Chapter 3 of The Myth of Morality is called "Practical instrumentalism." One of its first claims is that, "Outside morality we know very well what reasons are," which is sort of true and false at the same time, it seems to me. Ordinary people, or anyone in ordinary life, has no trouble that I know of with the concept of reasons. Philosophers, on the other hand, seem to be able to disagree and get confused quite well about reasons. And moral reasons don't appear to be different in this regard. I think I can understand someone's doing something for moral reasons even when I disagree with their morals, but giving a philosophical account of what this means would be difficult.
Anyway, Joyce goes on to distinguish between objective and subjective reasons. You have an objective reason to do something when it will further your ends, even if you don't know it. You have a subjective reason to do something if you are justified in believing that you have an objective reason to do it.
So (normal) reasons depend on what one's ends happen to be. But moral reasons don't, Joyce argues. There is (according to what we might call folk morality) a reason not to murder people even if doing so will further your ends and not doing so will not help you in any way to get what you want. You don't, for instance, have a conscience that will be troubled if you murder them for their money.
Your ends seem to be conceived entirely subjectively here (so that you don't have any ends simply because you are a human being, say, and therefore both rational and social), but it still isn't entirely clear to me what they are. On p. 66 Joyce writes:
it is only if one is allowed to pay attention to substantive empirical belief -- that most people will not choose their "total ruin" over trivial preferences -- that one can conclude that a rational person will act in accordance with what he takes to be morally required.His claim is not that Hume is necessarily right about reason not telling us what ends to pursue, but more that it is not obvious that a rational person will be moral. It is not, he says, a truism that we expect rational agents to act morally. There is some unclarity here, too, it seems to me (at least in my head when I try to think about this stuff).
I would think that we do expect rational people to behave morally most of the time. This does not depend on any logical connection between rationality and morality. It has to do with normality, and immoral acts are the exception rather than the rule, surely. (Although, of course, in certain situations they might be the rule. But then these situations are surely exceptional.) What about the supposedly rational person who happens to desire their own total ruin? Without good reason for wanting their ruin, I think this pretty much rules out their being rational. But there are different notions of rationality, and if this is all that Joyce means, or needs for his argument, then fair enough. He does, though, want to call desires reasons, even when they are irrational, so that one can have an "irrational reason" to eat something bad for you, or, I suppose, to destroy yourself or the whole world. This seems like a misuse of the word 'reason,' but if it is meant in a technical way then, again, I guess it's OK.
I'm not sure in the end how important this is for Joyce's main argument, but that should become clearer when I look at the next chapter (which is also about reasons) and the last couple of chapters, where Joyce defends his fictionalist thesis.