The starting point for many discussions of Kantian ethics and virtue involves either a presumption or a challenge to the effect that an account of the virtues, or an ethics of virtue, is both essential to a proper understanding of morality and alien to the Kantian enterprise. I think we should be puzzled about how that idea gets going.

After all, Kant has a “Doctrine of Virtue”; there are clear virtues that a Kantian moral agent should have; there is a nuanced and insightful, if not fully formed, story about how moral competence is acquired that is, partly, about the development of elements of a moral character necessary for good action. While it was for a long time thought that Kantian morals were legalistic, a matter of fixed and rigid rules of duty, requiring no more of the agent than conformity, that is no longer the prevailing view. There are duties and obligations, perfect and imperfect, but they do not subtend actions automatically: they depend on deliberation and situation-sensitive judgment. Where there are rules of action it is because there are regions of moral or social life where having rules is essential to successful human activity – what counts as property and wrongful taking, for example. For the same reason Aristotle also regards the virtue of justice as rule-normed. But perhaps nowadays we do not think justice is paradigmatically a virtue.