What to Do When You Don't Know What to Do When You Don't Know What to Do…[Note 1. For their helpful feedback I thank Benj Hellie, Tom ...]

There’s been renewed interest in the question of what it’s rational to do in the face of normative uncertainty—uncertainty among moral theories, claims about what to do in particular cases, and even accounts of rationality itself.2 But some fascinating problems have been raised not for any particular answer to this question, but to the very project of trying to answer this question. One of these problems invokes agents who’ve tried to answer the question for themselves, but have arrived at answers that we might regard as mistaken. Consider: Solomon must determine the fate of the murderer Norman. Solomon knows that punishing Norman will achieve none of the “forward-looking” aims of punishment, and wonders what to do. On one hand, he thinks retribution is probably a sufficient, if weak, ground for punishment. But he thinks it might be no ground whatsoever, in which case jailing someone for sheer retribution would be horribly wrong. It would be, essentially, locking a person in a cage for no good reason. In desperation, he consults a book about rationality under moral uncertainty. It reads: “ . . . if an agent thinks it is more probable than not that doing A is better than doing B, it may still be more rational to do B, if the difference in moral value between doing B and doing A is much greater, on the supposition that B is better, than the difference between doing A and doing B is, on the supposition that A is better. In other words, it can be rational to ‘hedge your bets’ in cases of moral uncertainty, and act on views that are less probable, but have more moral value at stake.” On this view, let us assume, it would be most rational for Solomon to free Norman. But Solomon just doesn’t buy it. He thinks this author’s “moral hedging” idea is crazy. He comes to believe instead that it is rational simply to do the action that is most probably right. On this view, Solomon rationally ought to jail Norman. And so he does just that. The author of the moral uncertainty book is told about all of this, and is asked about the rationality of Solomon’s action. What should our author say? Perhaps she should say it would have been more rational for Solomon to free Norman. After all, that’s what the theory in her book implied. If Solomon didn’t want to believe that theory, that’s fine, but it doesn’t make the theory false, any more than someone’s disbelief renders a scientific theory false. This answer is not entirely convincing. Rationality, as I shall understand the notion, is supposed to be sensitive to the subject’s perspective. So how can the