the week of october 8

pragmatic strengthening: exhaustive interpretations

We investigated a method of pragmatic strengthening that tries to capture the more general phenomenon of exhaustive interpretations, of which strengthening via scalar implicatures is just a special case. There are three major exhaustification methods in the literature: (1) interpretation in minimal verifying models (circumscription); (2) strengthening via exclusion of contextual alternatives; (3) interpretation in minimal verifying situations. We went through a number of examples illustrating exhaustive interpretations before taking a closer look at the minimal models method. This also gave us the occasion to practice the use of models. The second exhaustification method is represented by Benjamin Spector's work. It requires more background in semantics than I can presuppose here, but I placed a couple of links below. I will talk about the third method next week. I will illustrate it within a Davidsonian event semantics. The combined handout for Wednesday and Friday is here.

Here are some references on the topic of exhaustification. The two classics on circumscription are McCarthy's "Circumscription: a form of non-monotonic reasoning" and "Applications of circumscription to formalizing common sense knowledge." Schulz and van Rooij's "Pragmatic meaning and non-monotonic reasoning: the case of exhaustive interpretation" is the standard modern reference on circumscription in natural language semantics (there are also related papers by the same authors). Benjamin Spector's 2005 dissertation Aspects de la pragmatique des opérateurs logiques is the representative work on exhaustification via exclusion of contextual alternatives. The dissertation is in French, but there is an English appendix ("Scalar implicatures: exhaustivity and Gricean reasoning"), which lays out some of the major ideas. I have argued in talks and lectures over the last five years (see e.g. my Paris lectures) that the bulk of the exhaustification data falls out without any further assumptions from modules that are independently needed, including talk about minimal verifying situations, as in Davidsonian event semantics and situation semantics. My article on situations in natural language semantics gives a summary of those ideas. Simplifying slightly, I will illustrate the main insights within a Davidsonian event semantics next week.

All three approaches to exhaustification face considerable empirical challenges: you want to predict the correct exhaustive interpretations in a maximally general way. Homework 6 gives you a first taste of what those challenges look like. The topic of exhaustive interpretations is one of the hottest topics at the semantics/pragmatics interface these days. This is an area full of puzzles where discoveries can still be made.