I have been asked by several people to post my "official" formulation of the metaethical theory that I favor, what I call "Expressive-Assertivism." (You can read the long version here.) So here it is.
Expressive-Assertivism is the conjunction of the following three principles.
Central Tenet (CT): If a speaker properly and literally utters a basic ethical sentence, S, of the form 'x is right' or 'x is wrong' then the speaker performs one direct expressive illocutionary act and one direct assertive illocutionary act.
Extensionality Principle (EP): If a speaker properly and literally utters a sentence that contains an ethical predicate in an extensional context, then the speaker performs a direct expressive illocutionary act.
Generality Principle (GP): If a speaker properly and literally utters a sentence that contains an ethical predicate in an extensional context, then the speaker performs a direct expressive illocutionary act expressing some conative attitude toward things of a certain kind, viz., things that have the property picked out by the ethical predicate.
Several points are in order, so, for now, I'll just state these briefly and rather dogmatically. First, (CT), (EP), and (GP) are modeled on what I take to be the correct semantic theory for "thick" ethical predicates and for "emotionally charged" predicates, like racial, ethnic, religious, sexist, or political slurs. For example, if a person properly and literally utters the sentence 'If Bob is a _____, then I'd be surprised', where what fills in '____' is some kind of emotionally charged predicate, then it looks like the speaker (i) performs a direct expressive and a direct (complex) assertive illocutionary act, (ii) the attitude is expressed even though the emotionally charged predicate occurs in the antecedent of a conditional, and (iii) the attitude expressed is directed towards things of a certain kind, namely, the group of people who have the relevant property picked out by the predicate.
Second, what it means for an attitude to be directed "towards things of a kind" is tricky question, though I think it is clear that we can have attitudes with this kind of directedness. Here's a first approximation: An attitude is directed towards things of a kind just in case a person with that attitude is disposed, upon being presented with things of the kind that he or she believes have the properties rightness or wrongness, to approve or disapprove respectively of those things.
Third, the formulation of Expressive-Assertivism as the conjunction of (CT), (EP), and (GP) does not entail that the speaker actually have the attitude that is expressed. So it allows that a person can be performing a direct expressive and a direct assertive illocutionary act in properly and literally uttering an ethical sentence, even though the person is doing so insincerely (I'm taking the sincerity condition to be merely having the appropriate attitudes).
Fourth, Expressive-Assertivism implies that, if a speaker is sincere in properly and literally uttering an ethical sentence, then that person has a pro- or con-attitude toward something. Thus, Expressive-Assertivism entails some very strong version of motivational judgment internalism.
Fifth, (CT), combined with the previous point, strongly suggests, if not implies, that a moral judgment in the form of a moral thought is neither a simple belief nor a simple desire, but a complex psychological state consisting of a belief and a desire. (It is not a besire, which I take to be a simple psychological state with two different directions of fit.) So, if a speaker properly, literally, and sincerely utters the sentence 'Lying is wrong', then the speaker has a belief that lying has the property wrongness and some kind of con-attitude toward things that are wrong. Recognition that a moral thought can be a complex psychological attitude is precisely what dissolves Smith's version of The Moral Problem. That is, only on the assumption that a moral thought is a simple psychological state is The Moral Problem really a problem.
Sixth, the official formulation does not address what the properties rightness and wrongness are. That is, Expressive-Assertivism, as articulated above, is neutral with respect to any account of rightness or wrongness. There are, however, accounts of rightness or wrongness that, if true, I would be inclined to accept as rendering the expressivist element of Expressive-Assertivism unmotivated--specifically, those that, if true, account very well for the practicality or action-guidingness of ethical discourse.
Seventh, adoption of (CT) is what makes Expressive-Assertivism a dual-use or complex expressivist theory most similar to Stevenson's emotivism and, I think, Copp's Realist-Expressivism. Adoption of (EP) and (GP) is what makes Expressive-Assertivism most significantly different from these other theories.
Finally, as I see it, the main reasons in favor of Expressive-Assertivism are these: (i) it accounts for all the data that support the "descriptivity" and "practicality" of sincere, ethical discourse; (ii) because Expressive-Assertivism is modeled on three uncontroversial features of predicates from other parts of natural languages, the three principles adopted by Expressive-Assertivism are unsurprising, credible, and realistic; (iii) Expressive-Assertivism is consistent with whatever is the correct view about truth, so it can, but is not required to, accept minimalism about truth, a view that I take still to be controversial; and (iv) though I'm still working on this, I think (CT), (EP), and (GP) work together to shield Expressive-Assertivism from all embedding objections, which, as a family, I take to be the strongest objection to Expressivist theories.