Two main foundations have been proposed for the side-constraints that deontologists think make it sometimes wrong to do what will have the best effects. Thomist views agree with consequentialism that the bearers of value are always states of affairs, but hold that alongside the duty to promote good states are stronger duties not to choose against them.1 Kantian views locate the relevant values in persons, saying it is respect for persons rather than for any state that makes it wrong to kill, lie, and so on.2 The central innovation of Stephen Darwall’s Welfare and Rational Care is to extend this Kantian idea from side-constraints to the concept of welfare, or of what is good for a person.3 As a good-to-be-promoted, welfare is usually understood as located in states of affairs. Darwall agrees that a person’s welfare involves her being in certain states, but argues that the value in these states derives from her value as a person. More specifically, his “rational care” theory of welfare equates a person’s welfare with those states it would be rational to want for her insofar as one cared for her for her sake, so an attitude to her is primary and to her states is derivative. Whereas standard theories take the concept of welfare to come first and define care as a desire for that, Darwall reverses this ordering. Alongside this account of the metaethics of welfare, Darwall defends a substantive or normative theory about which states are in fact good for persons. He proposes that welfare consists primarily in “valuing activities,” ones that involve appreciation of objects of independent 1 worth; such activities include playing music and raising children. This theory, too, has Kantian elements, since the appreciation is of objects rather than of states of affairs and is also of their worth. I will examine Darwall’s two theories of welfare, asking of each how Kantian it is and how far, when it is Kantian, it is plausible.