In this paper I argue for a new analysis of belief statements in which the complement clause denotes not the putative content of a person’s beliefs but a proposition that can be inferred from that content. The notion of inference that makes this analysis possible operates within a world rather than across worlds, setting it apart from the possible worlds analysis of entailment. Our analysis unifies standard de dicto, de re and de qualitate attributions while also distinguishing among beliefs in different necessities or impossibilities and accounting for possible and impossible inferences that can be drawn from such beliefs. I accept that belief statements are systematically ambiguous, but the only ambiguity unique to attitude attributions I take to be the de translato/non-de translato ambiguity argued for by Tancredi and Sharvit (forthcoming) : the embedded clause can be interpreted as the attributor takes the subject to interpret it (de translato) or as the attributor herself interprets it (non-de translato). I show how our analysis leads to the dissolution of Kripke’s puzzle about belief while making it possible to reject Putnam’s and Burge’s claims that beliefs are not in the head.

論文