Abstract

Davidson’s Surprise argument promises to resolve a dispute that has arisen in contemporary formal semantics over the proper semantic value for a semantic theory. At issue are doubts that Pietroski raises about the compositionality of truth-conditions, and thereby about truthconditional semantics, which treats a truth value as the semantic value for a sentence. The dispute is recalcitrant because, as I show, Pietroski’s evidence that truth-conditions are not compositional can be explained away with attention to Cappelen and Lepore’s distinction between the truth of what is semantically expressed by an utterance and the truth of its speech act content. While the Surprise argument would, if it worked, support truth-conditional semantics, I demonstrate that it fails; in fact, it is peculiarly vulnerable to Pietroski’s concerns.