First my standard disclaimer: this started out as a reply to Aidan's reply, in comments, to my last post (for which I thank him). But again, my rule is that anything of this length goes out front. I'll start out in the third person, but subsequent occurrences of "you" may refer to Aidan, in his post or reply.
In his comment, Aidan quotes Scanlon as finding the trivial answer unsatisfying because "it simply takes the reason-giving force of moral considerations for granted." This seems to me to conflate two different things. Consider the ambiguity of "simply" here (just as an example, to see what I'm getting at). I agree that the trivial answer does this, and that it is unsatisfactory. But its unsatisfactory nature is not due (simply!) to its doing this at all, but instead that that is all that it does – it "simply" points to it. This is trivial; we wanted more.
Reading "simply" the other way, it says that it does this (i.e., takes the r-g f. of m. c. for granted) undeniably: it does indeed do it; it simply does; there's no way around it. It does indeed do this undeniably. But that, as I've said, is not what's wrong with it. A better answer will do that too. It is not that it does it at all which renders it trivial, but the unhelpful way in which it does it. There's no explanation there, just affirmation (however correct).
So it's not that "merely asking 'why does an action's wrongness give me reason not to perform it?' shows some kind of scary conceptual deficiency." That may indeed be what one finds puzzling, and it is not at this point, nor to the very idea of moral philosophy, that I object. My problem is with the implied requirements on acceptable answers. As you say, questions of the form "why is this a reason for that?" need not imply skepticism about whether this is indeed a reason for that. But that's my line. For certain requirements on the form of acceptable answers seem to me to amount in practice to skepticism. One is required to act, in generating and evaluating such answers, as if one did not (yet) believe the conclusion – as if one were a moral ignoramus. That's what makes it seem as if there is a "scary conceptual deficiency" – not that one asks the question at all. Puzzlement does indicate "conceptual deficiency" of a more everyday kind, not necessarily scary, and we treat it with a better account of how the concepts in question function – even if this does entail telling you what you already know, albeit now in a more perspicuous form which makes (the proper use of) the relevant concepts clear.
So the disagreement, it seems to me, concerns the purpose and nature of philosophical explanation. If you demand an "explanation" for something (here, the reason-giving force of moral wrongness), and you also specify that it not appeal to the truth of that which is being explained (whether or not you claim to believe it already), then this "explanation" can only take the form of an argument from the one (thus narrowly construed) to the other. This is effectively a skeptical requirement. This is why it looks like there's a dilemma: the non-trivial horn consists of such arguments, the unacceptability of which is that in scrupulously refraining from taking the conclusion for granted (in order to meet the requirement), these arguments, however effective they may be in showing that the conclusion holds, fail to explain. We are still in the dark about what puzzles us. That much is true: neither horn of the dilemma is acceptable.
I originally put this by saying I didn't think the dilemma bit: there are other, better, explanations which, in spurning that questionable requirement, are impaled on neither horn. But I could say instead that I do think the dilemma bites those committed to the questionable (I'd probably say "Cartesian," but that's my universal term of abuse) conception of philosophical explanation, but I'm not one of those people. It is indeed a puzzle how an acceptable explanation of that form could be enlightening. Compare the skeptical paradox: it does bite those with Cartesian conceptions of knowledge – an important fact, which should indeed be rubbed into the relevant complacent faces – but the rest of us (however few we may be) may shrug it off. (On the other hand, it's not "Prichard's paradox" but his "dilemma.")
Another way to see my worry is to compare it to McDowell's rejection (e.g. in "In Defense of Modesty") of Dummett's analogous demand in the theory of meaning, i.e., that we explain the phenomenon of meaning without appealing to it at all, taking up, as McDowell puts it (either there or in "Anti-realism and the epistemology of understanding"), the perspective of a "cosmic exile." For McDowell and me, "modest" explanations (from a "participant perspective," although I'm not sure McDowell uses this term) are perfectly good, and "explanations" to cosmic exiles inconceivable. I think this is Anscombe's Aristotelian point too ("Modern Moral Philosophy").
As for your second point. (Boy, careful explanation takes time, doesn't it?) My point is that speaking of "p" allows elision of a key distinction: between propositions I believe and those I don't. Any specific example falls into one but not the other of these, and in such cases I will know which. If I have no idea whether it is true (Andromeda example), I fail to see how its bare obtaining can constitute a reason to believe it. Any reason to believe something must itself be something I am aware of. Or at least that's the "internalist" way to put it. "Externalists" would complain not of my failure to be aware of it, but of the lack of any reliable process that caused the belief. After all, I don't believe it. You were the one referring (ie in the example) to its truth: so why do you believe it? Again, "because it's true" does nothing here (just as you naturally must have demanded more than that when you came to believe it yourself, as you claim to, in the example, by referring to its "truth"). Reasons for believing p, internal or external, concern not p (by) itself but our cognitive access to it.
Now (as with the other case) this is not to say that you may not appeal to the truth of p in giving your extended explanation of how we can know this (the reasons, in either sense, for our belief). But this is (typically) in cases in which I do (already) believe the truth of p. And here again, "because it's true" does nothing by itself.
We have reason to believe that electrons are running through the wire not simply because electrons are indeed running through the wire, although they are, but also because this caused the ammeter to spike, and ammeters are built in such-and-such a way, and every alternative explanation for why such things act that way in such circumstances has been ruled out (for such and such reasons), etc. (After all, this is why I believe it.) You may appeal to the existence of electrons; but you don't have to – you can tell the story however you like. Here a scrupulously non-question-begging argument for the truth of p would be fine as well – it would, if it worked, constitute a reason to believe p. But so might a "question-begging argument," even if it (thereby) fell short of convincing us of the truth of p, as even in such cases I might still see what you were getting at, enough to see it as a (non-conclusive) reason for belief.
In either case the bare truth of p constitutes no reason to believe it (so "because it's true" is false, not trivial). There are plenty of truths we have no reason to believe; we just can't pick any of them out, because if we could that would mean we knew they were true – and thus had reason to believe them. (Again, speaking of "p", while perfectly understandable, makes this harder to see.)
I am no blind acolyte of McDowell, but here again let me compare our views to remove a potential confusion. McDowell is famous for insisting, contra Davidson and Rorty, that our relations to the world are normative, not "merely causal," and that we do indeed aim, in inquiry, to "get the world right." (It is strange that he should have to do this as, as McDowell himself points out, it was the whole point of Davidson's action theory that "reasons can be causes.") So in this sense the "world itself" (not the Cartesian objective world/Kantian noumenon, but the world qua object of our beliefs, as opposed to our beliefs themselves) can constitute a "reason" for belief. This may seem to contradict what I said above. But this only means, I take it, that, as I said, we may refer to the fact that p in explaining not only how we come to know that p but even why we should believe that p. (It is Cartesians who rule such things out; I think this is the source of the externalist animus toward "evidentialism," but I'm not sure, as externalism itself seems Cartesian as well, albeit in other ways.)