Coordinative constructions in Arabic: aspects of morphosyntax as an indicator of interpersonal status.
In current linguistic theories, the complexity of linguistic expressions and the dependencies (or lack thereof) between their constituents are generally defined, as is well-known, on the basis of purely formal criteria: a complex expression is an expression which contains more than one clause; a clause member in a complex expression is said to be independent if it is merely juxtaposed or coordinated and dependent if it is linked to the 'matrix clause' by means of an overt formal subordinator.
However, when complexity and (in)dependence are considered from a Functional Discourse Grammar point of view, two crucial facts emerge: (a) a distinction must be made between interpersonally and structurally complex constructions on the one hand and between interpersonally and structurally (in)dependent constituents on the other, and (b) more often than not, there is a discrepancy between interpersonal and structural status in the sense that no parallelism holds between the two kinds of status.
Elaborating on the description of the interpersonal level given in (HENGEVELD; MACKENZIE, 2006), I will discuss some morpho-syntactic means provided by Standard Modern Arabic (hereafter SMA) to indicate the status of interpersonal units and the different relationships that may hold between them in coordinative constructions. I will concentrate on verb mood, interrogative particles and case marking. Some general implications of the proposed analysis for other morpho-syntactic areas such as non-restrictive term modifiers will be briefly examined.
2 Interpersonal vs structural complexity
Interpersonal complexity differs from structural complexity in the sense that the basis on which the former is defined is semantic and pragmatic rather than morpho-syntactic. More specifically, interpersonal complexity resides not in the number of the clauses a sentence can contain but rather in the number as well as the communicative status of the Acts that a Move can consist of.
Hengeveld and Mackenzie (2008, Forthcoming) give a complete and precise characterization of interpersonal structure and the relationships that can hold between its units. On the basis of this characterization, the following definition of complex Moves can be drawn up:
(1) Complex Move:
'A complex Move is a Move which contains more than one discourse Act.'
As regards the relationships that the elements of a complex Move can entertain, Hengeveld and Mackenzie (2008, Forthcoming) point out that they can be of two kinds: 'equipollence' and 'dependence'.
Two Acts are said to be 'equipollent' when they are given the same 'communicative status' by the speaker. Conversely, two Acts entertain a relationship of dependence when the speaker gives them 'unequal communicative status'. A dependent Act is typically a Non-Nuclear Act bearing a rhetorical function. However, Hengeveld and Mackenzie (2006) show that, in some cases such as non-restrictive relatives, an Act can be both nuclear and dependent at the same time. Moreover, Mackenzie (p. c.) assumes that any parenthesis would have the status of a dependent nuclear Act.
Dependent Acts can be assigned the rhetorical functions Motivation, Concession, Orientation and Correction., the latter two being typically assigned to the Acts expressed by the constituents referred to in FG as Theme and Tail. The list of rhetorical functions can be taken as open since other functions can also be postulated as well. In this respect, Hannay and Keizer (2006) propose a classification of non-restrictive nominal appositions based on the different rhetorical functions they can fulfil in discourse as autonomous discourse Acts. In this classification three main types of functions are distinguished: 'Reference identification', 'Justification' and 'Labelling'.
In Hengeveld and Mackenzie (2008, Forthcoming), non-restrictive relative clauses are analysed as expressing a dependent Nucleus Act which bears no specific rhetorical function. However, Mackenzie (p. c.) assumes that even a dependent nuclear Act can be labelled with a rhetorical function if it is proven that some formal aspect is relatable to this function.
In the rest of this study, I shall propose assigning specific rhetorical functions to dependent Acts only if the assignment of such functions turns out to be required by the adequate description of the formal characteristics of the relevant constructions.
Interpersonal complexity and structural complexity may coincide but this is far from being a general rule: on the one hand, structurally complex constructions may have a simple underlying interpersonal organization (2); on the other hand, interpersonally complex structures can surface as simple morphosyntactic configurations. The reason for this 'discrepancy' between the interpersonal and the structural levels mainly resides in the fact that there is, as established in (KROON, 1997), no one-to-one correspondence between their respective minimal units (i.e. the Act and the Clause respectively).
The interpersonal-structural discrepancy phenomena in coordinative constructions are illustrated in the following sentences:
(2) ma Zaydun faylasufan bal sa'iran. NEG Zayd.NOM philosopher.ACC but poet.ACC 'Zayd is not a philosopher but a poet.' (3) ma Zaydun faylasufan bal sa'irun. NEG Zayd.NOM philosopher.ACC but poet.NOM 'Zayd is not a philosopher. He is rather a poet.'
In both (2) and (3), two predicates of the same clause are coordinated. However, as I shall argue below, in (2) the coordination takes place within the same Act while in (4) it involves two full discourse Acts.
On the basis of the facts exemplified in (2) and (3), we can draw two conclusions. First, interpersonal complexity must be distinguished from structural complexity. Second, languages are expected to provide specific means for the realization of interpersonal complexity features.
In SMA, as in other languages, these features can be signalled by lexical means (i.e. modifiers) as well as morpho-syntactic and/or prosodic means. I shall restrict myself here to some specific features of Arabic morpho-syntax, namely verbal mood, interrogative particles and case marking, as they occur in coordinative constructions
SMA, like many other languages worldwide, has several coordinators. The Arab Grammatical Tradition (hereafter AG) mentions wa (and), fa/thumma (and then), 'aw/'am (or), lakin (but), bal (but rather), la (not) and hatta (even). The occurrence properties of these coordinators have been discussed at length in (MOUTAOUAKIL, 1988) within the FG framework. In brief, they are codetermined by the meaning of the coordinators themselves (as sequential/non-sequential, conjunctive/disjunctive, alternative, or adversative) and the number as well as the category of the elements they serve to coordinate. What we will be concerned with here is the coordination of clauses and (nominal/adjectival) predicates.
As is well-known, coordinated structures are thought of in formal grammars as complex constructions consisting of independent elements linked by a coordinative conjunction. Such a characterization is clearly insufficient and partially false: although they are structurally not embedded in each other, the members of a coordination sequence may entertain pragmatic-semantic dependence relationships.
In contrast, the FDG model, in taking underlying interpersonal status into account, allows a more adequate description of coordinative constructions. First, it allows one to determine the exact type of interpersonal complexity as well as the (equipollence vs dependency) relations that can obtain between the coordinated units. Second, it makes it possible to account in a more complete and precise way for the structural properties of coordinative constructions (including the complementary distribution of the coordinators) on the basis of their underlying interpersonal features.
In some languages, the interpersonal status of the coordinated clauses can be mediated through the selected coordinator itself. In this connection, Hengeveld and Mackenzie (2008, Forthcoming) point out that French car, for example, is specialized for the coding of dependent Acts. No such a specialization seems to hold in SMA: all the available coordinators may be used to relate terms, predicates or clauses expressing either full (in)dependent Acts or mere Propositional Contents. As we shall see below, the task of encoding the interpersonal status of the coordinated units is rather fulfilled by verb mood, interrogative particles in clause coordination and case marking in (nominal/ adjectival) predicate coordination.
4 Verb mood and equipollence vs dependence status
As proposed by Cuvalay-Haak (1997), there is a morphological opposition in Arabic between two sets of verb forms: suffix verb forms and prefix verb forms. The verb forms belonging to the latter set take one of three endings: Indicative, Subjunctive and Jussive, whose typical surface realizations are u, a and O (absence of vowel) respectively.
Our main aim here is to argue that the verb endings serve not only to realize TMA-categories but also to indicate the interpersonal status of the coordinated clauses.
Let us consider sentences (4a-b):
(4) a. liyarhal Zaydun fa-yaqdimu 'Amrun PART.leave.OPT.3.SG.M Zayd.NOM and-come.IND.3SG.M 'Amr.NOM 'May Zayd leave! And then 'Amr will come.' b. liyarhal Zaydun fa-yaqdima 'Amrun PART.leave.OPT.3SG.M Zayd.NOM and-come.SUBJ.3SG.M. 'Amr.NOM 'May Zayd leave! So 'Amr can come.'
As pointed out above, coordinated Acts can be either equipollent or dependent. In (4a-b), for instance, a DECLarative Act is coordinated with an OPTative Act. In (4a) the two Acts are equipollent while in (4b) the second Act depends, as being the purpose of Zayd's leaving upon the first one.
This difference is expressed by the verb mood: in (4a), the verb in the second clause takes the Indicative mood whereas in (4b) it takes the Subjunctive mood. This supports the claim that the communicative status of the coordinated members determines not the choice of the coordinator but rather the mood of their verbs.
The underlying interpersonal representations of (4a-b) are (5a-b):
(5) a. ([M.sub.1]: ([A.sub.1]: [(OPT [F.sub.1]) ([P.sub.1]).sub.S] ([P.sub.2]) A ([C.sub.1]: [[([T.sub.1]).sub.NewFoc] ([dR.sub.1]: Zayd [([R.sub.1])).sub.Top]] ([C.sub.1]))] [([A.sub.1])).sub.Nucl] < > ([A.sub.2] [(DECL [F.sub.2]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [[([T.sub.2]).sub.NewFoc] ([dR.sub.2]: 'Amr ([R.sub.2]))] (([C.sub.2]))] [([A.sub.2])).sub.Nucl] ([M.sub.1])) b. ([M.sub.1]: ([A.sub.1]: [(OPT [F.sub.1]) ([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.NewFoc] ([dR.sub.1]: Zayd [([R.sub.1])).sub.Top]] ([C.sub.1]))] ([A.sub.1])).sub.Nucl] < ([A.sub.2]: [(DECL [F.sub.2]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [[([T.sub.2]).sub.NewFoc]([dR.sub.2]: 'Amr ([R.sub.2]))] (([C.sub.2]))] ([A.sub.2])) Dep ([M.sub.1]))
5 Verb mood and interpersonal status shift
By 'interpersonal status shift', I mean the phenomenon exhibited by the following sentences:
(6) a. liyarhal Zaydun fa-yaqdima 'Amrun PART.leave.OPT.3SG.M Zayd.NOM and-come.SUBJ.3SG.M Amr.NOM wa yuqfala l-babu. and close.PASS.SUBJ.3SG.M the-door.NOM 'May Zayd leave ! So, 'Amr will come and the door will be closed.' b. liyarhal Zaydun fa-yaqdima 'Amrun PART.leave.OPT.3SG.M Zayd.NOM and-come.SUBJ.3SG.M. 'Amr.NOM wa yuqfalu l-babu. and close-PASS.IND.3SG.M the-door.NOM 'May Zayd leave ! So 'Amr can come. And the door will be closed.'
In sentences (6a-b), two coordination processes take place, the first with fa and the second with wa. In (6a), the last coordinated element has the same interpersonal status as the second one. They are Nuclear Acts relating to each other. With respect to the first element, both are dependent Acts. In sentence (6b), the status of the last coordinated element is different: it no longer expresses a dependent Act as in sentence (6a) but rather a new equipollent nuclear Act with respect to the first and second elements taken as a whole. According to this approach, the analysis of (6a-b) will be (7a-b):
(7) a. ([M.sub.1]: [([A.sub.1]: [(OPT [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.NewFoc] ([dR.sub.1]: Zayd [([R.sub.1])).sub.Top]] ([C.sub.1]))] [([A.sub.1]:)).sub.Nucl] < ([A.sub.2]: [DECL [F.sub.2]) ([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [[([T.sub.2]).sub.NewFoc] ([dR.sub.2]: 'Amr ([R.sub.2]))] (([C.sub.2]))] [([A.sub.2])).sub.Nucl] < > ([A.sub.3]: [DECL [F.sub.3]) [([P.sub.1]).sub.S [([P.sub.2]).sub.A] ([C.sub.3]: [[([T.sub.3]).sub.NewFoc] [([dR.sub.3]).sub.Top]](([C.sub.3]))] [[([A.sub.3])).sub.Nucl]].sub.Dep] ([M.sub.1])) b. (M1: [([A.sub.1]: [(OPT [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.NewFoc] ([dR.sub.1]: Zayd [([R.sub.1])).sub.Top]] ([C.sub.1]))] [([A.sub.1])).sub.Nucl] < ([A.sub.2] [DECL [F.sub.2]) ([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [[([T.sub.2]).sub.NewFoc] ([dR.sub.2]: 'Amr ([R.sub.2]))](([C.sub.2]))] [[([A.sub.2])).sub.Dep]].sub.Nucl] < > ([A.sub.3]: [DECL [F.sub.3]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.3]: [[([T.sub.3]).sub.NewFoc] [([dR.sub.3]).sub.Top]](([C.sub.3]))] [([A.sub.3])).sub.Nucl] ([M.sub.1]))
This change in communicative status is reflected by a change in verb mood. As expected, in (6a), the verb yuqfala takes the Subjunctive mood which expresses the dependence of the coordinated Acts, whereas in (6b) it takes the Indicative mood which signals that the thid coordinated element is an independent full Act. (3)
6 'a and hal communicative distribution
In SMA, the expression of INTERrogation is mediated through prosodic and morphological means. Prosodically, interrogative constructions have a specific (rising) intonational contour. Morphologically, they exhibit a clause-initial particle ('a/hal) or an interrogative pronoun such as ma (what), man (who), mata (when) and 'ayna (where).
Let us concentrate on the particles 'a and hal and, in particular, on their occurrence properties in coordinative constructions like the ones exemplified in (8a-c) and (9a-c):
(8) a. 'a fazat 'am rasabat l-fatatu PART succeed.PAST.3SG.F or fail.PAST.3SG.F the-girl.NOM 'Did the girl succeed or fail? b. 'a kitaban qara'at l-fatatu 'am mazallatan? PART book.ACC read.PAST.3SG.F the-girl.NOM or journal.ACC 'Was it a book that the girl read or a journal? c. *'a qara'at l-fatatu l-kitaba PART read.PAST.3SG.M the-girl.NOM the-book.ACC 'am 'a namat? or PART sleep.PAST.3SG.F (9) a. Hal qara'at l-fatatu l-kitaba? PART read.PAST.3SG.M the-girl.NOM the-book.ACC 'am hal namat? or PART sleep.PAST.3SG.F 'Did the girl read the book? Or did she go to bed?' b. 'adat Hindun come.back.PAST3.SG.F Hind.NOM fa-hal 'uqfila l-babu? and-PART close-PASS-PAST.3SG.M the-door-NOM 'Hind came back. And has the door been closed?' c. *Hal fazat 'am rasabat l-fatatu? PART succeed.PAST.3SG.F or fail.PAST.3SG.F the girl-NOM d. *Hal kitaban qara'at l-fatatu? PART book.ACC read.PAST.3SG.F the-girl.NOM 'am mazallatan? or journal-ACC
It becomes clear from these examples that the occurrence of the two particles in coordinative constructions takes place as follows: 'a appears in interrogative constructions with two coordinated focussed Ascriptive or Referential sub-acts, as in (8a) and (8b) respectively. As for the particle hal, it typically initiates interrogative constructions involving two discourse Acts which can be equipollent as in (9a) or entertain a dependence relationship as in (9b) where the first Act functions as a Motivation of the second one.
(8a), (8b), (9a) and (9b) are given the underlying representations in (10a), (10b), (10c) and (10d) respectively:
(10) a. ([A.sub.1]: [(INTER [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.ContrFoc] [([T.sub.2]).sub.ContrFoc] [([dR.sub.1]).sub.Top]] ([C.sub.1]))]([A.sub.1])) b. ([A.sub.1]: [(INTER [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [([T.sub.1]) [([dR.sub.1]).sub.Top] [(I[R.sub.2]).sub.ContrFoc] [(i[R.sub.3]).sub.ContrFoc]] ([C.sub.1])) ([A.sub.1])) c. ([M.sub.1]: ([A.sub.1]: [(INTER [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [([T.sub.1]) [([dR.sub.1]).sub.Top] ([dR.sub.2])]] [([C.sub.1])).sub.NewFoc]] [([A.sub.1])).sub.Nucl] < > ([A.sub.2]: [(INTER [F.sub.2]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] [([C.sub.2]: [([T.sub.2]) ([dR.sub.3])] [(([C.sub.2])).sub.NewFoc]] [([A.sub.2])).sub.Nucl] ([M.sub.1])) d. ([M.sub.1]: ([A.sub.1]: [(DECL [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] [([C.sub.1]: [([T.sub.1]) [([dR.sub.1]: Hind [([dR.sub.1])).sub.Top]] [([C.sub.1])).sub.NewFoc] ([A.sub.1])).sub.Motiv] ([A.sub.2]: [(INTER [F.sub.2]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [([T.sub.2]) ([dR.sub.3])] [(([C.sub.2])).sub.NewFoc]]] [([A.sub.2])).sub.Nucl] ([M.sub.1]))
The oddity of (8c), where 'a appears in each of the coordinated clauses, shows that this particle cannot be used to carry out a coordinated full discourse Act. Conversely, the ungrammaticality of (9c-d) makes it clear that hal cannot take either an Ascribing or a Referential sub-act in its scope.
Notice that the ungrammaticality of constructions such as (8c) can be related in the FDG framework to the fact that they involve a 'heterogeneous' coordination, i.e. a coordination of non-congruent members (namely a full Act and an Ascriptive sub-act).
The constraint conceived of in FG (MOUTAOUAKIL, 1988; DIK, 1997) as an illocutionary congruence constraint could be reformulated in FDG as a more general 'Interpersonal Status Congruence Constraint'. In its extended formulation, this constraint could also hold for constructions like (11):
(11) Sa'usafiru l-yawma wa bi s-sayyarati. travel.1.SG the-day.ACC and with the-car.GEN 'I will travel today. And I will do that by car.'
Constructions such as (11) are viewed in FG (DIK, 1997, p. 192) as ill-formed since they involve a coordination of two functionally non-equivalent terms, i.e. a Temporal and a Means satellite. Such constructions pose no problem, however, if they are analysed in the FDG framework on the basis that the second member of the coordination is set off from the clause to the extent that it carries out an autonomous discourse Act. In other words, the constructions exemplified in (11) involve a coordination of two equipollent discourse Acts rather than two terms within a single Act, so that it satisfies the 'Interpersonal Status Congruence Constraint'.
7 Case marking in predicate coordination
Coordinated (Nominal/Adjectival) predicates can be linked by the coordinator wa in positive clauses as in (12) and by bal in negative clauses, as seen in (2) and (3) repeated here for convenience:
(12) Zaydun faylasufun wa sa'irun Zayd.NOM philosopher.NOM and poet.NOM 'Zayd is a philosopher and a poet.' (2) ma Zaydun faylasufun bal sa'iran. NEG Zayd.NOM philosopher.ACC but poet.ACC 'Zayd is not a philosopher but a poet.' (3) ma Zaydun faylasufan bal sa'irun. NEG Zayd.NOM philosopher.ACC but poet.NOM 'Zayd is not a philosopher. He is rather a poet.'
The interesting fact about constructions like (2) and (3) is that the second member of the coordination can take the same case as the first member as in (2) or a different case, namely the Nominative, as becomes clear from example (3).
From the FDG point of view, bal in (2) coordinates two Ascriptive sub-acts within one Communicative Content while in (3) it is used to conjoin an autonomous Act and the Act formulated in the host clause, as becomes clear from representations (13a-b):
(13) a. ([A.sub.1]: [(DECL [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.ContrFoc] [([T.sub.2]).sub.ContrFoc] ([dR.sub.1]: Zayd [([R.sub.1])).sub.Top]] ([C.sub.1])) [([A.sub.1])).sub.Nucl] b. ([M.sub.1]: ([A.sub.1]: [(DECL [F.sub.1]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.1]: [[([T.sub.1]).sub.ContrFoc] ([dR.sub.1]: Zayd [([dR.sub.1])).sub.Top]] ([C.sub.1])) [([A.sub.1])).sub.Nucl] < > ([A.sub.2] [(DECL [F.sub.2]) [([P.sub.1]).sub.S] [([P.sub.2]).sub.A] ([C.sub.2]: [[([T.sub.2]).sub.ContrFoc]] (([C.sub.2]))] [([A.sub.2])).sub.Nucl] ([M.sub.1]))
The independence of the Act expressed by the predicate sa'irun in (3) is signalled by the Nominative case. The case marking at work in the coordination exemplified in (2) and (3) is one of the significant instantiations of the role that case variation can play in encoding the communicative status of interpersonal units, as we will see below.
8 Do we need rhetorical functions?
As mentioned above, dependent Acts and Communicated Contents can be assigned rhetorical functions. Of theoretical importance here is Hannay and Keizer's (2006) observation that although rhetorical functions have an explanatory power which must be taken into account in discourse-oriented grammars like FDG, their assignment should be justified by the influence that they can have on the formal properties of the different kinds of apposition constructions. This requirement relates to the general principle in F(D)G according to which only elements that have impact on the form of linguistic expressions can be represented in their underlying structure.
As far as coordinative constructions in SMA are concerned, the assignment of specific functions turns out to be without any impact on the morpho-syntactic properties of the coordinated members: dependent Acts take the same formal expression irrespective (of) the function that they could be assigned.
On the basis of these observations, we may represent the underlying interpersonal structure of the coordinative constructions at hand without any function assignment, the dependence status (represented by >) sufficing to determine the Subjunctive mood borne by the verb. More specifically, the general functional indices Nucl and Dep will trigger, during morpho-syntactic encoding, the assignment of the Indicative/Nominative or Subjunctive/Accusative suffixes.
Notice that this position could be taken for all the constructions where the correlation between dependence and the Subjunctive mood holds. However, I would like to leave open the possibility of assigning specific functions to coordinative constructions if it is proven that other formal aspects are sensitive to these functions.
9 Unified interpersonal morphology
It has become clear from the data discussed above that, in SMA, some morphological means such as case marking and verb mood play a role not only in grammar but also in discourse organization.
Case marking expresses syntactic (or semantic) functions, as it is well-known, but it also serves to indicate the interpersonal status of discourse units. Nominative case generally signals independent Acts whereas Accusative case is the typical formal expression of dependent Acts. This is evidenced by the fact that the correlations (Independence with Nominative and Dependence with Accusative) hold for various kinds of constructions. Let us briefly examine the case pattern exhibited by examples (14a-c):
(14) a. l-taqaytu bi-r-razuli s-samini l-'asmari meet.PAST.1SG with-the-man.GEN the-fat.GEN the-brown.GEN 'I met the fat brown man.' b. l-taqaytu bi-r-razuli s-samini, l-'asmara meet.PAST.1SG with-the-man.GEN the-fat.GEN the-brown.ACC 'I met the fat man, I mean the brown one.' c. l-taqaytu bi-r-razuli s-samini, l-'asmaru meet.PAST.1SG with-the-man.GEN the-fat.GEN the-brown.NOM 'I met the fat man. He is the brown one.'
In sentence (14a), the adjective l-'asmar functions as an internal modifier within the NP r-razuli s-samini l-'asmari. In sentences (14b-c), the same adjective behaves as a detached constituent expressing an autonomous discourse Act different from the one containing the host NP. In other words, while sentence (14a) embodies no more than one Act, sentences (14b-c) express two distinct Acts.
Let us concentrate on the case-marking properties of the modifier l-'asmar in examples (14a-c). In (14a), the modifier agrees in case with the head noun of the host NP while in (14b) and (14c) it takes the Accusative and the Nominative cases respectively. One possible way to account for this variation in case is the following: in both examples (14a-c), the detached adjective carries out an autonomous Act with the difference, however, that, in (14c), this Act is independent while in (14b) it depends upon the host Act.
In (14b), the detached adjective embodies a specification of the entity designated by the host NP while in (14c) it rather initiates a completely new discourse Act. In this respect, (14c) is synonymous with (15):
(15) l-taqaytu bi-r-razuli s-samini. meet.PAST.1SG with-the-man.GEN the-fat.GEN Huwa l-'asmaru l-ladi ... He the brown.NOM the-who ... 'I met the fat man. He is the brown one who ...'
If we take into account that Nominative and Accusative cases are assigned in grammar to independent arguments (Subject/Topic) and to dependent modifier terms respectively, we can generalize over the roles of these two cases in grammar and discourse as follows:
(16) 'Nominative case marks independent units; Accusative case marks dependent units.'
The endings of the prefix verb forms are commonly taken as indicating TMA values. What is not mentioned in the modern Arabist literature, as far as I know, is that these endings can also function as discourse markers, in particular, as devices indicating interpersonal status as well as a shift in interpersonal status: the a ending expresses dependence while the u ending marks equipollence or signals interpersonal status shift.
Interestingly enough, Ancient Arab grammarians do not differentiate these two verb endings from their homonymous forms functioning as Nominative and Accusative case marks, on the basis of the traditional claim that the prefix verb form (unlike the suffix verb form) is a 'hybrid' form displaying some 'nouniness' features including case marking. This claim can be re-interpreted as follows: the same morphemes a and u function as case marks on nouns and adjectives and as TMA marks on prefix verb forms.
Such a re-interpretation allows us to re-formulate generalization (16) as (17):
(17) "Grammatical/discourse independence is expressed by the (nominal, adjectival or verbal) suffix u; grammatical/discourse dependence is expressed by the (nominal, adjectival or verbal) suffix a."
Needless to say, the morpho-syntactic sub-component of the FDG of Arabic could be significantly simplified by generalizations like (17).
Coordinative constructions can involve a coordination of either equipollent or dependent discourse Acts, the interpersonal status differences being expressed by verb mood or case-marking distinctions and complementary distribution of interrogative particles.
Some generalizations on the relationships between the status of the interpersonal units and their formal expression can be captured. More importantly, interesting correlations between verb mood and case marking, such as those between Indicative and Nominative or between Subjunctive and Accusative, can be taken as relevant in this connection.
The notion of 'complex structure' should be revised on the basis of a clear distinction between interpersonal and structural complexity and the fact that the status of the interpersonal units is not necessarily reflected by the morphosyntactic configuration in which they are encoded. More specifically, the notion of coordination as well as the traditional sets of coordinators should be re- defined rather on the basis of the interpersonal properties of linguistic expressions.
The organization of the grammatical component of FDG and, in particular, the autonomy of the four levels it involves provide a highly suitable framework for such a revision.
Of considerable interest to (the history of) the epistemology of linguistics is the fact that the treatment of discrepancy phenomena is one of the issues where Arab Grammatical Tradition and FDG converge. This could be taken as an 'external' evidence for the FDG conception of the interpersonal and structural levels and their autonomy.
I am grateful to Prof. J. Lachlan MACKENZIE and Prof. Mike HANNAY for correcting the English of an earlier version of this article and for their valuable comments which have markedly improved the text.
Received June 2007 Approved October 2007
CUVALAY-HAAK, M. The verb in literary and colloquial Arabic. Berlin: Mouton de Gruyter, 1997.
DIK, S.C. The theory of Functional Grammar, Part 2: complex and derived constructions. Berlin: Mouton de Gruyter, 1997.
HANNAY, M.; KEIZER, E. A discourse treatment of English non-restrictive nominal appositions in Functional Discourse Grammar. In: MACKENZIE, J. L.; GOMEZ-GONZALEZ, M. A., Studies in Functional Discourse Grammar. Berlin: Peter Lang, 2006. p.159-194.
HENGEVELD, K.; MACKENZIE, J. L. Functional Discourse Grammar. In: BROWN, K. (Ed.). Encyclopedia of language and linguistics. 2nd ed. Oxford: Elsevier, 2006.
--. Functional Discourse Grammar. Oxford: Oxford University Press, 2008. Forthcoming.
KROON, C. Discourse markers, discourse structure and Functional Grammar. In:. CONNOLLY, J. et al. Discourse and pragmatics in Functional Grammar. Berlin: Mouton de Gruyter, 1997. p.17-32.
MOUTAOUAKIL, A. Essais en Grammaire Fonctionnelle. Rabat: SMER, 1988.
HENGEVELD, K. The architecture of Functional Discourse Grammar. In: MACKENZIE, J. L.; GOMEZ-GONZALEZ, M. A. A new architecture for Functional Grammar. Berlin: Mouton de Gruyter, 2004. p.1-21.
MOUTAOUAKIL, A. Reflexions on the layered underlying representation in Functional Grammar. Casablanca: Afric-Orient, 2000.
--. Function independant morpho-syntax. In: AERTSEN, H.; HANNAY, M.; LYALL, R. Words in their places: a festschrift for J. Lachlan Mackenzie. Amsterdam: Faculty of Arts, 2004. p.143-151.
--. Exclamation in Functionnal Grammar: sentence type, illocution or modality? In: DE GROOT, C.; HENGEVELD, K. Morpho-syntactic expression in Functional Grammar. Berlin: Mouton de Gruyter, 2005. p.351-379
--. Functional Grammar and Arabic. In: EID, M. et al. (Ed.). Encyclopedia of Arabic language and linguistics. Leiden: Brill, 2006. v.2, p.143-150.
SADOCK, J. M. The pragmatics of subordination. In: GEEST, W de; PUTSEYS, Y. Sentential complementation. Dordrecht: Foris, 1984. p.205-213.
(2) Examples of structurally complex but interpersonally simple constructions are performative utterances. Such constructions exhibit a surface morpho-syntactic configuration with a matrix and a subordinate clause. They express, however, no more than one discourse Act (HENGEVELD; MACKENZIE, 2008, Forthcoming).
(3) In Arab Grammatical Tradition, this phenomenon is called 'isti 'naf'. In its ordinary use, 'isti 'naf' means that a process is re-starting (possibly in a new, different direction) after a short or long interruption. Ancient Arab grammarians conceive of the conjunction wa in constructions like (6b) as a 'shifter' rather than a coordinator
Ahmed MOUTAOUAKIL (1)
(1) Mohammed V University--Faculty of letters-Rabat--BP 1040--Morocco. E.mail address:email@example.com
|Printer friendly Cite/link Email Feedback|
|Publication:||Alfa: Revista de Linguistica|
|Date:||Jul 1, 2007|
|Previous Article:||The lexical-grammatical dichotomy in functional discourse grammar.|
|Next Article:||Basic illocutions in the native languages of Brazil.|