Shadows of Forgotten Ancestors

Consider two groups, one composed of unrelentingly selfish individualists, the other of solid citizens who are occasionally willing to sacrifice themselves for (even distantly related) others. Against a common enemy, can we not imagine circumstances in which the latter group fares better than the former? Obvious disadvantages also accrue to a community of strict altruists constantly throwing their lives away in order to benefit total strangers. Such a group would not last long—if only because any tendency toward selfishness would quickly spread.

What if there’s a critical size for the group to work? When membership is below some rough threshold, certain functions of the group begin to fail. For example, the bigger the group, the better huddling together for warmth works,¹⁵ or mobbing a predator;¹⁶ and below a certain size, group benefits become increasingly unavailable. It’s not hard to imagine wholly selfish genes that cause defections from community service—a refusal to mob a predator, say, because it might be dangerous. If these genes proliferate, the point will be reached where almost nobody has the gumption to mob, and the danger posed to everyone by predators has increased. Thus, for longer-term reasons that are selfish at the level of the genetic instructions, short-term altruism may be adaptive, and might be selected for—even if the members of the group are not near relatives. In closely knit communities, individual selection and what looks very much like group selection are both elicited.

Many examples thought to demonstrate group selection have, with an almost maddening ingenuity, been explained at least equally well by a new school of biologists and game theorists. Some explanations seem quite plausible, but not all. For example, when a predator threatens a group of Thomson gazelles, one or two may leap in conspicuous high arcs near the predator. This is called stotting. The group selectionist view is straightforward: The individual calls attention to itself and risks being eaten in order to save the group. (But suppose stotting were never invented; could the predator eat more than one Thomson gazelle anyway? Compared to other species of gazelles ignorant of stotting, are fewer eaten thanks to stotting?) The prevailing individual selectionist view is that the stotter is advertising its own gymnastic abilities and reminding the predator that less athletic gazelles are easier to eat. It stots for crassly selfish reasons.¹⁷ (But then why don’t most Thomson gazelles stot when stalked? Why doesn’t such selfishness spread through the herd? Does the predator in fact turn its attention from the stotter to a less conspicuous gazelle?)

Like the classic optical illusions—is it a candelabra, or two faces in profile?—the same data can be understood from two quite different perspectives (although neither may be fully satisfying). Each may have its own validity and utility.¹⁸ Individual selection and group selection must ordinarily go together (or, in scientific speech, be highly correlated); otherwise evolution would never occur. We might argue that individual selection must have some primacy, because you can have individuals without a group, but not vice versa. However there are many animals, primates among them, where the individual cannot survive without the group.

Strict selfishness and strict altruism are, it seems to us, the maladaptive ends of a continuum; the optimum intermediate position varies with circumstance, and selection inhibits the extremes. And if it’s too difficult for the genes to figure out on their own what the optimum mix is for each novel circumstance, might it not be advantageous for them to delegate authority? For this again, brains are needed.

Consider kin selection once more. Never mind the nagging question about how well birds, say, can distinguish uncles from cousins; especially in small groups, it doesn’t much matter—everyone’s a pretty close relative, and kin selection works in a statistical sense, even if you occasionally put yourself on the line for some unrelated neighbor. It makes sense, in terms of the preservation of multiple copies of closely related genetic instructions, to accept a 40% chance of dying to save the life of a sibling (who has 50% of the same genes you have); or a 20% chance to save an uncle or a niece or a grandchild (who share 25% of your genes); or a 10% chance of dying to save the life of a first cousin (who has 12.5% of exactly the same genes that you do). Well, then, what about giving up the means of affording another child in order to preserve the families of many second cousins? What about donating ten percent of your income so a gaggle of third cousins have enough to eat? Might it pay to abstain from a few luxuries so fourth cousins can be educated? What about writing a letter of recommendation for an undistinguished fifth cousin?