One of the most powerful arguments in meta-ethics today is that of Treatise, Book 3, in which Hume seeks to show that morality's practical influence precludes its being based on reason. H.O. Mounce, in his reviewFootnote ¹ of my Hume, Reason and Morality: A Legacy of Contradiction,Footnote ² rejects my central contention that this argument contains a contradiction. This review is however flawed on several counts.

The contradiction I point to occurs in Hume's formulations of the intermediate premise of his argument and concerns reason's role in action. Only in some formulations of this premise does Hume qualify “reason” with the adverb “alone” (“of itself”, “in itself”). In others, there is no such qualification. Thus he states both that “reason alone can never have any.. influence [on action]” and also, simply, that “reason has no influence [on action]”, or “is perfectly inert”. I argue that, far from being equivalent, these ‘qualified’ and ‘unqualified’ formulations have contradictory implications. I write (6): “the claim that reason alone can never have any influence on action would seem to imply that reason has an influence [when combined with something else]”. But this is incompatible, I say, with those formulations of the premise which, lacking the qualification “alone”, imply that reason has no influence, or cannot move us at all. Mounce however, in characterizing my position, erroneously attaches this implication of the unqualified formulations to the qualified ones. Thus he writes that, on [Botros's] view “To say that reason alone cannot move us implies that reason cannot move us at all”. But if the claim that reason alone cannot move us were to imply that it cannot move us at all, there would not even be an appearance of contradiction, since both the qualified and unqualified formulations would then have identical implications. Not only would this make nonsense of my book, it would make nonsense of Mounce's subsequent attempt to show that the contradiction is only apparent.

By the time he reaches this stage of his review however Mounce seems to have got my position clear. He evidently now realizes that reason's inability to move us at all is, on my view, an implication of the unqualified formulations of the premise not, as he had earlier led us to believe, of the qualified ones. For he objects that, while Hume may indeed state without qualification that reason cannot move us, or is “perfectly inert”, he does not mean that it cannot move us at all; he means that it can move us, though only with the help of desire, on which it is wholly reliant.

Mounce's solution has two elements—a “proof” adduced, on Hume's behalf, for this claim that reason “depends for its activity wholly on passion or desire” and an appeal to metaphor to show how something which is inert can also move us. The “proof”, Mounce declares, is that: “The same reasoning can lead to entirely different acts”. He illustrates Hume's thought with the following example: “Suppose a person reasons that a certain water is poisoned and that if he drinks it he will die. He will avoid the water, if he wishes to preserve his life. But if he wishes to end his life, he will drink it. The reasoning is the same in both cases. But the acts are entirely different. Since the reasoning is the same, it cannot in itself produce the different acts … that, roughly, was Hume's view”.

This “proof” requires us to believe that Hume was stupid. How otherwise could we imagine that he would have failed to anticipate the following, parallel, “proof” that passion or desire “depends for its activity wholly” on reason: “Suppose a thirsty traveller comes upon water but wishes to preserve his life. He will avoid the water if he reasons that it is poisoned, and that if he drinks it he will die. But he will drink it if he reasons that it is not poisoned. The desires are the same in both cases. But the acts are entirely different. Since the desires are the same, they cannot in themselves produce the different acts. Therefore desire depends for its activity wholly on reason”.

Nor can it be replied that Hume ought to have anticipated this turning of the tables, but did not. For Hume himself (as I point out in my book, 11) asserts that where a desire is constant, different reasoning can lead to different actions. He writes: “as our reasoning varies [so]our actions receive a subsequent variation”, or as in Harrison'sFootnote ³ paraphrase: “Action is the product of reason and desire in conjunction and can be changed by a change in either”. Reason's role in action is, for Hume, just as necessary as desire's; each is dependent upon the other if action is to come about. The problem, as I explain (12), when dismissing this type of solution, is to reconcile this interdependence with reason's being dependent for its activity wholly on desire, but not vice versa.

Let us consider whether anything can be salvaged for Mounce from his comparison of reason with a brick. Here he argues that reason is inert because “it cannot move itself”. “It does not follow” he goes on “that nothing can move it, or that once in movement it cannot move other things. A brick is inert or wholly inactive, but one can move it easily enough if one throws it, and one can use it to move other things if one throws it against them”. His point seems to be that reason depends on desire in that only desire can get reason moving in the first place, but not vice versa. Therefore desire, not reason, is the true instigator of action.

As I point out (176–8) the use of metaphors is a common device for coping with the apparent contradiction. Those I consider, and eventually reject, at least do justice to the distinctively Humean idea that reason directs desire as to how to fulfil its ends. As DancyFootnote ⁴ for instance suggests, reason provides the intelligence whereas desire provides the “blind urge seeking a guide, an external agent to give it direction”. But a brick thrown at an object by an agent is brute matter given direction and momentum by an outside intelligence. Mounce's metaphor, partially— and it seems unwittingly—reversing the roles of reason and desire, makes nonsense of a crucial part of the traditional Humean theory.

But even if Mounce's metaphor is incongruent with a key aspect of Hume's account of motivation, the idea of reason as inert, and only prompted to activity by desire, is captured, an apologist for Mounce might finally urge, by the idea of the brick likewise getting its momentum from elsewhere. Since Hume never contemplates reason in its turn prompting desire, do we not have here the required reason-desire asymmetry, and so a resolution of the alleged contradiction?

But it is just false (as I observe, 13) that Hume never states that reason can prompt desire. On the contrary, he cites (3.1.1:12) “prompting” desire as one of two ways in which reason or judgement may cause action (the other being by “directing” desire). This happens when (2.3.3:3) the expectation of deriving “pain or pleasure from any object” leads us to feel “an emotion of aversion or propensity” so that we are “carry'd to avoid or embrace what will give us this uneasiness or satisfaction”. PenelhumFootnote ⁵ usefully cites other examples where a particular desire is, for Hume, the outcome of a perceptual judgement. Nor is this account of reason's activity confined to Book 2 and 3. In Book 1 (3.10:4) Hume declares that inductive reasoning or belief “is almost absolutely requisite to the exciting our passions”. Those who resent being weaned off that caricature of Hume's views which often passes, in meta-ethical circles, for direct acquaintance with the text will naturally cast around for ways of discounting these remarks. They certainly clash with the portrayal of Hume as an uncomplicated instrumentalist who restricts reason's role to that of informing us how to satisfy a desire we already possess, or of producing a secondary desire for that which is a means to satisfying a primary one. Far from being disquieted by these remarks however, Mounce makes no mention of them.

I would not wish to suggest that ingenious attempts to resolve the contradiction cannot be found in the literature. It is simply that the above solution, as my book makes clear, is not one of them. Moreover it is unjust to give the impression that I overlooked it.