UC Berkeley

Theories of discourse dynamics provide an account of the update or context change potentials of the sentences in a given fragment of language; how assertions of the sentences in question affect the context or the common ground of a conversation. The literature on the discourse dynamics of various fragments falls roughly into two camps. On the one hand, we have a non-semantic camp. According to members of this camp we should strictly distinguish between theories of meaning and theories of discourse dynamics. On the other hand, we have a semantic camp. According to this camp meanings are context change potentials — that is, providing a theory of the discourse dynamics of a fragment of language just is to provide a theory of meaning for the fragment in question. In this dissertation I hope to contribute to both the question of what the discourse dynamics of various fragments of language should look like as well as the question of whether we should side with the semantic or the non-semantic camp. With respect to the former question, I hope to defend a range of non-standard approaches to the discourse dynamics of conjunction, negation, conditionals and modals. With respect to the latter, my aim will be to convince the reader that, contra the popularity of dynamic approaches to the meaning of conditionals and modals, studying the discourse dynamics of such expressions provides an incentive to side with the non-semantic camp. The main argument is found in Chapter 2. Here we will see that given some plausible constraints on the context change potentials of indicative conditionals, there is a prima facie conflict between the idea that meanings are context change potentials and the idea that meanings behave compositionally. As I will argue, on popular and widespread views about context change and what it is for a theory of meaning to be compositional, we have reasons to believe that context change potentials do not behave compositionally. So, given that meanings behave compositionally, meanings are not context change potentials. Moreover, we will see that we can build a non-compositional theory of the context change potentials of conditionals on the basis of a truth conditional approach to meaning that is compositional in the relevant sense. In chapter 3, I focus on a purported advantage of a certain approach to the discourse dynamics of conjunction. According to the view in question, asserting a conjunction is like a successive assertion of its conjuncts. On this view the badness of certain conjunctions that are purportedly hard to explain on other accounts, reduces to the badness of certain discourses in an elegant way. However, I argue that this elegance is paid for with a range of implausible consequences. I conclude the chapter by discussing ways of giving up on the view and show that a version of the view developed in Chapter 2 does so in a particularly satisfying way. Chapter 4 develops a positive proposal for a theory of discourse dynamics for ‘might’ and ‘must’ claims. Here I argue that ‘must’ claims are genuinely informative in the sense of corresponding to assertions that eliminate some ways the world might be but not others. ‘Might’ claims, however, are never informative in this sense. Still, sometimes ‘might’ claims are non-trivial, for they highlight possibilities that have not been highlighted before. The approach to the discourse dynamics developed here is non-compositional in the same sense as the view discussed in Chapter 2.