From this interview:
Julian Sanchez: What is it, then, that evolution can tell us about ethics? Why should an evolved moral sentiment be any more normatively binding than, say, the urge to reproduce? Does explaining ethics as one more biological response suck out the normative force, the authority of morality?
Robert Nozick: Well, there are two places where I discuss that in detail. I refer to literature in the Kantian ethical tradition that's wanting a certain strength of bindingness to ethics -- including writing by my colleague Christine Korsgaard. I doubt that they can get it, and I argue that they're requiring something stronger of ethics than we have in the case of ordinary factual belief. A comparable question to that of the bindingness of ethics -- "Why should I always do what's right?" -- is "Why should I believe the truth?" In general it's a good policy to believe the truth, but in particular cases someone might actually be better off not believing what's true.
Examples one finds in the philosophical literature are somebody who's seen the trial of a child of theirs, where they're being proved guilty of some crime that would drive the parent into a depression, maybe a suicidal depression. They'd be better off not being convinced by this evidence. Or the literature that seems to show that optimistic or even overly optimistic attitudes towards one's chances at succeeding at something, or recovering from a disease, or something like that, actually increase the chances. Maybe not up to the level of optimism one feels, but there one would be better off not being a perfectly accurate assessor of chances. In fact there's some psychological literature that seems to indicate that when people are asked by psychologists what other people in their social circle think of them, and then the psychologists check with these other people about what they actually do think, that the people who have more accurate views of what other people think of them are less happy, less successful in life, cope less well with various things, than the people who have rosier views of what people think of them than is actually the case. Now, here's another case where one may be better off believing what's not strictly true. Parents raising children might think: "Well, do I want my child to have a disposition to believe exactly what's true about other people's opinions of him or her? Or to have, not an out-of-touch- with-reality view, but a more optimistic than is actual view, a rosier view, of what people think of them, so that their life will go smoother, more easily, and so on?"
Now, there's at least a question there. And in the case of factual truth, there's not any knockdown argument, in a particular case where one would be better off believing what's not strictly speaking true, for saying that a person is epistemically required to believe the truth. If we can't do that in the factual case, why would we expect to do, and think we have to do, the comparable thing in the ethical case? That is: have a notion of bindingness so strong that in every possible situation, it requires one to do what in general is to our benefit and right, and what we've been shaped generally to do, including to be reasonably cooperative agents in social cooperation. Through the evolutionary process, those who are able to engage in social cooperation of various sorts do better in survival and reproduction. So I'm questioning the demand for bindingness in ethics. One, because nobody's delivered on that demand yet, and secondly, even outside of ethics nobody's delivered on the comparable demand.
JS: Which does seem to take some of the wind out of the objection? and yet we don't want ethical conclusions to just be these sort of interesting facts.
RN: Correct. I search around for something that is more binding than merely an interesting fact. I do say that self consciousness is something that's crucial to guiding one's behavior in an ethical way, at least according to norms and principles, and that self- consciousness is something that people often take as the distinguishing mark of being human. Maybe in that realm, if the distinguishing mark of being human is something we've come to have because of its usefulness in having us adhere to norms, that is enough of a punch behind bindingness to leave us in a satisfactory situation.
JS: At the risk of asking something so counterfactual it's not useful? does the emphasis there on the historical source of consciousness as something selected by evolution for this purpose mean that, say, my Swampman counterpart [i.e. an exact duplicate of me who springs into existence by chance, not through an evolutionary process], who arises as a mere fluke, has no reason to be moral?
RN: Well? there'd be all of the general reasons that we have for being moral in a society where we're not good at deceiving others and others can detect when we're being hypocrites. But is there some desert island case where somebody can get away with things and those reasons don't apply? Yes. The strongest kinds of argument people make when you look in the Kantian literature seem to have something to do with preserving one's own identity, or something like that. But that's presupposing a concern with one's identity. Why should one be so concerned about that? And if one is so concerned about one's personal identity and integrity and keeping that identity, it's hard to reconcile that with the strong attack on self-interested motives that the Kantians mount. To say that self-interested motives are insufficient to ground the normative force of ethics, but somehow it's based on a concern with one's own identity? that's an uncomfortable position, a position in tension.