Abstract

Consider a superagent, a being with extraordinary rational capactities. Not only are a superagent’s beliefs closed under entailment, but a superagent also has a wonderful kind of introspective awareness: whenever she believes that she doesn’t believe something, she is right—she doesn’t believe it. A superagent, then, is a being who satisfies the following two principles: (I) If p entails q, and if S believes p, then S believes q. (II) If S believes that she doesn’t believe p, then S doesn’t believe p. I don’t suppose anyone thinks that there is such a thing as a superagent, or that there ever will be. But I do suppose that there could be a superagent, in the sense that it’s not conceptually impossible for a superagent to exist. Our concept of belief, whatever it is, should not be such as to rule out the possibility of a being of this kind. In this paper I expose a paradox: although our concept of belief shouldn’t rule out the possibility of a being of this kind, it apparently does.1 I end the paper by arguing that the key to solving this paradox lies in the recognition that the question of whether an agent could believe so-and-so can sometimes diverge from the question of whether it is possible that she believes so-and-so.