Hume's Morality: Feeling and Fabrication (review)

Hume's project, in Book 3 of the Treatise, of showing that virtue and vice are discerned by feeling, not reason, is notorious for its contradictions. Armies of Humean scholars have fought valiantly, ingeniously, but unsuccessfully, to resolve them, and in the first half of Hume's Morality, Cohon shows herself an admirably doughty follower in their footsteps. The second half concerns Hume's division between natural and artificial virtues. We learn how self-interest is redirected, and moral sentiment strengthened to provide artificial virtues with the motives they lack naturally, so as to enable them to facilitate impersonal co-operation, though a particular virtuous act may benefit no-one. The author explains why, for Hume, their real nature as prosthetic virtues compensating for our human deficiencies must be hidden, and the paradoxes thereby raised. Part 2 is less controversial, and I shall concentrate on the meta-ethical arguments of Part 1 about which I have a number of critical reservations.

Cohon's subject in the first three chapters is Hume's argument from morality's link with action to the conclusion that it cannot be grounded in reason. A major challenge, as she sees it, is to render the other premise—that reason is inert—compatible with Hume's claims elsewhere that beliefs can, without help from independent passions, influence passion and action. She suggests (66, 68) that reason is for Hume distinct from belief. The former always signifies a process of reasoning and is causally inert, the latter an outcome, whether of reasoning or an irrational process, and is causally efficacious. It does not follow, she argues, from the fact that reasoning can result in beliefs, and that beliefs can alone produce passion and action, that reasoning itself can do so. This would only follow if reason caused the belief which caused the passion, and only assuming "a transitivity of causes" (for which she sees no evidence in the Treatise). For "there is no [parallel] transitivity of process" (74).

Having assured us that, once moral discrimination is also treated as a process the motivational argument is valid, she moves to Hume's positive phase. A second potential inconsistency arises from her interpretation of Hume as both a moral anti-realist, holding that vice and virtue do not exist independently of human psychological responses, and a moral cognitivist (100). The positions can, she believes, be reconciled, within a "moral sensing view" which likens moral attributions to those of secondary qualities. We form true beliefs about an object's colour though we do not suppose that its colour inheres in it, but is a relational property existing between us and the object. She endorses, on Hume's behalf, moral impressions which are felt sentiments, and moral ideas which copy them. This tiered approach is used to explain how Hume can consistently hold that moral discriminations are feelings, and yet accord a major role to inferential reasoning in the adoption of that steady general standpoint distinctive of morality.

Cohon is right that, for Hume, beliefs alone produce passions and action. But this is not novel. Thirty years ago, Penelhum (1975; since followed by this reviewer among others) scorned those who sought to prevent a conflict between this claim and the characterization of reason as inert by inventing a pre-existing free-floating desire to explain why just believing that there was within reach a ripe fruit could, according to Hume, generate a desire for it.

Her attempt to resolve the contradiction is more novel but less obviously right. Of course, Hume sometimes means by "reason," the process of reasoning, just as we do when the context demands it. But he plainly attributes to it the causal force she denies. At Treatise 3.1.1.16 (SBN 462-63), he states: "Reason . . . may . . . be the mediate cause of an action, by prompting . . . a passion." Here, as at 3.1.1.12 (SBN 459-60) (cited by Cohon) where Hume asserts that "reason . . . can have...