ABSTRACT
Cataloguing what is obligatory and what is optional is not a particularly difficult task. The Arabic grammatical tradition for instance distinguished between the ʕumda, the obligatory subject (agent)-predicate, and the rest, the faḍalaat. A real challenge, however, is to define the conditions under which grammatically alternative ways of expressing the same propositional content are most appropriate, and under what conditions no choice at all is allowed. Reflecting the basic status of grammar in linguistics, six chapters in this volume explore this question. The explanation of pronominalization, anaphors and reflexives has sometimes been treated as a grammatical phenomenon, sometimes as derivable from pragmatic principles. Owens et al. follow the latter tradition (as in Levinson 1987) in explaining how null subjects in Arabic verbs find their referents. Using a corpus of spoken Arabian peninsular Arabic, the authors apply a Goldvarb analysis to tease out the factors influencing the presence or absence of overt subjects. Basing themselves on Levinson’s general anaphora principle, itself derived from Grice’s maxim of quantity, they confirm the commonsense intuition that null subjects tend to pattern with identical preceding subjects. The quantitative analysis, however, reveals a very large number of “exceptions” to this tendency – null subjects for instance often have a different subject from the previous verb – to the extent that a set of four conditions are developed to explain them. To confirm the fluidity of boundaries between grammar and pragmatics, the
authors suggest that what in spoken Arabic are largely pragmatically based constraints, pattern in ways analogous to grammaticized switch referent systems. A standing issue in the study of information structure is the interpretation of alternative word order. A number of dialects have variable SV/VS word order. In isolation, any sentence can have either order. The contributions of Clive Holes and Bruce Ingham explore the conditions favoring one order or the other. Examining a corpus of Bahrain Arabic, Holes observes a basic difference between the two: VS is the norm where narrated events are presented in sequence (see Myhill 1992). In the nearby Najdi Arabic of Saudi Arabia, Ingham in his contribution suggests that VS will be used when the entire sentence conveys new information (cf. Sasse’s 1987 contrast between thetic and categorical predicates). By contrast Holes finds that SV structure is used not to carry forward a narrative, but rather to “halt the story in order to focus on the actors,” or to provide background information to what has been narrated. Ingham notes a similar function of SV order in Najdi, suggesting that the order is used to indicate a contrast of subjects (focus on actors). Holes’ observation that SV occurs where there is no dynamic carrying the action forward provides a generalizing perspective: the introduction of an overt S in SV order signals an association with a discourse-old event. It is noteworthy that S in this order is often a pronoun, quantifier (like kill “all”) or general reference noun like naas “people,” underscoring the discourse-structural status of S of SV as conveying information of secondary status. As far as the subject position goes, discourse oldness is marked either by a repeated subject (in Ingham’s data) or by a participant-irrelevant subject (in Holes’ data). Holes’ chapter contains another interesting observation on the nature of SV/ VS order. In the only detailed study to date on this topic Dahlgren (1998) suggests that VS strongly correlates with perfective verb form. Holes’ data, however, shows that imperfect verbs are equally well represented in VS order. This observation underlines the importance of collecting a broad database on which to base generalizations. Dahlgren’s work, taken entirely from published texts, relies strongly on traditional folktales (Erzählungen), a dominant collecting tradition among Arabicists (and many working in non-western languages). Holes documents everyday conversation, which served as the basis of his extensive sociolinguistic investigations (e.g. Holes 1987). It may be that what the frequency of VS in his texts indicates is that it is narrative type rather than verb form which is the most basic determiner of VS order in Bahraini Arabic, a finding confirmed in Owens et al. (2009). Malcolm Edwards’ is the fourth paper in this volume to address word order among major sentence constituents. Taking an opposite tack from the previous two papers, Edwards argues that Egyptian Arabic word order is best understood in terms of a basic SVO grammar. He shows using a series of test criteria that Egyptian Arabic (quite different from the varieties described by Ingham and Holes) is constrained by a fairly rigid SV word order. With these criteria he furthermore argues, against previous work (Brustad 2000), that Egyptian Arabic at
least cannot be considered a topic-prominent language, since subjects and topics in the language are grammatically quite distinctive. Edwards emphasizes that there are grammatical constraints which are not amenable to pragmatic manipulation. In a fifth chapter, Mustafa Mughazy deals with the appropriateness of using the existential expressions in Egyptian Arabic. This is marked formally by the predicate fi “there is.” The presentative subject cannot simply be construed as discourse new, since there are obvious instances where the existential subject represents information known to the interlocutor. Previous analyses assume that prior knowledge of the addressee of the topic is the key to resolving the issue of matching the definiteness of the sentence initial NP (the topic) and the pragmatic function of relaying new information about this topic. Mughazy argues that an existential sentence is acceptable only if it provides information that the speaker believes to be new to the hearer. In his proposal, two conditions license the subject: a presentational proposition which establishes the discourse relevance of the subject, and, following Sperber and Wilson (1995), a relevance proposition which assures that the referent is integrated into the addressee’s state of knowledge. The chapter is an example of how examining a certain grammatical structure in the light of its pragmatic function can enhance our understanding of the links between syntactic rules and structuring information. Finally, Mohammed Farghal brings to light a fundamental contrast between Standard Arabic (SA) on the one hand and the spoken dialects on the other. Working within a Gricean framework, he identifies a class of tautological expression such as muhándis, muhándis “engineer, engineer,” whose pragmatic intention is more than the sum of the semantic parts. In this case, said with an intonation contour _ – _ _ – _, it indicates the speaker’s indifference to the state of affairs expressed in the proposition, “I don’t give a hoot if he’s an engineer.” Following Searle, Farghal terms this an “indifference tautology.” What he brings out is that tautological expressions by and large are available only within the spoken language. Standard Arabic has them only sparsely, a fact he attributes to an impoverished intonation system in SA, as well as to the fact that tautological expressions frequently violate the canonical sentence structure rules of SA. While there have been many studies, beginning at least with Al-Toma (1969), that describe structural differences between Standard Arabic and dialects, this study is one of the few to establish that there are fundamental differences in the pragmatic resources available to the different varieties.
