constituent with the pragmatic function Topic or Focus; a Subject; an Object; a constituent with neither a syntactic nor a pragmatic function capable of conferring a special position on it; and a nominal or adjectival predicate in a non-verbal sentence. Constituents are allocated to the positions corresponding with their functions by placement rules. The following general formulation relates to the 'internal' positions:" (24)
Rl. R2. R3.
CO Focus Subj
-
PI
^ P0 -
S, Obj - O, Verb - V
where the arrow indicates 'is placed in' Recall that the placement rules, which operate on an unordered structure, are not transformational rules. They are placement rules, not displacement or movement rules of the type proposed in the framework of transformational grammars^". For each of the pragmatic functions to be considered in this book, I shall seek to specify which placement rules are required and what constraints are to be placed on the operation of these rules. Let us exemplify the application of placement rules with reference to (19), a functional structure with case-specifications. Terms (xi) and (хз) are placed in positions S and О in accordance with R3. Term (хг), however, is not placed in P0 by R2, even though it is has Focus: in Chapter 1, we shall see that only Contrastive Focus is placed in P0. In other words, a term with New Focus such as (хг) in (19) occupies the position corresponding with its semantic (or, if it has one, syntactic) function. The result of the operation of these placement rules may be seen in (17) above. The organization of a Functional Grammar may be displayed as follows:
12 Pragmatic Functions in Functional Grammar of Arabic Figure 1. FUND Lexicon predicate frames
predicate formation
derived predicate frames
terms
basic predicate frames
basic terms
derived terms
term formation
extension of predicate frames i = term insertion t PREDICATIONAL STRUCTURE
{
syntactic function assignment rules
I
pragmatic function assignment rules Í FUNCTIONAL STRUCTURE expression rules form of constituents order of constituents stress and intonation
1
CONSTITUENT STRUCTURE ( = prephonological representation)
In the five chapters that now follow, each of the pragmatic functions in Arabic will be studied with reference to the properties of the constituents to which they are assigned. The first two chapters will deal with the 'internal' functions, Focus and Topic respectively. The remaining chapters will deal with the functions Theme, Tail and Vocative, in that order. The discussion of each of the five functions will follow the major stages in the generation
Introduction
13
of a sentence in FG. After the definition of the function in question, there will be an examination of the rules governing the assignment of the function and the constraints to which these are subject, followed by a study of the referential, case-marking, positional and, in some cases, anaphoric properties of the constituent(s) to which the function is assigned.
NOTES 1. In the 'Extended Standard Theory', the Focus function is regarded as a semantic notion and is consequently treated as an aspect of semantic interpretation (cf. especially Jackendoff 1972). The phenomena of 'topicalization' and 'left/right dislocation' are analysed from a purely syntactic angle. The work of Fassi-Fehri on different sentence-types in Arabic is carried out in 'Lexical-Functional Theory', which also falls under the formal paradigm (cf. in particular Fassi-Fehri 1982, which will often be a source of inspiration in this book). 2. The following presentation of F G will concentrate on the essential properties of the theory. It will itself be 'functional' in the sense that it will contain no more information than is necessary for an understanding of the chapters dealing with the five functions to be examined. For a fuller account of the methodological principles of F G and of the organization of a Functional Grammar, the reader should consult Dik (1978), (1980a) and (1980b). Detailed aspects of the various rules brought into play in the generation of sentences in F G will emerge from the specific analyses to be presented. 3. See, for the concept of 'psychological reality', Bresnan (1978). 4. For the same reason, lexical items are not represented at any level in terms of primitive semantic concepts. The defining predicates in F G belong to the language that is being described (see Dik 1978: 12). 5. In this respect, F G resembles 'Relational Grammar' and 'Lexical-Functional Theory', in which 'grammatical functions' like Subject, Object, Indirect Object, ... are postulated as primitive notions in opposition to the models proposed by Chomsky, who continues to define these functions in terms of configurations. 6. The standard names for these three structures in F G are 'predication', 'functionally specified predication' and 'syntactic form' respectively. My choice of more suggestive labels (^oes not imply any modification of the three concepts. 7. For the definition of the notion of a 'synchronically productive rule', see Dik (1980b: 26-28). 8. See, for more detail on the typology of 'states of affairs' and the semantic functions associated with each type, Dik (1978: Ch. 3). 9. In Arabic, as in the Semitic languages generally, words are formed not by the affixation of morphemes (i.e. prefixes, suffixes and infixes) to a root, but by the attribution of a 'scheme' (consisting of vowels) to a 'root' (generally consisting of three consonants). Thus, for example, kataba 'write', kätaba 'correspond with', kâtib 'writer', kitâba 'writing', etc. are formed by attributing to the tri-consonantal root k.t.b. the schemes fa'ala, fä'ala, fäll and fì'àla respectively. Since matters of tense are not of immediate importance in this book, verbal predicates will be given not as roots but, for ease of presentation, as tensed forms resulting from the application of a scheme to a root. 10. For more detailed information on term structure, see Dik (1978: Ch. 4). 11. One of the arguments that may be advanced for applying syntactic function assignment before pragmatic function assignment is that the assignment of certain pragmatic functions is partially determined by the syntactic function of the relevant term. Thus, there is a tendency
14
Pragmatic Functions in Functional Grammar of Arabic
for Topic to be assigned to a term to which Subject has already been assigned (cf. Dik 1978: 178-181). 12. Arguments have been put forward for the introduction of a third syntactic function. Secondary Object: cf. Kanno (1983). 13. Justification for assuming the presence of an Object function in Arabic is presented in Moutaouakil (1985b), where it is shown that Object is assigned to Go, Ree, Loe and Temp, and is crucially involved in causative and object-raising constructions. 14. This links up with the traditional Arabic grammarians' claim that the subject function of passive constructions is assigned to the constituent with object function. The assignment of this function to other constituents is regarded by these grammarians as an 'extension' ('rawassu"). Indeed, sentences (ii) and (iii) are marked as compared with (i): (i)
(ii)
(iii)
du'iya Zaydun (Go) was-invited Zayd„om 'Zayd was invited' jîma yawmu l-jumu^ati (Time) was-fasted day nom the-Fridaygcn O n e fasted on Friday' mukita fl d-dâri (Loe) was-stayed in the-housCgcn 'They (etc) stayed at home'
Moreover, the assignment of Subject function in passive constructions to constituents other than the object (in the sense of Goal) is liable to increasing constraints. Thus, the Subject of such constructions may be definite or indefinite if it is Goal, but is obligatorily definite if it is a Locative: (iv)
a.
b.
duriba Zaydun (Go) was-beaten Zaydnom 'Zayd was beaten' duriba rajulun (Go) was-beaten
man„om
Ά man was beaten' (v)
a.
b.
ruqija ñ d-däri (Loe) was-danced in the-housegt„ 'There was dancing in the house' »ruqija ñ darin (Loe) was-danced in housegen 'There was dancing in a house'
15. See, for a typology of different kinds of Focus, Dik et al. (1981). 16. One of the principles governing the assignment of cases in natural languages is that one and the same constituent cannot receive two cases at once. 17. In certain languages (such as Japanese), however, some pragmatic functions, among them Topic, confer on the constituents with which they are associated a case that 'masks' their semantic and even their syntactic function (cf. Dik 1980b: 18-19). 18. Further evidence for this distinction may be derived from (i) the fact that it is possible not to realize the abstract case (cf. the nouns referred to in Arabic as ' m a q j ü r ' and 'manqûf'); (ii) the fact that the 'surface' case-marking of a constituent in certain constructions (e.g. certain Vocative constructions in Arabic) does not correspond with the case assigned to the constituent in question at the underlying functional structure level.
Introduction
15
19. These rules are based on those formulated by Dik (1980b: 21). 20. To take a simple example, the constituent t-tuffâlfata in (vi) is placed clause-initially, according to the transformational models, by a movement rule (called 'Topicalization') which operates on the ordered structure (vii): (vi)
t-tuffâljata ?akaltu the-appkacc ate-Is 'It was the apple I ate (, and not...)'
(vii)
?akaltu t-tuffähata Ί ate the apple'
In FG, the constituent t-tuffäijata occupies clause-initial position (i.e. position P0) as a result of a placement rule operating on the unordered structure (viii): (viii)
?akalav (dxi: pl(xi))A|SubjTop (dx2: tuffäljat(x2))G
P0
->
P0
opt
obi = obligatorily
Top
(R3)
clause operator
(R4)
Subj
opt = optionally PI
->
obi
Top
->
Foe
Rules (Rl, R2 and R4) are subject to the Single Occupancy Constraint and (R2 and R4), where the constituent in question is Topic, to the Referentiality Constraint. 2.4.2.3. Topic placement in copular sentences 2.4.2.3.1. Formation of copular sentences We saw above that copular non-verbal sentences are distinguished from their non-copular counterparts by the presence of the copula kâna (or any of the forms in its paradigm) inserted by the rule of Copula Support. The workings of this rale were illustrated in 2.4.2.1 above. Note that Copula Support applies before the case-marking rales, since the case of the predicate is dependent upon whether copula insertion takes place: in non-copular non-verbal sentences, the predicate takes the nominative case, but in copular non-verbal sentences, the accusative case. Thus, in (54a), repeated here for convenience, the case-marking rules assign nominative case to the Subject and accusative case to the predicate: (54a)
kâna Zaydun musafìran was Zaydnom travellinga cc 'Zayd was travelling'
As to the order of constituents, I hypothesize that the positional pattern for copular sentences is as follows: (86)
P2,
PI
P0
cop
S
N/A
О
X ,
P3
In keeping with this pattern, the constituents of (54a) are placed as follows: the copula käna in cop position, the Subject in S and the predicate in its normal position. After the operation of these placement rales come the rules for stress and intonation, which yield a prephonological representation of the sentence.
Topic
93
2.4.2.3.2. Positions of the Topic Copular sentences share certain properties with non-copular non-verbal sentences and certain properties with verbal sentences. They resemble the former in their predicational and functional structures and the latter in their constituent structure. Considering constituent structure more closely, we may observe that the same positions as are required to account for the order of constituents in verbal sentences return, in the same order - with the addition of the cop slot - in copular sentences (compare (28) and (86)). Accordingly, Topic placement in copular sentences is governed by the same rules as apply in verbal sentences. Just as in verbal sentences, the Topic of a copular sentence occupies either the position corresponding to its semantic (or, if relevant, syntactic) function or the position P0. In sentences (87) and (88) Topic occupies S and X respectively, and in sentences (89), it is placed inP0: (87)
käna Zaydun (Top) Sujä^an was Zaydnom bravcacc 'Zayd was brave'
(88)
käna "Amrun fî d-dâri (Top) was "Лтгпот in the-houscgen '"Amr was at home'
(89)
a.
l-bäriljata (Top) käna Zaydun fî yesterdayacc was Zayd„om in 'Yesterday Zayd was at home' b. fî 1-maktabi (Top) käna Zaydun in the-officegcn was Zaydnom 'In the office Zayd was waiting for
d-dâri the-housCgen munta^iran 'Amran waitingacc 'Amr,acc "Amr'
The Single Occupancy Constraint also applies to the P0-placement of a Topic in a copular sentence; in other words, a Topic can be placed in P0 only if this position is not occupied by another Topic or by a Contrastive Focus, witness (90): (90)
a. *l-bäriljata (Top) fî d-däri (Top) käna Zaydun munta?iran yesterdayacc in the-houscgen was Zayd nom Waitmgacc 'Yesterday at home Zayd was waiting' b. *l-bärilfata (C-Foc) fî d-däri (Top) käna Zaydun munta?iran It was yesterday, at home, that Zayd was waiting'
94 Pragmatic Functions in Functional Grammar of Arabic P0-placement of a Topic is furthermore subject to the Referentiality Constraint; (91) are ungrammatical because a non-referential Topic has been placed in P0: (91)
a. *fì darin käna Zaydun munta?iran in housegen was Zaydnom waitingacc 'In a house, Zayd was waiting' b. *fi laylatin kâna Zaydun munta^iran in nightgen was Zayd nom waitingacc O n e night Zayd was waiting'
In sentences in which Topic function is assigned to the Subject, the SubjTop must appear in S position, as in (54a), repeated here for convenience: (54a)
käna Zaydun (SubjTop) musäfiran yvas Zaydnom travelling,cc 'Zayd was travelling'
If a 'Subject' appears to occur before the copula, it must be in P2 and therefore is a Theme rather than a Subject; the following sentence can only be an example of the Theme, Predication construction: (92)
Zaydun, käna musäfiran Zaydnom was-3s travellingacc 'Zayd, he was travelling'
This is confirmed by the fact that the term in question can occur before clause operators, which always occupy the absolute first position in the predication: (93)
a.
b.
Zaydun, ?a käna musäfiran Zaydnom Q was-3s travellingacc 'Zayd, was he travelling?' Zaydun, ma käna musäfiran Zaydnom NEG was-3s travelling.cc 'Zayd, he wasn't travelling'
As in non-copular sentences but in contrast to verbal sentences, the Subject of copular sentences may be referential only if it is also Topic, witness the ungrammaticality of (94) as against (95): (94)
*käna rajulun marii^an was шаппот üUcc Ά man was ill'
Topic (95)
95 a.
b.
kâna г-rajulu тап(^ап was the-таппот ilLc 'The man was ill' kâna Hâlidun mañean was Hälidnom ilUcc 'Hälid was ill'
If the Subject of a copular sentence is assigned Contrastive Focus, it may be a non-referential expression; thus (96) is perfectly grammatical if the Subject is interpreted as Focus: (96)
kâna rajulun fî d-däri (là rajulâni/ mraPaturí) was таппога in the-housegen nottwo:men„om/womannom 'It was one/a man that was in the house (not two men/a woman)'
As in verbal sentences, the P0 placement of Topics is easier for some constituents than for others. The predicate, notably, prefers to occupy its normal position: (97)
a. ??fî 1-maqhâ (Top) kâna Zaydun in the-cafégcn was Zaydnom 'In the café was Zayd' b. ??hazînan (Top) kâna Hâlidun sad acc was Hâlidnom 'One person who was sad was Hâlid'
The sentences under (97) are unexceptionable, however, if they are interpreted as having a Contrastive Focus constituent in P0-position. This is confirmed by the naturalness with which expansions of the type used to test for a Focus interpretation can be added to these sentences: (98)
a.
fî l-maqhâ {C-Έoc) kâna Zaydun (lâ fî d-dân) in the-cafégen was Zayd„om not in the-houscgen 'It was in the café that Zayd was (not at home)' b. hazlnan (C-Foc) kâna Hâlidun (lâ fariffan) sadacc was Hâludnom not happyacc 'Hâlid was sad (,not happy)'
96 Pragmatic Functions in Functional Grammar of Arabic 2.5. TOPIC A N D A N A P H O R A
2.5.1. Pronominal binding vs. positional binding In FG, constituents placed in one of the special positions P0 and - to a certain extent - P2 are regarded as being linked by anaphora to a position or pronoun within the predication. In other words, constituents in P0 (and P2) may be said to 'bind' a position or pronoun in the predication. I shall refer to the two types of binding as 'positional binding' and 'pronominal binding' respectively. 2.5.2. Positional binding Constituents placed in P0 - interrogative pronouns, constituents with Contrastive Focus or Topic - bind a position inside the predication. This position is the one that the constituent would have occupied, by virtue of its semantic or syntactic function, had it not been placed in P0. Thus, in sentences (99), the constituents mädä, Zaydan and f l l-maqhâ, all of which occur in P0 because of their pragmatic function, bind the postverbal positions (indicated as (0i)) that they otherwise would have occupied: (99)
a.
mâdâi (Foc) Saribta (0i)GoObj what drank;-2s 'What did you drink?' b. Zaydan, (Foe) qäbaltu (0i)GoObj Zaydacc met-Is 'It was Zayd I met' c. fî 1-maqhäi (Top) qäbaltu 'Amran in the-cafégen met-Is "Атгасс 'In the café I met °Amr'
(0Í)LOC
When a Topic is placed in P0, the binding is usually positional, witness: (100)
a. fî S-ääri'ii (Top) qäbaltu Hälidan (0Í)LOC in the-streetgen met-Is Hâlidacc 'In the street I met Hälid' b. fì gabàbi gadini(Top) Zaydun musäfirun (0i)Time in morning tomorrowgen Zaydnom travellingnom 'Tomorrow morning Zayd will set out'
There are certain construction types, however, in which a P0-placed Topic does not bind a position but a pronoun, as we shall see in the following section.
Topic
97
2.5.3. Pronominal binding Pronominal binding is characteristic of those Theme, Predication constructions in which there is an anaphoric pronoun in the Predication referring back to the Theme, cf. (101): (101)
"Amruni (Theme), sâ'adahui Hälidun "Атгпош helped-3sA Hâlid„om '^Amr, Hälid helped him'
In Classical Arabic, however, a Topic placed in P0 could also bind a pronoun in certain types of construction, as in (9), repeated here for convenience: (9a)
Zaydan qäbaltuhu Zaydacc met-ls-3sA Ί have met Zayd'
(9b)
1-kitäba qara'tuhu the-bookacc read-ls-3sA Ί have read the book'
The following explanation may be put forward as motivating the presence of such constructions in Classical Arabic. As we saw in the preceding section, there are certain constituents that cannot occur naturally in P0 unless they either are interrogative pronouns or carry Contrastive Focus; sentences in which such constituents occupy P0 and have Topic function are, it will be recalled, of dubious grammaticality. However, these sentences become perfectly grammatical if, occupying the postverbal position that the Topic otherwise would have filled, there is an anaphoric pronoun referring back to the Topic in P0. Thus sentences (102), which are difficult to interpret with an initial constituent that is not Contrastive Focus but Topic, become entirely natural, under the same interpretation, if the P0constituent is picked up again by an anaphoric pronoun, as in (103): (102)
a. ??Zaydan (Top) qâbala ""Amrun Zaydacc met АтГпот '"Amr met Zayd' b. ??°Amran (Top) ?a'ta Zaydun 1-kitâba "Атгасс gave Zaydnom the-bookacc 'Zayd gave "Amr the book'
98
Pragmatic Functions in Functional Grammar of Arabic
(103)
a. b.
Zaydan, qäbalahu, "Amrun '^Amrani ?a°tähui Zaydun 1-kitâba
On this basis we may hypothesize that the presence of the anaphoric pronoun in this type of construction represents a strategy allowing the speaker to position in P0 Topic constituents that it otherwise would be difficult to place in a special position. 2.5.4. Constraints on binding The 'island constraints' on the P0-placement of Focus constituents discussed in 1.3.2.4 above apply in equal measure to the P0-placement of Topic constituents; in the context of the present section, these constraints may be regarded as constraints on the relation between the constituent in P0 and the position it binds in the predication. Consider, as examples of violations of the Complex NP Constraint, the Co-ordinate Structure Constraint and the A-over-A Constraint respectively, (104), (105) and (106): (104) *Hindani (Top) qabaltu r-rajula 1-ladî Paljabba (0i)Goobj Hindacc met-Is the-тапасс the-one liked 'Hind I met the man who liked' (105) *bi "Amrini (Top) marra Zaydun bi Hâlidin (0i) with "Amrgen met:up Zaydnom with Hälidgen and 'With 'Amr Zayd met up with Hälid and' (106) *R-Ribati (Top) zara Zaydun kulliyyata (0i) Rabatgen visited Zayd nom UniVCrSltyacc 'Of Rabat Zayd visited the University'
NOTES 1. See, for the tendency for Topic function to be preferentially assigned to Subject, Dik (1978: 143). 2. For a consideration of further aspects of Topic placement in Arabic, see also Moutaouakil (1987). 3. See, for a pragmatic approach to reference, Dik (1978: 55-56). 4. This constraint, which we shall call the 'Single Occupancy Constraint', may be formulated as follows: Single Occupancy Constraint Any one position may be occupied by only one constituent. 5.
The following sentences are examples of non-verbal sentences with an Adverb as predicate:
Topic
99
(i)
1-muqâbalatu 1-yawma the-meeting„om the-daya^ 'The meeting is today'
(ii)
d-dahâbu gadan the-departure„„m tomorrowa« 'The departure is tomorrow'
6. According to the Arabic grammatical tradition, the constituent occupying initial position in this type of construction has to be 'definite'. It can, however, be indefinite provided it is preceded by an interrogative particle or a negative particle, as in the following sentences: (iii)
?a rajulun й d-dàri Q man„om in the-housege„ 'Is it a man that is at home?'
(iv)
mä rajulun fi d-däri NEG man„om in the-housCgen 'It is not a man that's at home'
The traditional grammarians should have generalized their analysis of this type of sentence to sentences of the type exemplified by (70) and (71) since in these two cases the constituent in question is not a Topic but a Focus, and consequently need not be a referential expression. 7. The predicate of non-verbal sentences, both copular and non-copular, is assigned a pragmatic function (either Topic or Focus). Dik (1980b: 105) postulates that adpositional predicates in sentences of the following type have a semantic function: (v)
Zaydun fi d-däri Zayd„om in the-housege„ 'Zayd is at home'
However, he does not ascribe a syntactic function to non-verbal predicates, be they nominal, adjectival or adpositional. This poses a serious problem for case-marking in Arabic, since all non-verbal predicates are marked for case, witness: (vi)
'Amrun Ijazlnun •^Атгпои sad„om '^Amr is sad'
(vii)
Zaydun lugawiyyun Zayd„