The Gricean view of linguistic communication represents the interlocutors as expressing and recovering communicative intentions. According to it, this task is performed not only by a decoding of the linguistic sentence meaning, but also by a complex inferential activity. Many recent works in pragmatics are concerned with how much inferential activity is needed to explain the understanding of linguistic expressions, and how this activity is triggered and guided. In particular, it is commonly agreed that the success of linguistic communication owes much to expectations and heuristics that have evolved in the human cognitive system. Two main approaches can be distinguished: the neo-Gricean, represented here by Stephen Levinson and the post-Gricean, developed by Dan Sperber and Deirdre Wilson. My aim here is to examine their competing conceptions of expectations and heuristics. What I want to show is that (i) the two competing accounts of both the evolution and the current organization of the pragmatic cognitive mechanisms involved in linguistic understanding attribute very different roles to the heuristics in the retrieval of the speaker’s communicative intentions and (ii) both of heuristic solutions proposed are flawed.
