MUST THE LANGUAGE OF KNOWLEDGE BE USED IN EXPLAINING KNOWLEDGE OF LANGUAGE? (DUMMETT, QUINE, CHOMSKY)

Few thinkers in the past three decades have exerted more influence on the philosophy of language than Quine, Dummett, and Chomsky. No investigation into the current state of philosophy of language can omit consideration of their views. Yet I believe that their work has often been seriously misinterpreted. I begin by trying to clear up some unfortunate and prevalent misunderstandings. In particular, I examine in detail the relationship between Quine's and Chomsky's thought and argue that rumors of their incommensurability have been greatly exaggerated. I lay out the many affinities between their approaches (including their agreement over the sort of data for which a theory of language must account ultimately, over the need for innate mechanisms, and over the posits to which such a theory will eventually be committed) and isolate the crucial junction at which they part company (over the explanatory structure of scientific theories, and, for this reason, over the conception of knowledge admissible in such accounts of language). I also reconstruct Dummett's arguments against truth-based theories of meaning and in favor of verificationist ones, and argue that some recent criticisms of these (e.g. Devitt's and Putnam's) have rested on misunderstanding. During my exposition of Dummett's thought, various puzzling impasses are noted. I show that these are symptomatic of a deep tension within his view. Essentially, this consists in his desire to follow both Quine and Chomsky, to adopt simultaneously a Chomskyan position and a Quinean one on the very issue that I argued fundamentally divided these two thinkers. The resultant schizophrenia about the nature of linguistic knowledge is studied closely and suggestions for its cure are tendered. Its significance for the future of the philosophy of language is assessed.