Daniel Laurier

The purpose of this paper is to offer an account of what an agent's being rational to do or think something might amount to, which doesn't reduce to saying that it consists in this agent's doing or thinking something that is rational for him. In the first section, I call attention to the fact that such a distinction between agent rationality and action or belief rationality is widely admitted, I reject the idea that it could be interpreted as a distinction between the rationality of tokens and the rationality of types, and I suggest one natural way in which a notion of objective agent rationality could be informally characterized in terms of action or belief rationality. But this first, rough, characterization depends on further uses of the notion of rationality which I try to make sense of in the second section, at least in a preliminary way. The burden of this second section is then to determine whether the intuition behind the informal proposal introduced in the first can be substantiated, i.e., whether a substantial and coherent notion of agent rationality can be worked out, and at what cost. In the concluding section, I try to "deflate" some of the worries that could be raised by the account of agent rationality I end up with.

The theory of radical interpretation, as based on the principle of charity, sets a priori limits on the possibility that different agents have different beliefs, and on the possibility that one has false beliefs. David Papineau put forward a teleological approach to intentional states which, he claims, doesn’t have these unacceptable consequences. Having distinguished half a dozen of different forms that the problem of radical interpretation might take, I show that Papineau’s approach is not radically different from those based on the principle of charity. Finally, I suggest that the consequences of the principle of charity with respect to the problems of error and divergence are in fact both unavoidable and acceptable.