This is in response to Robin's comments on rationality. I agree with Robin that institutions shape the selves whose lives they structure, and in this context that the instrumentally rational agents that people rational choice models tend to be produced by the universal reign of markets. I think an appreciation of this, though, must lead to an attempt to formulate a thicker notion of rationality, if we are to do what Robin wants to do--evaluate institutions. Once we appreciate the shaping of agency and preferences by institutions, the standard welfare criteria, such as Pareto-efficiency, become simply tests of "coherence" between the agent-creating institutions and the agents they create. Arguably, coherence in this sense is neither necessary nor sufficient for an institutional structure's being judged good. Not sufficient: reread Brave New World. Not necessary either, because we might argue for the superiority of less coherent institutions which create unsatisfied Socratess to those which create satisfied pigs. (Apologies to Mill.) The point is that to evaluate institutions is to evaluate a package: institution-plus-types-of-agents, and we're then doing at the theoretical level what people do practically when they think not so much about how to get what they transparently want, but what sort of people they want to be, what sort of wants they should have. And the theorist should follow the practical agent in this: trying to articulate the sorts of criteria that are employed by the latter, and push for "reflective equilibrium". Unless one thinks that reason has no place here, we as practical agents employ and we as "social scientists" need to articulate a thicker kind of rationality. In Susan Hurley's phrase, a theory of rationality for human agents is "an ethic". I put the quote marks around "scientist" advisedly: this sort of endeavor won't pass muster under conceptions of science which make it the neutral examination of some independently existing reality.