"Two Roles for Propositions: Cause for Divorce?" ABSTRACT The question addressed in this talk is the appropriate attitude for nondescriptivists, whether about morality, conditionals, epistemic modals, or other topics, to take toward propositions in their domain. Traditionally, nondescriptive theories are often thought of as denying the existence of propositions of a certain sort - moral propositions, or epistemic modal or conditional propositions. Some nondescriptivists have nevertheless claimed to be able to 'earn the right' to say that there are propositions in their domains, but only by offering a deflationary construal of this thesis. I'll summarize five general pieces of theoretical work that propositions do, but merely deflationary talk about propositions is unable to do - which correspond to five problems for theories which offer a deflationary construal, at best, of talk about propositions. But rather than arguing that these problems by themselves pose an insuperable challenge to nondescriptivist theories, I'll be aruing that they place productive constraints on how we should understand the best versions of such theories. I'll distinguish between two sets of theoretical roles that have traditionally been associated with propositions, as well as a set of arguments which suggest that the same entities play both sets of roles. But I'll argue that these arguments are not conclusive, and hence that a theory which takes the right view about the nature of belief can divorce these two roles. I'll go on to illustrate what such a view might look like, and show both why it is a recognizably nondescriptivist theory, and why it recognizably should count as believing in nondescriptive propositions. This gives us a fruitful way of thinking about the descriptivist/nondescriptivist divide not as between a proposition-based and a non-proposition-based way of understanding meaning, but as a surprising dispute about what sort of things propositions /are/.
