Wednesday, February 22, 2023

A Refutation of Hume’s Law

 Hume's law or Hume's guillotine[1] is the thesis that, if a reasoner only has access to non-moral and non-evaluative factual premises, the reasoner cannot logically infer the truth of moral statements. (Wikipedia)

I do not claim that I can infer the truth of ought statements from is statements but I claim that it would be possible to do it if a suitable set of is statements turned out to be true, hence that Hume’s law is not true in the sense in which it is usually imagined to be, as a conclusion depending only on logical argument not on factual claims.

My argument starts with intuitionism, the philosophical position that holds that just as humans have senses such as sight and hearing that imperfectly sense physical facts so we have a moral sense that imperfectly senses moral facts. For a book length exposition see Ethical Intuitionism by Michael Huemer, for a short sketch, including my response to obvious counterarguments, Chapter 61 of The Machinery of Freedom.

The argument for intuitionism, beyond the fact that it describes how most people feel about morality — that certain acts really are wicked — is consistency. We believe that our physical senses are not lying to us about physical reality because what they report usually passes all the consistency tests we can subject them to, consistency both between information reported to us by sight, hearing, smell and touch and between the observations of physical reality made by different people. It is possible that it is all an illusion — what I know about what other people perceive ultimately reaches me through my senses, which could all be lying to me — but it is the best evidence we have available. If moral perceptions are similarly consistent, if almost everyone, given a sufficiently well described situation and action, will have about the same moral response, that would be evidence that there is a moral reality out there which we are perceiving.

Whether that situation exists, whether almost everyone has about the same moral perceptions, is a fact of reality, a non-moral fact. Suppose it does. One might still reject the conclusion on the basis of an alternative explanation. Perhaps there are no moral facts, just moral beliefs, consistent because they were produced by biological evolution hard wiring into us beliefs that cause us to behave in ways that lead to reproductive success, or societal evolution producing societies that indoctrinate their population into the set of moral beliefs that that make a society more likely to survive. 

This alternative explanation, however, is subject to factual tests. One could imagine evidence showing that some widely held moral belief did not contribute to reproductive success or societal survival. I do not claim to have any such evidence but it is at least logically possible. If it existed, and if we observed consistency across humans of moral judgement, that would be evidence for a moral reality that humans could perceive; their common perceptions would be evidence of moral facts just as ordinary perception is of physical facts. Hence it is logically possible to deduce ought from is. 

It may be that moral nihilism is correct, that intuitionism and other forms of moral realism are wrong, that the necessary facts are not true, but they could be. Hume’s law is not a claim about what facts exist but about the logical impossibility of deducing moral facts from physical facts and I believe I have shown that it is false.

I have shown the possibility of evidence, not of proof, but that is true of all our factual beliefs. I cannot prove that the sun will rise tomorrow or that the Earth is round, I can only offer very strong evidence for those claims. I believe that the is-ought claim as commonly understood, certainly as I understood it, applies to evidence for moral facts as well as to proof of moral facts.


Unknown said...

I think Huemer makes a similar argument in his paper, "A Liberal Realist Answer to Debunking Skepticism: The Empirical Case for Realism" in sections "4.1. A Darwinian Trilemma" and "4.2. Changes in Gene Expression". However, I think he also accepts Hume's Law as discussed in Ethical Intuitionism, section 4.3 The is-ought gap: "I see Hume's Law as an important lemma in my larger argument that one cannot account for moral knowledge without appealing to something like ethical intuition."

Emil Karlsson said...

The argument for intuitionism presented here seems to me to be an abductive (inference to the best exlanation) argument. Not a deductive one. The way I've understood Hume's Law its only meant to be applied to deductive inferences. So as I understand it, I don't think you've refuted it. Given a wider interpretation of Hume's law I think I agree with you though.

SB said...

I've never encountered this epistemological definition of "intuitionism". It feels analogous to the concept of "mathematical realism", the hypothesis that mathematical concepts are really "out there" waiting to be discovered rather than being a pure invention of the human mind. (Neal Stephenson's novel Anathem is set in a world in which debates over this are as consequential as, say, Catholicism vs. Protestantism in 16th-century Europe.)

Interestingly, mathematics also has something called "intuitionism", but it's very different. It can be oversimplified to say "we reject non-constructive proofs": you can't claim to have proven "P or Q" unless you know which one you've proven, you can't claim to have proven "there exists an x such that P(x)" unless you can point to a specific such x, and so on. It turns out that most familiar proofs are constructive, or can be tweaked to make them so, but there are notable exceptions.

SB said...
This comment has been removed by the author.
SB said...
This comment has been removed by the author.
William H. Stoddard said...

I don't find this persuasive.

Consider the idea of folk sciences. There is, for example, folk physics, which contains such intuitions as that when an arrow is fired, it shoots out in a straight line until its impetus is exhausted, and then plummets to the ground; various experiments have purported to show that this is how people naturally think. But in fact its course is a parabola. There is folk taxonomy, which divides animals into beasts, birds, snakes, fish, and crawly things (nicknamed "wugs" by some anthropologists); but that's not actually how animals are classified in scientific taxonomy.

So if ethics is a science, there is no reason to suppose that our intuitions about what is ethical necessarily correspond to what IS ethical. Folk morality may not be any more accurate than folk physics.

David Friedman said...

The intuitionist claim is not that we correctly perceive moral theories, any more than the usual claim about objective reality is that we correctly perceive physical theories. It is that we correctly perceive moral facts. We can then try, successfully or not, to build theories on those facts, just as we build, successfully or not, physical theories on perceived physical facts.

Matthew Muñoz said...

Hume would not allow for empirical consistency as a criterion; he’d throw the problem of induction at you (and then he'd shrugs his shoulders and admit your position is psychologically inevitable).