It turns out you can't prove any conclusion by logical argument, if your opponent simply refuses to accept it. At most, you can reveal a logical contradiction in their views. And at this point, someone who stubbornly refuses to accept your viewpoint -- unless they've agreed to follow some rule that forbids it -- can change their assumptions to fit their desired conclusion. This actually happened in mathematical set theory, so it has a respectable history. In the case of morality, let's say we show that everyone has a natural instinct to follow our preferred morality. If Odysseus wants to go on killing people, he can argue that true virtue requires fighting our instincts. Actually he'd just hit me, but you see my point.
Some will see this as a flaw in reason. People like Leon 'bring down the Towers of secularism' Kass seem to think that reason can never produce morality -- at least not morality that religious people can accept -- and that we need theocracy to fix the problem. (Or he may want anarchy, I can't tell which.) Well, I've already addressed the claim that God provides believers with 'objective' or absolute morality. And I think this focus on abstract objections misses the point entirely. Rational morality doesn't mean deriving morals from set theory, or even string theory. It means giving an argument that most humans will accept, for rational reasons.
REBT seems to argue against the claim that we need religion to teach people morality. For those of you who skipped my rough drafts, I found out about this therapy while researching general-semantics. It outperforms other non-chemical therapies in controlled studies. And it works by helping people to pursue "rational goals". I admit that if someone disputes the rationality of these goals, we have no way to prove it. But does this actually matter in practice? It seems like pretty much everyone does agree with the goals. Indeed, some critics call REBT hedonistic and egoistic because its goals contribute to long-term personal happiness. (To which Dr. Ellis might reply that if you look at actual studies instead of arguing about noises, you'll see that his therapy also reduces antisocial behavior.) Humans don't like to hurt people who we see as similar to ourselves, and reason tells us to regard everyone as similar to ourselves. So, to get back to moral education and the religious argument, it seems to me that reason can take care of teaching morality without using the name of any deity. We could convince children of my preferred morality using reason and experience. (This may not always convince adults who believe in a different morality, but I think it has a better chance than converting them to a hostile religion!) Using their instincts to silence abstract objections doesn't seem dishonest, because if we teach them REBT and some form of general-semantics, the slightest thought on the matter will lead anyone who cares back to those objections. And I think my preferred "rational goal" morality can easily survive such investigation. Some will say that, if we go back far enough in history, Christianity helped give one instinct the largest role in morality. (Odysseus would have stressed the virtue of other desires.) This seems plausible enough, but what of it? Thinking about the influence of my environment does not lead me to go and kill people, because it does not change my desires (learned or otherwise). It does not change my True Will in this regard. And however my morality started, I don't think we need religion to teach it.