A system of duties. In addition to this significant change in the meaning of its title, here are a number of other things in the Metaphysics of Morals which ought to surprise anyone whose image of Kantian ethics is based on the earlier, more foundational works. In effect, however, this means that the Metaphysics of Morals ought to be both surprising and enlightening to most Anglophone moral philosophers, since their image of Kantian ethics is derived almost exclusively from the Groundwork and the second Critique. Even those who have dipped into the Metaphysics of Morals have seldom let it shape their conception of the basic principles and standpoint of Kantian ethics as they have gotten it especially from the first fifty or so pages of the Groundwork. They have almost never let it significantly influence their interpretation of what they have already read in those pages. Consequently, the familiar image of Kantian ethics is in serious error on some fairly basic points.
For example, it is almost universally supposed that Kant’s conception of ordinary moral reasoning is that when considering a course of action, we should formulate the appropriate maxim and decide whether it can be universalized. Kant’s admirers, in fact, as well as his critics, tend almost by reflex to think of the universalizability test as his most (or even his only) significant contribution to moral reasoning. But the universalizability test is used very seldom in the Metaphysics of Morals. In fact, it is used exclusively in connection with a single duty: the ethical duty of beneficence to others (MS 6:393, 453). Perhaps this should not have come as a surprise, since the Metaphysics of Morals is a system of positive duties and the universalizability test is almost exclusively of negative import, used mainly in deciding whether a given maxim is permissible or impermissible rather than in establishing positive duties. The case of beneficence to others is in fact the only one where it can be used to ground a positive duty, since in Kant’s view there is only one end which all human beings have necessarily, namely that of their own happiness. Hence it is only in this one case that we necessarily adopt a maxim (that of self-love) and therefore have a duty to adopt it only in a form which may be universalized – namely, that which includes also having the happiness of others as an end (MS 6: 453).
In the Metaphysics of Morals, Kant conceives of ordinary moral reasoning instead as the prioritizing, weighing and balancing of duties – and of the obligating reasons (rationes obligandi, Verpflichtungsgründe) based on them (MS 6:224). Some duties, those which are strict or perfect, require specific actions or omissions; others, the wide or imperfect duties, require only the setting of ends. With wide duties there is consequently considerable latitude for different agents, or for the same agent at different times, to decide how far and by which actions she will promote these ends. It is only in the case of strict duties that the performance or nonperformance of an action is wrong or blamable; actions in promotion of the ends grounding our wide duties are meritorious, but the omission of any specific action of this kind is not wrong unless it involves the general abandonment of the required end. Kant’s category of wide duties thus encompasses much of what others prefer to categorize as ‘supererogation’ and regard as falling altogether outside the scope of duty properly speaking. Kant, however, thinks that the concept of duty applies to such actions because we can, and sometimes must, rationally constrain ourselves to perform them.6