Abstract

What accounts for the facility that masters of a language have with their language? The view I call the "Linguistic Knowledge Thesis" answers: knowledge. That is, it is because those who have mastered a language know the truths formulated in the semantic theory for that language that they have the facility that they do. Devotees are drawn to the Thesis out of a distaste for behaviorist or dispositional alternatives, but they encounter trouble when pressed to explain the nature of linguistic knowledge. Indeed, a common complaint against the thesis is that it threatens to overintellectualize language use in just the ways Wittgenstein taught us to suspect. Advocates of the Thesis frequently characterize a master's linguistic knowledge as "tacit" or "implicit", but this merely labels the problem. At the other extreme, there are those who explain mastery as the possession of a practical ability. This view has definite advantages: it does not require attributing to language users the contents of speculative and highly technical linguistic theories, and it may even make for a simpler theory of language acquisition. Yet it waits on a suitable account of practical abilities: mastery is not a practical ability in the sense that cycling and swimming are. ;The resolution of this debate consists in realizing that each side is, more or less, right about one half of language use. A non-knowledge-based view is an attractive conception of "hearing" a language; we can account for the comprehension of the sentences one encounters in books or on the radio without invoking linguistic knowledge. Yet we do need to ascribe such knowledge to language users in order to make sense of speaking--producing sentences in normal speech situations. This speaker-based approach to the Linguistic Knowledge Thesis is sharply at odds with those inspired by Davidson, Chomsky, and Fodor, whose views are grounded in facts about sentence comprehension. There are at least two important consequences of this argument: first, it turns out that speakers know, at best, only scattered bits of the highly speculative linguistic theories in question, and second, the knowledge they do have is the ordinary sort of knowledge we ascribe to agents in the course of rationalistic explanations of their behavior