Oh, so you don't trust me, huh? Well, at least trust Martin Daly and Margo Wilson, two of the finest evolutionary psychologists in the known universe. It was in conversation with them that I first realized how thoroughly kin-selected genes fail to act in their own self-interest once they permeate a population. As Daly and Wilson have noted, by the time a gene for discriminatory altruism permeates the population, and could, in theory, help itself by becoming more diffusely altruistic, it's too late for this sort of self-help; the gene can't "change its mind" about how it affects behavior. It is, after all, just a bit of code, no smarter than the strings of ones and zeros that underlie each word in this sentence.

Daly, Wilson, and Catherine Salmon have analyzed this issue in a chapter of a forthcoming book (Evolutionary Social Psychology, edited by Douglas T. Kenrick and Jeffry A. Simpson), extending a seminal analysis originally performed by Richard Dawkins in a famous 1979 paper about common misunderstandings of kin selection. The upshot of the Dawkins-Daly-Wilson-Salmon analysis is to answer two questions that might occur to anyone who still doubts that natural selection could so spectacularly fail to optimize genetic self-interest.

First question: Granted that our altruism genes fail to exploit the opportunity to serve themselves by encouraging diffuse altruism. And granted that it's too late for them to change. But couldn't some other gene come along and exploit this same opportunity?

No, because when another gene first "comes along," it resides in only one family. So, diffuse altruism will get it nowhere. The only kind of altruism that will get a gene up the first few rungs of the ladder toward general acceptance is the discriminating kind. To put the point in technical terms: A gene for diffuse altruism cannot successfully "invade" a population full of genes for discriminatory altruism or a population full of genes for no altruism.

Second question: Since the logic behind kin-selected genes has changed now that they pervade the species, so that they're no longer optimally self-serving, why don't they fade away? The answer is that "self-serving" is a relative term. True, these genes aren't as self-serving as genes for more diffuse altruism would be if they pervaded the species. But, again, that's not an option anyway; mutant genes don't suddenly arise in all members of a species at once. Any "replacement" gene that could make the discriminatory-altruism genes fade away would have to start out in one family and work its way outward. Now, what kind of gene would do that? Certainly not Loveless Bob's "indifference" gene. That gene would let copies of itself perish, and would die out just as it did the first time. And, as we've just seen, a gene for diffuse altruism can't pull itself up by its bootstraps, either.

In other words, our genes for familial altruism, however suboptimal, seem nonetheless to be "evolutionarily stable." Once they permeate the population, they are immune to "invasion" by rivals. At least, they are immune to invasion by rivals at both ends of the altruism spectrum--genes for indifference and genes for diffuse altruism. And, as Daly, Wilson, and Salmon note, it's hard to imagine any other plausible gene that could launch a successful invasion. At any rate, none has.

One final question: Doesn't a randomly selected member of the population have at least a slightly lower chance of sharing your altruism genes than your offspring and siblings do? Yes. Even if all the basic genes for kin altruism permeated the population (reached "fixation") 1 million or 2 million years ago, some of them might occasionally be displaced by some mutant gene. So of the (say) 50 genes directly governing kin altruism, maybe one is different from yours in a significant number of people out there. And the chances of your sibling being one of those people are less than the chances of Joe Random being such a person. So maybe you share, on average, 99.998 percent of your altruism genes with your siblings and 99.8 percent with Joe Random. But, from the standpoint of the genes' rational self-interest, such differences are too small to justify the magnitude of the discriminations they actually encourage. In terms of genetic bang per buck, your genes gain more when you spend $10 buying food for a beggar than when you spend it buying your child's 15^th plastic Christmas present.