Hernan, have you read about his social intuitionist model of moral judgment? (You can read about it here: http://www.nd.edu/~w.../Haidt 2001.pdf) That gets into more particulars about what he believes intuitions are, and how people often use them. For the most part, Haidt isn't making any explicit moral claims.
I have read a couple articles about him but not directly his work, I will read the above pdf.
However, I think we must be careful about how we view his work. It may well be true that he is making no "explicit moral claims", but it seemed pretty clear from the articles I read about him that he was making a claim about moral claims, i.e. that they are no more than intuition and that there is no role for higher reason other than to justify intuiton ( your #3 below.) or for regarding any intuitively derived moral claim as more valid than another. I will read his own article and see if that interpretation of his work is true or not.
If you are interested in this, however, I strongly urge you to read Thinking, Fast and Slow
by Kahneman. I think Kahneman puts Haidt in context, namely, that Haidt's observations of how people moralize is generally true but that the exceptions are more interesting than the rule. Objectivism largely operates in this exceptional space in which people at least try
to reason about morality beyond relying on intuition.
If you aren't familiar with the model, I'll explain it: (1) An eliciting situation (like being asked if incest between two consenting siblings is moral) brings about an intuition. (2) The intuition produces a judgment, absent any deliberate reasoning. (3) The individual justifies the intuition like a lawyer defending a client, which Haidt labels ex post facto reasoning. In other words, Haidt is saying that reasoning does not produce judgments, only justifications after the fact. From here, (4) your reasoning affects the intuitions of others. He also gives some caveats that reasoning (private deliberation) can alter intuitions, but only someone with the level of expertise as a philosopher can do this. Also heavy reflection may affect judgments. But I don't think Haidt explains well enough what allows private deliberation to work at all, which if evaluated further, may weaken Haidt's claims, making his model closer to how people often don't think about how an intuition originated. Post #10 by Hairnet is a good explanation of how I suspect intuitions originate (I was thinking about posting my explanation, but Hairnet covered it all). Perhaps we can rely on intuitions for some decisions, provided our prior reasoning is good. In some sense, intuitions are just automatizations of knowledge. If your knowledge is wrong, your intuition will always be wrong. When to trust an intuition is a good question, but really that has to do with knowing when to suspect you need to reject old ideas, or expand current ones. By the way, since reading more and discussing Haidt a lot in my philosophy class, I've taken the viewpoint that intuitions are not emotions. Intuitions have a relationship between emotion and reason, but is itself a unique process that is different from both.
So I have no great difference with Haidt as an observation of how people generally moralize, though I think Kahnman better explains why, but it's worth spending a moment on the bolded sentence above, and as you yourself seem to recognize when you say "people often don't think about how an intuition originated". This is, I think, a key issue and here Haidt is dead wrong in his assertion that "only someone with the level of expertise as a philosopher can do this". In fact, this is nothing more than habit formation applied to moral views. In general, we are training our intuiton all the time in a myriad of ways (inclding, but limited to, previous deliberative thinking) and there is no reason to think that expertise as a philosopher is a necessary prerequisite to this. In fact, this is a key element of Catholic teaching, what is called "formation of conscience," which is the basis of moral instruction. This is in the direction of, but not the same as, Rand's concept of the human intellect as a blank slate.
In short, if Rand overestimates the role of reason in moral judgement, Haidt underestimates it or, to be more precise, underestimtes the potential for it.
The interesting question to consider is whether those who take a more deliberative approach to moral reasoning have any advantage over those who rely on their intuitions. I strongly suspect so.
Edited by hernan, 03 March 2012 - 12:14 PM.