Analyticity and the indeterminacy of translation

Abstract

The subject of this paper is the analytic-synthetic distinction and its relation to the indeterminacy of translation. In particular, I am con cerned with the question of whether a denial of the analytic-synthetic distinction need involve one in the general scepticism about meaning which is dramatized by the claim that translation is indeterminate. My answer to this question is complex, largely because of a complexity or unclarity in the way in which the analytic-synthetic distinction has been thought of. Proponents of that distinction have taken it to be an epistemological cleavage among statements; they have also assumed that analyticity is simply truth by virtue of meaning, and so must be acceptable if the notion of meaning is. They have thus assumed that the notion of meaning has the power to divide statements into two epistemologically quite distinct kinds. Once articulated in this way, this assumption is at least not obvious; I hope to cast doubt upon it. My strategy is as follows. Drawing largely on the work of Quine, I present two arguments (in sections 1 and 2, respectively). The second is an attack on the notion of meaning, and rests upon, or is equivalent to, the indeterminacy of translation. The first argument seems to rely upon nothing so controversial, and is explicitly an attack on the idea that the analytic-synthetic distinction is of general epistemological importance. Separating the two arguments illuminates the complex debate between Carnap and Quine over analyticity; it also suggests that one can accept the first argument (thereby rejecting the philoso phical use of the analytic-synthetic distinction) without accepting the general scepticism about meaning which is the burden of the second argument.1