Hume: The Is/Ought Gap

In this paper, I argue for an interpretation of Hume's Law that sees him as dismissing all possible arguments from is to ought on the basis of a comparison with his famous argument on induction.

Examining the moral sense theories of Francis Hutcheson, David Hume, and Adam Smith from the perspective of the is-ought problem, this essay shows that the moral sense or moral sentiments in those theories alone cannot identify appropriate morals. According to one interpretation, Hume's or Smith's theory is just a description of human nature. In this case, it does not answer the question of how we ought to live. According to another interpretation, it has some normative implications. In this case, it (...) draws normative claims from human nature. Anyway, the sentiments of anger, resentment, vengeance, superiority, sympathy, and benevolence show that drawing norms from human nature is sometimes morally problematic. The changeability of the moral sense and moral sentiments in Hume's and Smith's theories supports this idea. Hutcheson's theory is morally more appropriate because it bases morality on disinterested benevolence. Yet disinterested benevolence is not enough for morality. There are no sentiments the presence of which alone makes any action moral. (shrink)

This is a commentary on Adrian Heathcote’s interesting paper ‘Hume’s Master Argument’. Heathcote contends that No-Ought-From-Is is primarily a logical thesis, a ban on Is/Ought inferences which Hume derives from the logic of Ockham. NOFI is thus a variation on what Heathcote calls ‘Hume’s Master Argument’, which he also deploys to prove that conclusions about the future (and therefore a-temporal generalizations) cannot be derived by reason from premises about the past, and that conclusions about external objects or other minds cannot (...) be derived by reason from premises about impressions. Heathcote raises an important question. Having (apparently) argued that our inductive inferences are not justified by reason, Hume puts them down to Custom, and seems to suggest that we OUGHT to indulge this propensity but NOt the superstitious propensities that lead to religious belief. (Query: Why is it right to indulge one non-rational propensity but not the others?) Finally Heathcote argues that just as there are valid, but not formally valid, arguments taking us from claims about inferential relations to claims about what we ought to believe, so there may be valid, but not formally valid, arguments taking us from factual claims about some situation to claims about what we ought to do. -/- I reply that Hume does indeed have a Master Argument and that it does rely on logical principles but not on the logic of Ockham which had been largely forgotten by Hume’s day. Instead Hume relies on the idea widely believed in the 18th Century and taught to Hume at Edinburgh by his logic Professor Colin Drummond, that in a logically valid argument the conclusion is contained in the premises. I reconstruct Hume’s Master Argument using this principle. I draw a careful distinction between two theses: 1) that we cannot get from non-moral premises to moral conclusions with the aid of logic alone and 2) that we cannot get from non-moral premises to moral conclusions with aid of analytic bridge principles. Hume believed the first but not the second. What then is the role of NOFI in the larger argument of the Treatise? To show that the truths of ethics cannot be derived via logic from self-evident truths of some other kind and thus that they are not demonstrable. How can we make sense of Hume’s apparent belief that it is sometimes right to transcend reason and sometimes not? In the case of Custom, we live in a world governed by causal regularities, and, in such a world, induction is in fact a fairly reliable belief-forming mechanism. Thus a suitably qualified spectator (one aware of the kind of world we live in) would tend to approve of indulging it, even if it cannot be justified by reason. However, our superstitious propensities are (and can be known to be) unreliable, since they produce different and inconsistent results in different people. Thus it is it is wrong (something a suitably qualified spectator would disapprove of) to indulge the faculty of Superstition. I also take issue with Heathcote’s penchant for valid, but not formally valid, inferences. I supply the missing premises for Heathcote’s Is/Ought inferences and argue that they are either not true or not necessary. (shrink)

I argue 1) That in his celebrated Is/Ought passage, Hume employs ‘deduction’ in the strict sense, according to which if a conclusion B is justly or evidently deduced from a set of premises A, A cannot be true and B false, or B false and the premises A true. 2) That Hume was following the common custom of his times which sometimes employed ‘deduction’ in a strict sense to denote inferences in which, in the words of Dr Watts’ Logick, ‘the (...) premises, according to the reason of things, do really contain the conclusion that is deduced from them’; that although Hume sometimes uses ‘demonstrative argument’ as a synonym for ‘deduction’, like most of his contemporaries, he generally reserves the word ‘demonstration’ for deductive inferences in which the premises are both necessary and self-evident. 3) That Mr Hume did indeed mean to suggest that deductions from IS to OUGHT were ‘altogether inconceivable’ since if ought represents a new relation or affirmation, it cannot, in the strict sense, be justly deduced from premises which do not really contain it. 4) That in a large and liberal (or perhaps loose and promiscuous) sense Hume does deduce oughts and ought nots from observations concerning human affairs, but that the deductions in question are not inferences, but explanations, since in another sense of ‘deduce’, common in the Eighteenth Century, to deduce B from A is to trace B back to A or to explain B in terms of A; 5) That a small attention to the context of Hume’s remarks and to the logical notions on which they are based would subvert those vulgar systems of philosophy which exaggerate the distinction between fact and value; for just because it is ‘altogether inconceivable’ that the new relation or affirmation OUGHT should be a deduction from others that are entirely different from it, it does not follow that the facts represented by IS and IS NOT are at bottom any different from the values represented by OUGHT and OUGHT NOT. (shrink)

This is a response to Stephen Maitzen’s paper. ‘Moral Conclusions from Nonmoral Premises’. Maitzen thinks that No-Ought-From-Is is false. He does not dispute the formal proofs of Schurz and myself, but he thinks they are beside the point. For what the proponents of No-Ought-From-Is need to show is not that you cannot get SUBSTANTIVELY moral conclusions from FORMALLY non-moral premises but that you cannot get SUBSTANTIVELY moral conclusions from SUBSTANTIVELY non-moral premises. And he believes that he can derive substantively moral (...) conclusions from FORMALLY moral but SUBSTANTIVELY nonmoral premises. However his argument relies on what I call ‘taxonomic essentialism’, the thesis that sentences do not change their semantic characters from context to context or from world to world. In particular, they do not change their META-ETHICAL status from context to context or world to world. If a sentence is non-moral at one world it is non-moral at all the rest. This thesis leads to contradictions (with some propositions as both moral and non-moral) and even if (as is perhaps possible) these contradictions can be avoided, it leads to further consequences that are palpably absurd. Thus It may be possible to derive substantively moral conclusions from premises that are not substantively moral, but if it is, Maitzen has failed to prove the point. (shrink)

Frank Snare had a puzzle. Noncognitivism implies No-Ought-From-Is but No- Ought-From-Is does not imply non-cognitivism. How then can we derive non-cognitivism from No-Ought-From-Is? Via an abductive argument. If we combine non-cognitivism with the conservativeness of logic (the idea that in a valid argument the conclusion is contained in the premises), this implies No-Ought-From-Is. Hence if No-Ought-From-Is is true, we can arrive at non-cognitivism via an inference to the best explanation. With prescriptivism we can make this argument more precise. I develop (...) an account of imperative consequence that underwrites Hare’s principle that you cannot derive imperatives from indicatives. Thus if moral judgments contain an imperative component, it will be impossible to derive moral conclusions from indicative or non-moral premises. Given this account of imperative consequence, we can explain No-Ought-From-Is without appealing to anything as nebulous as the conservativeness of logic. Hence if No-Ought-From-Is is true, we have an inference to the best explanation for prescriptivism. Both lines of argument face problems from Prior. Given Prior’s counterexamples, No-Ought-From-Is as originally conceived is false. The version that survives is No-Non-Vacuous-Ought-From-Is. But the best explanation of this does not include non-cognitivism. With prescriptivism it is worse. For the version of No-Ought-From-Is that prescriptivism ‘explains’ – that is, the version of No-Ought-From-Is that prescriptivism implies – would exclude Prior’s counter-examples to Autonomy as invalid. But they are not invalid. Thus Prior’s counter-examples to No-Ought-From- Is refute prescriptivism. Thus from 1960 onwards R. M.Hare was a dead philosopher walking. But if non-cognitivism cannot be derived from No-Ought-From-Is, this suggests that it is not what Hume was trying to prove. I argue that what Hume was trying to prove is that moral truths are not demonstrable. To be demonstrable, a proposition must be either self-evident or logically derivable from self-evident propositions. By Treatise 3.1.1.27, Hume had proved to his own satisfaction that no moral propositions are self-evident. That leaves open the possibility that they are logically derivable from self-evident but NON-moral propositions. The point of No-Ought-From-Is was to exclude this possibility. If you cannot logically derive moral conclusions from non-moral premises, you cannot demonstrate the truths of morality by deriving them from self-evident but NON-moral truths. I also discuss why Hume abandoned No-Ought-From-Is in the EPM. He had no need of it since he thought he had a proof that (with some exceptions) no nontrivial truths are demonstrable. Hence no non-trivial MORAL truths are demonstrable. No-Ought-From-Is drops out as unnecessary. (shrink)

In this paper, I argue that David Hume's is-ought problem stems from morality as a simplifying framework. Morality is the attempt to describe human behaviour and its relation to nature through a court framework, dating to prehistoric times. Such a court does not really exist, hence we are not referring to anything directly when we make moral statements, and therefore 'is' and 'ought' do not align. The solution is to replace morality with what it's really trying to describe: patterns in (...) individual behaviour. (shrink)