Pragmatics and grammar (review)

Pragmatics and grammar is published in the well-known and long-lasting series 'Cambridge textbooks in linguistics', known in the discipline as 'the red series'. Neither original monographs nor coursebooks, the success of this series has shown that there is a place for quite high-level compendia of thought on a particular domain of linguistics with presentation of the author's own views. This volume by Mira Ariel fits the bill admirably, as it brings together a wide range of different positions on the grammar/pragmatics interface (much of it cutting-edge research), employs a wealth of pertinent utterance examples (virtually all attested), and presents A's own position on the interface.

At the outset, A claims 'The fact is our current grammar is often our pragmatics (of the past) turned grammatical' (xiii), and her aim is both to explain how this happens, how aspects of usage can penetrate language systems and change them, and to ensure that, nonetheless, a clear synchronic distinction between grammar and pragmatics is maintained. Two essential components of her view are that: (i) a key mechanism in language change is that of the 'salient discourse profile' of a linguistic form, that is, (roughly) a pattern of usage or discourse function of the form that has become routine and dominant among a community of language users; and (ii) the main locus for the grammar/pragmatics interface is a particular representational level of utterance content in which they both take part—'what is said' or 'what is explicated'—and so the pragmatic inferences that contribute to this content are the ones most likely to lead to semantic/grammatical change.

In Ch. 1, 'Introduction: Grammar, pragmatics and what's between them', A establishes her position on the semantics/pragmatics distinction as a distinction between the meaning delivered by the linguistic code (expression-type meaning) and the meaning derived by defeasible inferential processes guided by presumptions about rational communicative behavior (the code/inference distinction, in brief). A takes the (now fairly standard) view that communication, no matter how linguistically explicit, always requires pragmatic inference. She discusses the distinguishing properties of codes and pragmatic inferences, and outlines several different kinds of pragmatically inferred meanings and some of the mechanisms proposed for generating them. She ends by setting out the two big issues that centrally concern her: (i) how to draw the code/inference distinction, and (ii) how inferences can become coded.

Part 1 (Chs. 2 and 3), 'Drawing the grammar/pragmatics divide', focuses on the division of labor between grammar and pragmatics, and the kinds of arguments and evidence that indicate for any given phenomenon of utterance meaning whether it is a matter of grammar (linguistic encoding) or of pragmatic inference, and, if pragmatic, whether it contributes to the proposition explicitly expressed (explicature) or is an implicature. A makes clear that it is often no easy matter to establish a sure answer to these questions. She presents three specific case studies in their full complexity: (i) referring expressions, in particular, definite descriptions; (ii) temporal and causal interpretations of and-conjunctions; and (iii) upper-bounded interpretations of scalar terms, with a specific analysis of most.

In Ch. 2, 'Distinguishing the grammatical and the extragrammatical: Referential expressions', A discusses a number of issues concerning the meaning, function, and distribution of referring expressions (names, definite descriptions, pronouns, demonstratives, clitics, and zero forms). She presents her own take on the much disputed semantic/pragmatic status of the existential presupposition associated with definite descriptions. She argues that there is an encoded (hence context-invariant) component, namely 'givenness' (of the referent), and a pragmatically inferred (hence cancelable) component, namely speaker commitment to the existence of the referent. While the details of this analysis (33-44) and of several other analyses in this chapter are intrinsically interesting, more important for the theme of the book is the theoretical point they illustrate: for any regularly occurring form-function correlation, whether it is a stable feature of the code or a recurrent pragmatic inference is something that...