I attempt first to disentangle three aspects of Kant’s distinction between perfect and imperfect duty. There is the central distinction between principles of duty contrary to that which is contradictory in conception/consistent in conception but contradictory in will. There is also a distinction between essential and non-essential duties: those which cannot, or occasionally can, be passed over consistent with the requirements of morality. Finally, there is a distinction between duties that exhibit a scalar aspect – degrees of goodness or virtue – and duties that do not. My aim is to show how these distinct considerations can be reconciled as aspects of a single distinction, and I conclude that the remarkable complexity of Kant’s perfect/imperfect distinction is actually a strength, rather than a weakness.

Kant’s theory of friendship is crucial in defending his ethics against the longstanding charge of emotional detachment. But his theory of friendship is vulnerable to this charge too: the Kantian sage can appear to reject sympathetic suffering when she cannot help a suffering friend. I argue that Kant is committed to the view that both sages and ordinary people must suffer in sympathy with friends even when they cannot help, because sympathy is necessary to fulfill the imperfect duty to adopt others’ merely permissible ends (MPEs), and we ought to take friends’ MPEs as our own. MPEs are individuated in terms of concepts which include marks of the first person, and no marks of law other than permissibility. To adopt ends of others individuated in terms of such concepts rather than merely promote them as means to different ends, those concepts must engage with one’s feelings in a way that requires sympathy.