Letters to Language

Language accepts letters from readers that briefly and succinctly respond to or comment upon either material published previously in the journal or issues deemed of importance to the field. The editor reserves the right to edit letters as needed. Brief replies from relevant parties are included as warranted.

July 2, 2012
To the Editor:

According to Keith Allan (Language 88.2. 413-16), I am constantly 'admitting' things in my new book Defining pragmatics (DP) (Cambridge University Press, 2010): 'A ADMITS, "most, if not all language use involves both grammatical and pragmatic aspects"' (p. 413, emphasis added here and in following quotes); 'A ADMITS that codes often provide only rudimentary meanings' (p. 414); 'The pragmatic basis for grammar is ADMITTED' (p. 415). And although I profess to distinguish between grammar and pragmatics—based on a code versus inference division of labor—Allan finds me out: 'The squishy border between grammar and pragmatics CONTINUES TO BE UNCOVERED throughout...DP' (p. 414). Finally, I must be cutting the very pragmatic branch I am sitting on, for supposedly, my book 'charges that "pragmatics does not stand for a coherent set of issues"' (p. 415).

While Allan mostly perceives the propositional content of DP correctly, he somehow misses the stance taken, and thereby the point. Is pragmatics incoherent according to DP? Only if it is assumed that a host of criteria will all converge to neatly delineate the canonical list of topics, as in Stephen C. Levinson's Pragmatics (Cambridge University Press, 1983). A major point of the view proposed in DP is that multiple-criterion definitions (e.g. pragmatic interpretations are simultaneously inferred, nontruth-conditional, secondary, extrasentential, extragrammatical, etc.) are untenable, for they clash with each other. I show that quite often some phenomenon comes out pragmatic on one criterion, but grammatical on another.

But the problem runs deeper. I am specifically arguing against what I call the monolithic assumption, whereby specific topics must belong exclusively to either grammar or pragmatics (cf. Levinson's (1983:94) question of 'whether THE STUDY OF DEIXIS BELONGS TO SEMANTICS OR PRAGMATICS', emphasis added). Instead, I argue that while some aspects of speech acts, presupposition, and so on are grammatically specified, others are pragmatically inferred. There is, then, 'no pragmatic turf, with a predetermined set of topics that pragmatics has to include or exclude' (DP, p. xiv). Allan prefers a squishy distinction between grammar and pragmatics. But that presupposes a preexisting definition for the two. What is it?

Is DP's overall position that pragmatics is incoherent? Far from it. The book's main claim is that once we base the grammar/pragmatics distinction on a division of labor between directly ENCODED AND UNCANCELABLE associations between forms and meanings or conditions of use (grammar) and CANCELABLE associations mediated by INFERENCE (pragmatics), 'a unified view of the field can thus be construed' (DP, p. xiv). In other words, I assert the coherence of a pragmatics that is defined as inference, while rejecting as incoherent the older approach to defining the field.

The bulk of DP is devoted to specifically supporting (rather than admitting) the view that grammar contributes only a rudimentary meaning that must be pragmatically enriched (cf. relevance theory). I make a special effort to demonstrate that the grammatical/pragmatic divide, taken as code versus inference, is relevant to all so-called pragmatic topics, and even to topics deemed 'beyond pragmatics'. I proudly assert that code and inference virtually always work in tandem to get our message across to our addressee. Allan states that I do 'not succeed in distinguishing grammar as code from pragmatics as inference, partly because, as A[riel] admits, "most, if not all language use involves both grammatical and pragmatic aspects"' (p. 413). In other words, for Allan, if both code and inference are needed for language use, they cannot be distinguished. I beg to differ. It takes a left leg and a right leg to walk, and a female and a male to beget a child; yet their necessary interaction does not make them indistinguishable.

The other reason for my 'failure' to clearly distinguish codes from inferences according to Allan is that '[w]ere...