Abstract

Kai von Fintel and Anthony Gillies have proposed a revised contextual analysis of sentences that make use of “might” epistemically. On their view, when a speaker uses an epistemic modal term, several propositions are made available to his conversational partners and, as a result, there are several propositions that may be picked up on by those partners. Because there is no concrete “context of utterance,” there is no one proposition that the speaker could be said to have asserted. This is meant to resolve conflicting truth evaluations by different speakers of a single utterance. I argue that the position is unworkable for two reasons: First, on their view there is no single proposition that counts as being asserted or semantically expressed by a “might” utterance; this has several counterintuitive consequences. Second, their position does not address several of the original problems that led many to abandon a contextual account.