Who Needs Semantics of Quotation Marks?

This paper further develops the semantic approach to quotation marks first presented in Benbaji (2004a) and (2004b). The account defended here is a version of the neo-Davidsonian semantic theory of quotation recently revived by Cappelen & Lepore. I begin by providing two further pieces of evidence in support of a semantic account. I argue, contra Recanati, that quotation marks cannot be “pragmatic indicators”, namely “expressions which have certain conditions of use, and whose use indicates that the conditions in question obtain”. Facts about verb phrase anaphora and about the cancelability of conventional implicatures clearly show, I believe, that quotation marks contribute to what is strictly and minimally said by the sentence in which they appear. On the other hand, I argue, contra Cappelen & Lepore, that the semantics of these markers is not “innocent”. Within some contexts, the semantic value of quotation marks is a component of the proposition expressed by the sentence in which they appear, while within others it is part of the mechanism that determines which proposition is expressed by the sentence given a context.