In a practical philosophy, […] we have to do not with assuming grounds for what happens but rather laws of what ought to happen even if it never does, that is, objective practical laws."(Kant, G 4:427) 1 "We will not talk about 'oughts' at all: that is how you talk to children, or to nations in their infancy, not to those who have acquired all the culture of a mature age." (Schopenhauer, WWR 1, §53, 320)Kant's statement expresses the still widely accepted view that morality takes a prescriptive form: it concerns what in some sense ought to happen, even if it never does happen. An ought, Kant tells us earlier in the Groundwork, is that through which an imperative is expressed, and an imperative is the formula of a command or law (of reason) (G 4:413). Morality is thus from this perspective a system of oughts, imperatives, commands, laws. And on the common view, as on Kant's, these oughts or imperatives of morality are thought to bind us in a particularly strong fashion, that is, categorically or absolutely. In short, they tell us what we must do, or must not do, come what may. This conception of morality as a system of categorically binding oughts or imperatives has more recently come under fire, perhaps most notably at the hands of Elizabeth Anscombe (1958) andPhilippa Foot (1972). 2 To my mind, however, the most forceful and most interesting critique of the prescriptive conception of morality, though one not so well known today, was advanced more than a century earlier by Arthur Schopenhauer, first in his The World as Will and Representation (1818) and more fully in his unsuccessful prize-essay On the Basis of Morality (1839). My aim here will be to reconstruct and sharpen his critique, and to argue that it does in fact cast serious doubt on the prescriptivist conception of morality.I understand Schopenhauer's critique to consist of four main objections. According to the first, which I consider in Section 1, Kant begs the question by merely assuming that ethics has a prescriptive or legislative-imperative form, when a purely descriptive conception such as Schopenhauer's also presents itself as a possibility. In Section 2, I set the stage for the remaining objections by elucidating, sharpening, and 1 See also CPR A802/B830; MM 6:216. 2 See also Slote (1982); Williams (1985: 174-96); Taylor (2000: 139-77); Taylor (2002: 77-84); see Capaldi (1966) for an argument that Hume means to reject "ought" as a moral category. defending a principle of Schopenhauer's on which they all depend: namely, that a binding ought must be understood to presuppose, and be conditioned by, a threat of punishment or promise of reward. In Section 3, I turn to Schopenhauer's second objection, which anticipates Anscombe's point that the notion of a moral ought loses all intelligibility outside the framework of a divine law conception of ethics. I then note that, whereas Anscombe settles for arguing that we should jettison the moral ought (to the extent that we can) merely because, as a matter of historical fact, we have largely abandoned it...