Can, or Should, Dummett Solve the Delivery Problem?

Abstract

Michael Dummett has long argued that we should ascribe implicit knowledge of a meaning-theory to speakers, and that the task of a theory of meaning is to tell us what such knowledge consists in. But he also sees it as a problem that how implicit knowledge is actually used, that is, how a speaker's metalinguistic knowledge of a meaning-theory issues or delivers the speaker's knowledge of meanings of utterances (the delivery problem). In this paper I argue that Dummett's instrumental construal of implicit knowledge does not and cannot solve the delivery problem. However, I do not suggest Dummett to modify or abandon his instrumental construal; rather, I think he can dissolve the delivery problem by recognizing that knowledge of semantics for a language is not a necessary condition for mastering a language. I shall argue this point through Davidson's attitude towards the role of linguistic knowledge and his thesis in his (in)famous paper A Nice Derangement of Epitaphs. It is widely held in contemporary philosophy of language and linguistics that linguistic competence requires knowledge of a formal semantic theory, such as Davidsonian truth-conditional semantics or Chomskyan linguistic theory. Michael Dummett also subscribes to such a meaning-theoretic account of linguistic competence. He has long argued that we should ascribe (implicit) knowledge of a meaning-theory to speakers, and that the task of a theory of meaning is to tell us what such knowledge consists in. One way to tackle this task is to construct an articulated, correct, meaning-theory for a language, on the one hand, and give an account of the epistemic/cognitive relation mediated between the contents of the meaning-theory and a competent speaker of the language, on the other. Higginbotham wrote that Dummett dubbed Auslegung, Vol. 28, No. 1