Abstract This article suggests that the discomfort with translation theory felt by some translation scholars arises from the fact that translation theory has tended to undermine itself, and hence translation studies as such, by questioning the existence of its own subject matter. An attempt is made to ease the discomfort by defending Davidson's (1973; 1974) reply to the indeterminacy thesis proposed by Quine (1960). Finally, the article draws on Davidson's later theory of linguistic interaction (1986) in presenting a model of translation which highlights features which translation does not share with other types of linguistic interaction, and which may, consequently, merit particular attention in translation theory.