Okay, yeah, so belief in belief is a thing. We can profess opinions that we've been taught are virtuous to hold without deeply integrating them into our worldview; and that's probably increasingly common these days as traditional belief systems clank their way into some sort of partial conformity with mainstream secular ethics. But at the same time, we should not automatically assume that anyone professing traditional values -- or for that matter unusual nontraditional ones -- is doing so out of self-interest or a failure to integrate their ethics.

On the contrary, I think it's quite reasonable to assume that somebody who bases their morality on religious background has not integrated these preferences and is simply confused. My objection here is mainly in case somebody brings up a more extreme example. In these ethical debates, somebody always (me this time, I guess) brings up the example of Islamic sub-groups who throw acid in the faces of their daughters. Somebody always ends up claiming that "well that's their culture, you know, you can't criticize that. Who are you to say that they are wrong to do so?". In that case, my reply would be that those people do not actually have a preference for disfigured daughters, they merely hold the belief that this is right as a result of their religion. This can be seen from the fact that the only people who do this hold more or less the same set of religious beliefs. And given that the only ones who hold that 'preference' do so as a result of a belief which is factually false, I think it's again reasonable to say: No, I do not respect their beliefs and their culture is wrong and stupid.

Setting aside the issues with "terminal value" in a human context, it may well be that post-Enlightenment secular ethics are closer in some absolute sense to a human optimal, and that a single optimal exists.

The point is not so much whether there is one optimum, but rather that some cultures are better than others and that progress is in fact possible. If you agree with that, we have already closed most of the inferential distance between us. :)