Retour vers liste

Détail de la contribution

Auteur: Matthijs WESTERA

Meanings as proposals: an inquisitive approach to exhaustivity

Abstract/Résumé: A disjunction may defeasibly imply that only one of the disjuncts is true. (Neo-)Gricean accounts treat such exhaustivity implicatures as explanations why the speaker chose one utterance over stronger, relevant alternatives. However, it has proven difficult to characterise the set of relevant alternatives, and to describe a general-purpose Gricean reasoning pattern that applies to disjunctions like [p v q v (p^q)] and [p v q v r], disjunctions in embedded positions, number words, and more generally quantifiers with existential force. We think that not the Gricean approach is to blame for this failure, but rather the underlying, classical conception of meaning as information. An utterance alone cannot add information to the common ground; at best, an utterance is a proposal to do so. Conceiving of meanings as proposals commits us to Unrestricted Inquisitive Semantics (InqU) [1,2]. We use InqU as the semantic foundation for a Gricean pragmatics that overcomes the problems mentioned. Entailment in InqU is sparser than classical entailment. E.g., a disjunction (1a) is not entailed by a disjunct (1b), but only by (1c) (or something informationally stronger). (1) a. John or Mary will come to the party. b. John will come to the party. c. John will come to the party, or John and Mary. Intuitively, this reflects that (1b) is not 'maximally related' to (1a), because it fails to mention the possibility that Mary comes. In accordance with this intuition, we spell out the Maxim of Relation as follows: relative to the speaker's knowledge state, an utterance must entail the proposition under discussion (PUD). Suppose that (1a) is the PUD, and the speaker utters (1b). In accordance with Relation, she must either know that only John comes, or that both John and Mary will come. We call this a Relation implicature. Independently, the Quantity implicature is derived that it is not the case that the speaker believes that Mary will come. The Relation and Quantity implicatures together entail exhaustivity: the speaker believes that only John will come. We present a formalization of this idea and illustrate its empirical scope. There are two reasons why the approach works, both relying on the finer granularity of InqU. First, rather than comparing an utterance to informationally stronger alternatives, it is compared to alternatives that are equally informative, but more related. Second, although the pragmatics is globalist, InqU effectively computes the alternatives locally. [1] Ciardelli, I. (2009). Inquisitive Semantics and Intermediate Logics. MSc thesis. University of Amsterdam. [2] Westera, M. (2012). Meanings as proposals: a new semantic foundation for a Gricean pragmatics. Presented at SemDial.