John Hare’s God’s Command, Chapter 3, “Eudaemonism,” Section 3.2.3: Two Errors of Kant:

Eudaemonism is a single-source view. Before looking at four defenses of it, we need to face two difficulties with Kant’s account of morality and happiness. Fortunately we can modify Kant’s own account in order to overcome these difficulties without losing the argument from providence, and the modification will remove some distractions.

The first difficulty is with Kant’s account of happiness, and the second with his account of morality. Both problems come from Kant’s overstrict dichotomies. Happiness, he says, is “a rational being’s consciousness of the agreeableness of life uninterruptibly accompanying his whole existence,” and he goes on to say that “to make this the supreme ground for the determination of choice constitutes the principle of self-love.” Kant here ties happiness to pleasure, pleasure that derives from the satisfaction of one’s inclinations as a sum. In turn, inclinations are defined in terms of the lower faculty of desire, and their satisfaction is something empirical, and can’t therefore determine practical (necessary) laws. To try to make happiness the ground of morality would therefore be to lose morality altogether.

Despite first appearances, though, Kant is against what he calls a “morose” ethics, which sets morality in opposition to all pleasure and renounces all concern of moral persons for their own happiness. He distinguishes, it’s true, between what he calls “practical love” (the will’s obedience to the moral law) and “pathological love” (which is a feeling such as sympathy and compassion), and denies that the latter can be commanded, whereas the former is “the kernel of all laws.” This might lead one to think that the state that the moral law commands is one in which inclinations do not appear at all. But in fact Kant says that to love God with practical love (which can be commanded) means to do God’s commandments gladly, and that to love one’s neighbors means to practice all duties toward them gladly. He has in mind a translation of the theological doctrine of sanctification, in which our inclinations become over time more and more in line with duty. In the resultant state, our wills will be in conformity to duty for its own sake, and this deserves what Kant calls “esteem,” and our inclinations will also conform to what duty requires, and this deserves what he calls “praise and encouragement.” But merely including inclinations in this way is not enough.

There are two revisions we need to make to Kant’s account of happiness. The first is that Kant needs to acknowledge a kind of “gladness” that is not merely the satisfaction of sensuous inclination. He needs an account of not-purely-sensuous moral pleasures, such as the awe we feel in the presence of the moral law within, or a delight in goodness that is like the astonishment at the wisdom displayed in the order of nature, an effect “stimulated only by reason.” [Think of the pleasure that comes from reading great books.] But this kind of “higher” pleasure is never properly integrated into Kant’s account. The second revision is that it is better not to insist on tying happiness to pleasure at all, even if we continue to index the content of happiness to the agent. There are many self-indexed goods, such as accomplishment, which are only derivatively pleasures. That is to say, we get pleasure from them only because we antecedently think of them as good.

The difficulty with Kant’s account of morality is that he holds that motivation is either by self-indexed inclination or by universal moral principle. This dichotomy is a mistake, and it is interesting that Scotus makes very much the same mistake. After elaborating the distinction between the two affections, Scotus proceeds to argue that every motivation that is for justice rather than advantage is for God, and so the choice is always: God or self. But surely I can be motivated to achieve something for, say, Peter, without this being self-indexed by my caring essentially that Peter is in some special relation to me or that the result be achieved by me. My motivation here is indeed indexed to a particular and I may not be motivated to pursue similar good things for other similar people. On Scotus’s dichotomy, this motivation does not belong under either the affection for justice or the affection for advantage.

In the same way Kant holds that the first formula of the Categorical Imperative, the formula of universal law, requires the eliminability not only of self-reference, but reference to any particular person. But we need an intermediate category, of inclinations that are not universal but that are indexed not to the self but some other individual. This is not only a terminological question, whether to call a principle “moral” if it contains ineliminable reference to, say, Peter. There is a substantive question about whether to have the highest kind of admiration (what Kant calls “esteem”) for a person who acts on such a principle. Hare thinks we should, and will argue so in a later chapter. One way to put this point is that the two formulations of the Categorical Imperative can come apart on one plausible interpretation of the second (though not on Kant’s interpretation of it). It is possible to care for another person as an end in herself but not be willing to eliminate reference to her from the maxim of one’s action.