A week's worth of comments and emails about the "After Thought" project have reminded me that the words "rational" and "irrational" can't be bandied about loosely. Talk about the subject risks falling into what the historian David Hackett Fischer calls "the fallacy of equivocation"—where "a term is used in two or more senses within a single argument, so that a conclusion appears to follow when in fact it does not." Before this project can move on, we're going to have to get precise about the different uses of the word "rational."
As far as I can see, when people discuss "rationality" they can mean any of five different concepts.
1. Rational in the sense of explicable: When we say something has a "rational explanation," for instance, we mean simply that causes have effects, that logic works, and that the world operates in accord with laws that we can discern. If the world were not subject to physical laws—if the freezing point of water was different next week and evolution inapplicable on Wednesdays—then we couldn't discern any causes below surface appearances and temporary circumstances. So we couldn't really explain anything. Having a conversation about this topic, then, presumes that this meaning of "rational" is not in dispute.
Post-rationalists can't and don't deny that the universe operates according to physical laws. But they do deny that people use those laws to override the built-in biases of the mind. For instance, the laws of probability are pretty clear that buying lottery tickets is close to just taking the money out of your pocket and setting fire to it. But knowing this and owning it—using the knowledge to guide one's own behavior, with the sheer force of your conscious mind—are two different things.
2 Rational in the sense of objectively optimum according to logic. Suppose you decide to stop eating for a long time. You might for some reason think this made perfect sense. You might, for that matter, be part of a community of ascetics who support you and share your opinion. However, a good argument exists that you are wrong in some objective sense—that the best choice accords with reason, and that it involves not dying of hunger. In this case, we're saying "the rational thing to do" exists, whether or not any real person is actually choosing to do it.
Interestingly, this sense of rational is very important to behavioral economists—precisely the scholars who say Reason doesn't rule our mental roosts. After all, the narrative of a behavioral-economics argument boils down to this: "The rational thing to do is avoid obesity, save for retirement, shun the risk of catching a sexual disease, buy the sensible car, etc., but, hey, that's not what people do." How do they know what "the rational thing to do" is? Because they rely on "rational" in the sense I'm describing here.
3 Rational as in "this makes sense, just not at for the individual." Let's stick with the example of hungry you, engaged in your fast. Suppose your dropping dead of hunger is the key act in the process of overthrowing an awful dictatorship (not far fetched—this year's Arab uprising began with the self-immolation of a single man). Your sacrifice, then, would be irrational for you as an individual but a sensible move for your political party. How many hunger strikers have risked death and felt that their decisions were rooted in a sound, inevitable logic? For that matter, how many soldiers have died because it was rational for the army as a whole to sacrifice some fighters in one place?
Many evolutionary biologists invoke this kind of rationality to explain human behavior. The soldiers who sacrifice their lives for the army, it has been claimed, are increasing the prestige of their families, thus helping relatives get mates and resources; or they're doing it because it increases the odds that others will do the same for their kin; or they're doing it because culture has fooled their genes into acting as if fellow-citizens were actually cousins. I'm agnostic about these different ideas, but I want to note what they have in common: They say that the actions are rational for your genes even if they don't seem to make sense for you. When "rational" is used in this sense, it is a claim that human choices are logical at some level of analysis other than the individual. This sense of "rational" has moved the locus of reason from your mind to a different level.
4 Rational in the sense of self-interested. I think of this concept as the mirror image of the one above. Instead of pushing the locus of reason down to genetic level or up to the national stage, this concept insists, a priori, that only what benefits the individual counts as a rational choice. As such it doesn't captivate historians, political scientists, ethicists or biologists.
Who it does captivate, apparently, are economists. As far as I can tell, the notion that rational must equal self-interested is enshrined in economic theory. Here, it serves to defend not a general model of the human mind but a theory of how markets work. They are supposed to be efficient in allocating goods and services because all participants, each in pursuit of his own interests, collectively find the true value of goods and services. Being skeptical about this idea doesn't necessarily commit you to skepticism about any other meaning of "rational."
5 Finally, we have "rational" in the sense it's usually used in ordinary conversation: Contrary to innate biases and intuitions. This is the meaning we intend with images like "my heart tells me to take the job, but my head says no." It's a synonym for mastery of the self, based on facts and evidence as opposed to magic charms.
Of course, these different senses of "rational" aren't necessarily contradictory. In this passage from "Happiness Research and Cost Benefit Analysis" (pdf), the law professors Matthew Adler and Eric Posner in fact invoke all five meanings in laying out their definition of human well-being: "individual well-being," they write "consists in those things that individuals [Rational-4], with full information and deliberating rationally [Rational-5], contemplating the prospect of living different lives, converge [Rational-2] in self-interestedly [Rational-3] preferring."
But if rationality can mean a combination of the five senses I mentioned, it doesn't have to. Instead, Meaning 3 could imply a different choice than you'd get with Meaning 5 (your genes might want you to sacrifice your life for their rational reasons, but you might be able to master that impulse and run the hell away instead). Meaning 4 certainly has contradicted Meaning 2 in many people's experience.
In any event, any discussion of how people are, or aren't, rational, is going to stall unless the participants agree on which sort of "rational" they're debating.
Illustration: A section of Ingres' 'Oedipus and the Sphinx.' Source: Wikimedia.