The Price of Altruism

by Oren Harman

  Two very different examples help to show just how far kin selection has captured the imagination:

  Moving through soil by extending its pseudopods, most of the time the cellular slime mold, Dictyostelium discoideum, is a loner. Usually it engulfs and eats bacteria, but when times are rough and bacteria are scarce, something amazing happens: The starving amoebas secrete a chemical, cAMP, which attracts the others along a concentration gradient, until chains of tens of thousands of them merge into a mound. Soon the mound elongates into a slug that begins to crawl, as one multicellular body, across the forest floor. When it reaches a place with some heat and light it stops, and the amoebas that formed the front 20 percent of the body arrange themselves into a stalk, laying down tough cells of cellulose, just like plants, to make it nice and hardy. Then the remaining 80 percent climb up the stalk. When they reach the top they reorganize themselves into spores, forming a round glistening orb. It is this 80 percent that will stand a chance to live another day, sticking perhaps to the wings or legs of some insect, or otherwise being taken by the wind. The 20 percent that formed the stalk, on the other hand, will have sacrificed themselves altruistically for all the rest.7

  This is incredible, but what was discovered next is even more fascinating. In the wild most fruiting bodies form from a single clone: All the amoebas coming together to make the slug are virtually genetically identical. But when the husband-and-wife team Joan Strassman and David Queller mixed amoebas from different clones they uncovered the following: Able to recognize one another, members of one clone did their best to stick together at the backside of the slug. When the stalk was made, it was primarily they, and not the others, who shimmied up to become hopeful spores.8

  If amoeba can recognize and aid kin, so too, of course, can humans; this shouldn’t be all that surprising. What is surprising is that studies have shown that stepchildren are not only much less likely to be invested in than biological children, but also much more likely to be abused. Surprising, that is, if your names aren’t Martin Daly and Margo Wilson. This husband-and-wife team has taken kin-selection logic to its end: Just like the slime mold, they claim, and the spitting toad and the cuckoo, humans are simply following Hamilton’s rule.9

  But if genetic relatedness was a handmaiden to the gene’s-eye point of view, von Neumann games also proved a useful mountaineering partner. Soon its ropes, too, were being climbed by many a follower. The point of departure was George and Maynard Smith. Bolstered in The Selfish Gene, the concept of the ESS soon invaded the study of animal behavior. George and John, it transpired, had made an error in their paper: Retaliator, after all, was not an ESS. Since Dove did equally as well in a population of Retaliators, it could slowly drift into the population. When that happened, the true ESS would become a mixture of “Hawks” and “Bullies.”10 George, perhaps, might not have been glad to hear about it, nor to know that his “Mouse” had once again become a “Dove.”11 But considering that within a decade the application of game theory to evolution had revolutionized the field, perhaps he might have been assuaged nonetheless.

  Once more, two illustrations from the many serve to make the point. Male dung flies, it transpires, are aptly named: Like fierce elephant seals or bucking red deer, they too defend their territory, even if in their case this is nothing but a patch of smelly excrement. The reason they do so is that females lay their eggs on the dung, and the fresher (and thus smellier) the patch, the more attractive it is to them. Having arrived earlier, males fight over the best patches; he who secures the most attractive dropping will win the right to mate with the female as she deposits her eggs. The question is: For how long should a male fly defend a patch of fresh shit before moving on to another? After all, the drier and crustier it becomes, the less chance that a female will choose to land on it. Clearly, just as in a von Neumann game, the answer depends on the actions of the other male flies. It turns out that, fashioning the minute fly a strategist, an optimal ESS can be worked out. On paper it is forty-one minutes, and incredibly, in nature it’s just a few minutes away.12

  But if an ESS is good for flies, once again it is not too good for humans. In fact, game theory analyses of animal, and even plant and bacteria, behavior have been so successful that the modifications made specifically to fit evolutionary problems are now being retranslated back into economics. If neoclassical economic theory à la Milton Friedman assumed perfectly rational actors, it has since become clear that this is not really so: Risk aversion, status seeking, myopia, and other inbuilt cognitive biases are rampant in humans, and economic models of decision making need to take them into account. Introducing evolution-style games that assume minimal rationality, but whose dynamic depends on mutation, selection, and learning instead, has therefore become popular in economic theory. As an increasing number of theorists have found, this approach is helpful in figuring out problems like why firms don’t always act to maximize their profits, or whether in a given competitive market investors should be aggressive or lazy. Darwin owed a debt to Malthus, and his followers are paying it back.13

  Alongside kin selection and game theory, Trivers’s reciprocal altruism has also lowered a rope from Mount Modern-Evolutionary-Biology. One of the first to climb it, in fact, was Bill Hamilton himself. Trivers had sent him a draft of his 1971 paper, and, though Hamilton found the math flawed, he encouraged the young American to continue. It turned out that the two animal examples provided in that paper were not, in fact, good examples of reciprocal altruism: Cleaning fish not being swallowed by their larger hosts when danger came around was later repaid by hosts returning to the same cleaners, as were warning cries made by particular birds when predators were spotted lurking. These were more accurately instances of “return-effect” altruism rather than reciprocal altruism because the return benefit didn’t come from the second party’s choice to reciprocate but rather for other reasons. But despite the semantic imprecision and the weak math, it didn’t really matter; Trivers had thrown down the rope. A decade later, together with the American political scientist Richard Axelrod, Hamilton proved mathematically that, alongside perpetual defection, the strategy of tit for tat is a Nash equilibrium: Through iterated encounters natural selection would favor social behaviors that exacted a fitness cost in the short run. Reciprocal altruism had been welded to the prisoner’s dilemma. And while perpetual isolation was always an option, the rule of cooperation was simple enough for a bacterium. “The benefits of life are disproportionately available to cooperative creatures,” Axelrod and Hamilton began, and Trivers, for one, thought it of “biblical proportions.” “My heart soared,” he wrote to Bill after sitting down one night with classical music to read the paper.14

  Soon Mount Modern-Evolutionary-Biology was crowded with others climbing up the reciprocal altruism rope. The prisoner’s dilemma, these researchers found, was too simplified a version of natural interactions. But allowing for the inclusion of punishment and forgiveness, delicate cheating, observer effects (when a third party looking on has an impact on the two-person game—something called “indirect reciprocity”), and many other subtleties eventually inched the fit between nature and such models closer together. As the years progress the laws of cooperation gain steadily: Theory and observation alike place them firmly as a powerful motor in the evolution of altruistic behavior.15

  Pure direct reciprocal altruism between nonkin in nature, it must be said, has proved something of a rarity. For one thing the altruistic helpers might actually be related more often than Trivers and others suspected, rendering “reciprocal altruism” nothing but a version of Hamiltonian kin selection. Another problem seems to be that behaviors that were once interpreted as pure-cost assistance (baboons grooming each others’ backs for fleas, for example), may actually just be a form of mutualism (the baboons gain valuable nourishment from eating their friends’ fleas). Yet another impediment to the “you-scratch-my-back-I’ll-scratch-yours” theory comes courtesy of Oscar Wilde. “I can resist everything except temptation,” he qui
pped in Lady Windermere’s Fan, and most animals, experiments show, are not all that different. Immediate gratification is the custom of even the most intelligent and social of mammals, a thorn in the side of establishing the courtly conventions that serve as requisites for social restraint.

  Finally, a theory called “the handicap principle,” espoused by the Israeli zoologist Amotz Zahavi, argues that animals that perform ostensible acts of sacrifice do so to prove that they are worthy of reciprocation. Thus, when a gazelle spots a lion lurking in the grass and begins to jump up and down in the air (a behavior called “stotting”), she is advertising to her friends that she is “willing” to pay a price for being part of the group. The problem with this solution to reciprocation’s underbelly of deceit is that it is very difficult to falsify. What may seem like a selfless warning to her friends (or at the very least an act expecting reciprocation) might actually be a signal to the lion that he should focus his pursuit on a member of the troop less athletic and therefore more likely to end up on his palate. Likewise, a male peacock sporting a gigantic (and costly) colorful tail, or a bull elk showing off its large rack of antlers, may be signaling to potential mates that they need not look any further, that Numero Uno is stronger precisely because he carries a hindrance that would handicap a lesser fellow.16

  Still, alongside kin selection, reciprocity and the games it engenders have been valuable spectacles through which to gaze at Nature and her ways. There is a downside, however, for what all three scaling ropes share in common is a rather dubious conquest: Whether a monkey scratches it’s neighbor’s back, or a bee fatally loses its entrails when it stings an invader at the hive, altruism à la Hamilton, Trivers, and von Neumann is never what it seems. In fact, when it comes to altruism, the gene’s-eye view of evolution leads to positively uninspiring places. “The economy of nature is competitive from beginning to end,” the biologist Michael Ghiselin wrote,

  …the impulses that lead one animal to sacrifice himself for another turn out to have their ultimate rationale in gaining advantage over a third…. Where it is in his own interest, every organism may reasonably be expected to aid his fellows…. Yet given a full chance to act in his own interest, nothing but expediency will restrain him from brutalizing, from maiming, from murdering—his brother, his mate, his parent, or his child. Scratch an “altruist” and watch a “hypocrite” bleed.17

  The gene’s-eye view can lead to strange places, too. “Consider a pride of lions gnawing at a kill,” Dawkins asks us to imagine.

  An individual who eats less than her physiological requirement is, in effect behaving altruistically towards others who get more as a result. If these others were close kin, such restraint might be favored by kin selection. But the kind of mutation that could lead to such altruistic restraint could be ludicrously simple. A genetic propensity to bad teeth might slow down the rate at which an individual could chew at the meat. The gene for bad teeth would be, in the full sense of the technical term, a gene for altruism, and it might indeed be favored by kin selection.18

  “Tooth decay as altruism?” biologist and philosopher Helena Cronin asked, amazed, in her 1991 book The Ant and the Peacock. “That’s hardly how saintly self-sacrifice was originally envisaged.” And yet, she added, “The logic is unassailable.”19

  After years of debate, it seemed, the genes, evolution’s real scions, had finally provided the answer: Natural goodness was slowly being unmasked.20

  Until group selection started making a comeback.

  For both Williams and for Dawkins, the gene as replicator offered a strong argument against evolution working at the level of the group; if genes were the only permanent unit in nature, how could the eye of selection peruse anything else? As a bookkeeping device, scoring genes in populations helped keep good track of evolutionary change. But “selfish genes” were not fashioned merely as a metaphor: The point of view they engendered negated evolution at any other level.21

  And yet, when reconsidered by theorists beginning in the eighties, chief among them David Sloan Wilson, it began to become clear that the unit of selection and the level of selection depended on entirely different criteria. Of course genes were replicators, and clear units of permanence. But whether a certain level of life could be viewed by selection depended not on permanence but on where fitness differences resided in the biological hierarchy. Here is why: If a population is viewed as a nested hierarchy of units, with genes existing within individuals, individuals existing within groups, groups existing within populations and so on, fitness differences can exist at any or all levels of the hierarchy because heritable variation can exist at all these levels. The gene’s-eye point of view therefore says nothing about the possibility of group selection; genes, after all, can evolve by outcompeting other genes within an individual, or—via the traits they confer—by helping individuals outcompete other individuals within a group, or by helping one group outcompete another. Even if the genes are the replicators, it still needs to be determined whether they evolve by between-gene/individual selection, between-individual/group selection, or between-group/population selection. Despite all the history and hype, the gene’s-eye view and group selection are not, and never should have been, antithetical.22

  When it comes to altruism this multilevel selection approach is all the more important. For as Darwin himself perceived when contemplating the ants, altruism, like any other trait, is sure to have evolved. What makes altruism special is that it reduces individual fitness within the group while benefiting the group as a whole. The only real question, then, is whether between-group differences in fitness are ever strong enough for within-group altruism to evolve.

  This is an empirical question. And it is why, after thinking it over following their first cryptic telephone conversation, Hamilton wrote to George that he was enchanted with his formula. The covariance equation, he suddenly saw, had elegantly wiped away years of confusion. Over his own head, and the heads of Haldane, Fisher, Wright, and Allee, of Emerson and Kropotkin and Huxley, it had returned to Darwin’s own original insight: Evolution can occur at different levels simultaneously. The beauty of it was that this simple mathematical tautology could now help define where selection was acting most strongly for each and every trait.

  Hamilton thought that he was the only person in the world who understood just how momentous George’s formulation really was. For what covariance allowed to see was that while the different ropes scaling Mount Modern-Evolutionary-Biology seemed as though they had been thrown from the crest of an entirely different peak, in fact their climbers were simply making ascents up different faces of the very same mountain. Inclusive fitness made it seem as though “altruism” was always just apparent since by sacrificing himself an “altruist” might die, but his genes live on in the bodies of kin. Reciprocal altruism, on the other hand, also took the sting out of altruism, but the price in this case was always repaid to the very one who had made the sacrifice. Finally, variations on the prisoner’s dilemma fashioned reciprocation a game, often making this ascent, too, seem as though it were progressing up a distinct slope.

  The trick, of course, was to be able to see how each ascent related to the other, how scaling up one face could illuminate something about the mountain without having to exclude the lessons learned from alternative climbs. Inclusive fitness, for example, taught that genetic relatedness was important for the evolution of altruism. This was important. But what a multilevel frigate could do, armed with the ammunition of a covariance gunner, was to put it in perspective. Here are Sloan Wilson and the philosopher Elliott Sober:

  For all its insights, kin selection theory has played the role of a powerful spotlight that rivets our attention to genetic relatedness. In the center of the spotlight stand identical twins, who are expected to be totally altruistic toward each other. The light fades as genetic relatedness declines, with unrelated individuals standing in the darkness. How can a group of unrelated individuals behave as an adaptive unit when the members have no genetic in
terest in one another?

  Replacing kin selection theory with multilevel selection theory is like shutting off the spotlight and illuminating the entire stage. Genealogical relatedness is suddenly seen as only one of many factors that can influence the fundamental ingredients of natural selection—phenotypic variation, heritability, and fitness consequences. The random assortment of genes into individuals provides all the raw material that is needed to evolve individual-level adaptations; the random assortment of individuals into groups provides similar raw material for group-level adaptations. Mechanisms of nonrandom assortment exist that can allow strong altruism to evolve among nonrelatives. Nothing could be clearer from the standpoint of multilevel selection theory, and nothing could be more obscure from the standpoint of kin selection theory.23