Hume's Sentimentalism: A Path to Moral Realism? — Epoche B2
Corrigible Sentiment: What Hume's Moral Theory Delivers, and What It Does Not Two ordinary cases A neighbour helps me carry a wardrobe up three flights of stairs. In the same week I read that a famine has killed thousands of people I will never meet. The first event moves me; the second, honestly reported, moves me much less. Yet I do not for a moment judge the wardrobe the greater matter. Second case: a colleague quietly takes credit for work I did, and I am furious; a year later I do the same to someone else, and I feel almost nothing. I judge both acts wrong, and I judge them equally wrong. In both cases the felt sentiment and the moral judgement come apart, and the judgement corrects the sentiment rather than reporting it. That is the datum this essay starts from, because it is the datum that decides what David Hume's moral philosophy can and cannot support [1] . The question here is not the broad one of whether Hume was a non-cognitivist [2] . It is narrower: given that Hume's own theory explains the correction, does the explanation reach as far as moral realism? I argue that it delivers cognitivism — moral judgements admit of correctness and error — and stops short of realism , and that the fashionable route from the one to the other, response-dependence, does not close the gap. What would refute me is a specification of Hume's corrected standpoint in terms that are not themselves moral; the fifth section explains why I think there is none. Hume's sentimentalism Hume's central negative claim is that moral distinctions are not derived from reason. In A Treatise of Human Nature (1739–40) he argues that reason concerns relations of ideas and matters of fact, that it is inert, and that morals influence action; so morals cannot be the product of reason alone. The best-known formulation is Book II: Reason is, and ought only to be the slave of the passions, and can never pretend to any other office than to serve and obey them. Book III applies this to morality. Look for the viciousness in a case of wilful murder, Hume says, and you will not find it among the object's qualities; you find it only when you turn reflection into your own breast and find a sentiment of disapprobation. Vice and virtue, he concludes, are like colours and sounds as modern philosophy conceives them — not qualities in objects, but perceptions in the mind . The same section contains the famous observation that writers on morality slide without warning from statements joined by is to statements joined by ought , and owe an explanation of the transition. It is easy to see why this reads as an ancestor of twentieth-century non-cognitivism, and Simon Blackburn's projectivism explicitly presents itself as Hume's heir [3] . Non-cognitivism holds that moral judgements do not express beliefs and are not true or false; they express attitudes. If moral judgement just is a sentiment of approbation, the non-cognitivist reading looks compulsory. Why the non-cognitivist reading strains It strains against the two cases above, and Hume knew it. Sympathy, the mechanism by which we come to share the feelings of others, is his explanation of moral sentiment; but sympathy is stronger for those near us in place, time and relation. If a moral judgement simply reported the sentiment sympathy actually produces, my judgements would swing with proximity — the wardrobe would outrank the famine — and they would be worthless for the purpose Hume thinks morality serves, which is to give people who differ in situation something common to talk about. Hume's remedy is the general point of view. We learn to disregard our own position and to judge a person's character by its effects on those in his ordinary circle, from a standpoint we can expect others to share. Hume is careful and honest about how far this works: he notes that we succeed in correcting our language more reliably than our feelings , which are often stubborn. That admission is exactly what the cases require, and it is not what a pure expression theory would predict. A shout of "boo!" does not have a corrected form. A judgement does. Cognitivism, and why it is not yet realism Two claims must be kept apart. Cognitivism holds that moral judgements are assessable as correct or incorrect, so that "you are wrong about this" says something more than "I feel differently". Realism adds that what makes a moral judgement correct does not depend on our attitudes. The general point of view gives Hume the first. There is a standard; judgements made from a partial standpoint can fail to meet it; the failure is a mistake and can be pointed out, which is what happens in the wardrobe case. So Hume is a cognitivist in the sense that matters for ordinary moral argument. He is explicit in refusing the second. The standard is fixed by human sentiment. In the Enquiry Concerning the Principles of Morals (1751) he insists that the final verdict in morals rests on a sense or feeling that nature has made common to the species, and he treats this as the answer to the question about morality's foundation rather than as a difficulty. A standard fixed by the constitution of a species is not attitude-independent. Hume is not a realist, and it is not a defect in him; it is his position. Response-dependence, and why it is not realism The tempting move is to say that this is realism, of a response-dependent kind. A response-dependent property is one constituted by a disposition to produce a response: redness, on one familiar account, is the property of looking red to normal observers in normal conditions. Such properties are not "in the object" independently of us, but claims about them are perfectly truth-apt, and people get them wrong all the time. Read Hume this way and moral goodness becomes the property of eliciting approbation from a judicious spectator at the general point of view — objective enough for truth and error, grounded in nothing spooky. This is an illuminating redescription of Hume. It is not an upgrad