Given that Churchland focuses on the causal story (physiological, evolutionary, psychological, cultural) for where we get our moral sense, does that mean that the causal story is all there is to it, i.e. that by understanding the causal theory, you understand morality itself?
Certainly Kant thought not: the causal story is only relevant for him in figuring out how to teach people morality and the like, after the essence of morality has already been determined through a priori reasoning.
At the other extreme, you might think that we have these sentiments, and sometimes they agree, so in that sense there's something "objective" about rightness, in that it's intersubjectively verifiable, but really, it's just a shared fiction (like the value of money).
Certainly Churchland's view is closer to the latter than the former, but the vigor with which she took to discussion of actual ethical/social problems means that she's not an error theorist about ethics.
This for me was the gap in Churchland's book, and why I built Hume into the conversation. Scholars have the same trouble interpreting Hume in this respect as I'm running into here. Let me just quickly throw out my own view as I'm trying to formulate it:
I see a lot of parallels to Russell's view of mathematics here: it's conventional, but if taken seriously, it needs to be a full, coherent system, which means there's more to ethics than is given in the actual ethical views of individuals. Much like we have to posit the axiom of infinity for basic arithmetic to work (for Russell), moral reasoning involves positing an objective "right" a la Plato. So yes, for Russell, we can act as if numbers are Platonic entities, but we needn't actually adopt that as a metaphysical claim, and we shouldn't forget that the basic postulates are ultimately a matter of convention, which is underdetermined by experience. This is what pluralism in the moral realm amounts to: we start with the data of our own "moral sense," i.e. the experiences we have of judging things and those we observe in others, and our eventual moral theory has to accord in some way with those, but once you start looking at the internal logic in our disparate judgments and try to systematize them, then you end up rejecting some of your gut reactions (maybe even most of them).
The only question, then, is how much do you need to systematize these judgments? Enough to address actual moral questions you have to deal with, yes, but no more than that. I suggested in my example about abortion that we have two sentiments in play: 1. Only those beings that act like us (i.e. they talk and evidently reason) get treated as members of the moral community, and 2. Our offspring get treated as members of the moral community, especially when they're babies who can't speak for themselves. A moral realist would have to say the question "morality is morally wrong" needs to be true or false. I think Churchland ends up as a pluralist, in that the statement does have a truth value, but this can be different depending on the concrete circumstances re. the abortion in question. There are problems with that view, in that any decision re. an individual case seemingly has to rely on general principles if it is not to be arbitrary. I incline more towards moral skepticism: there simply is no truth value to the claim that abortion is wrong. Now, I'm not saying I have a great semantic analysis of this. I don't want to say that the word "wrong" needs to be stripped out of a logically perfect language and replaced with some convoluted expression referring to the speaker and his or her society and all that. The kind of ambiguity that the term has seems to me fruitful: much like "beauty," it points to the urge to systematize our intuitions and sentiments and conventions and is itself a useful tool in moving us along this road... or at least it CAN be useful in this way.