Coordination in conversation

Abstract

I give an account of the meaning of epistemic modals-words like 'might' and 'must', on a broadly epistemic interpretation-and how speakers use them to coordinate on their information. I begin by exploring what epistemic modals mean. Motivated by embedding data which are problematic for almost all existing accounts, I develop a new semantics for epistemic modals which I call the bounded theory. The bounded theory comprises a standard relational semantics together with a constraint which entails that local information is always taken into account in the evaluation of epistemic modals. I argue that the bounded theory makes sense of the subtle embedding behavior of epistemic modals-thus providing an adequate account of their meaning-and sheds new light on the way that local information is structured in natural language. In Chapter 2, I turn to the question of how speakers use epistemic modals to coordinate on their common information. I argue that we can make sense of the characteristic dynamics of epistemic modality within a relational semantic framework (like the one given in Chapter 1), together with a contextualist approach to the dynamics of conversation. The key to doing so is to take the relation in question to be determined by the interlocutors' prospective common attitudes. Chapter 3 explores the contribution of 'must'. Assertions of -Must p-1 and assertions of p alone seem to have the same basic goal, and yet their felicity conditions subtly differ. I argue for a new characterization of those differences, and a broadly pragmatic explanation of those differences based on the approach to epistemic modals advocated in the preceding chapters. In Chapter 4, I return to semantic questions. I develop formal tools to compare the expressive power of semantic theories with respect to the embedding data which they can account for. These tools show that the relational theory can account for all the embedding data which recent revisionary theories can account for, but not vice versa, a fact which necessitates a substantial shift in perspective on the debate regarding the semantics of epistemic modals.