The evolutionary game theory of conditional cooperation

In my mathematician’s incarnation, and in my loitering around the Institute Vienna Circle, I came across Austrian Karl Sigmund, the 2003 Gauss Lecturer. Along with John Maynard Smith (the “Etonian communist”) and American George Robert Price, he is at least one parent of evolutionary game theory, a fascinating branch of mathematics that applies game theory to biology or, rather, the evolving populations of life forms. Its tools are valuable to my interest in crowd phenomena. It defines a mathematical framework of contests, strategies and analytics for Darwinian competition. There are indeed mathematical criteria to predict the resulting prevalence of such competing strategies, and evolutionary game theory establishes a rational basis for altruistic behaviors within the Darwinian process. Unlike classical game theory, it centers on the dynamics of strategy change; its determinants that are not just competing strategies but, more importantly, the frequency of occurrence of these strategies within a given population.

Humans have superior adaptive abilities. They are far better than apes at imitation. And they are receptive to praise and reprimand. Man is the perfect pet – domesticated like no other, by ourselves. It is not unusual that a species practices selective breeding on its own kind. Sexual selection is well known since Darwin. An oft-cited example is the peacock’s tail: it does not facilitate survival but only impressing the female of the species – although recent research puts that in question. A male characteristic and the female preference for it spread across the population.

Of course, domestication does not only require selective breeding of just any given characteristic. Said characteristic must also have an economic benefit. What is the economic utility of humans? They don’t contribute commodities such as wool or eggs, but they contribute services. There are other service animals as well: horses serve as means of transportation, dogs as a hunting tool or an alarm device. What purpose do humans serve? They serve as partners of other humans. A partner is someone amenable to assistance, but only on condition of reciprocity.

Indeed, human readiness to cooperate with partners, their “conditional cooperation,” is a salient characteristic. And it is unique. While bees or ants also cooperate on a large scale, they do so only within their beehive or anthill, that is to say, with their own siblings. Humans are rather unique in that they are capable of cooperating also with individuals to whom they are not related.

Such cooperation is grounds for the success of our species. There appear to be no natural limits to the degree of our communal enterprise. That seems odd: should evolution not favor creatures that primarily maximize their own interests?

 

Of course, it can be in the best interest of an individual to participate in a common enterprise. But even greater would be the benefit of a free rider: an individual who contributes less than others and who thus exploits the collective. If this information spreads, there will soon be only free riders and the communal enterprise will fail. That happens, and in this case, maximization of the individual interest was counterproductive – a social dilemma.

The prescription formula against it is obvious: free riders must be penalized and compelled to cooperate. But such compulsion is expensive, because it creates an incentive to count on others bringing the free rider to heel. This, of course, is nothing but an advanced form of free riding. Game theory, which is the mathematical analysis of conflicts of interest, permits us to show that compulsion to cooperate works only if and when the enterprise is a voluntary one.

Voluntary submission to compulsion sounds like an odd thing, but it nonetheless forms the basis of all human societies. We sign a purchase-and-sale agreement; we enter into marriages. Even the social contract of the great moral philosophers is based on voluntariness. Jean-Jacques Rousseau commented on the voluntariness of ties: “Man is born free; and everywhere he is in chains. One thinks himself the master of others, and still remains a greater slave than they” (The Social Contract vol. I, ch. 1) (« L’homme est né libre, et partout il est dans les fers. Tel se croit le maître des autres, qui ne laisse pas d’être plus esclave qu’eux »). This explains why – and how – our social life is based on morals. Humans have a moral instinct just as they have an instinct for language. Of course, we are not born with a native language, but with a talent to learn it. Similarly, we are not born with morals, but with the talent to adopt moral norms. This includes our great disposition to recognize authority and to follow rules even when they are absurd. This requires great adaptive aptitude – thankfully, we are better at aping than any apes! Imitation supports learning. We love to drill, to teach, and to indoctrinate like no other animal species. We are receptive to praise and reprimand, we bristle at the lapses of others, and are – less readily – ashamed of our own.

We build reputation systems. A significant part of our daily flow of information consists of gossip: we judge others. Such reputation systems are key for our choice of sexual or business partner. Our economy is based on trust, and the safest method to gain a reputation of trustworthiness is to be virtuous, which is to say, to internalize social norms to such an extent that they are no longer perceived as compulsion or duty. These are virtues we instilled in ourselves. They include compassion, empathy, and a sense of fairness.

Moral norms are cultural products. But it can happen that genes adapt to cultural circumstances. For example, our genetic predisposition for the digestion of starch and milk products has changed dramatically since humans settled and built permanent residences. The same was true of dogs. Predispositions for our behavior are shaped by cultural development. We also bred virtues into dogs – love of humans is just one example. Thirty thousand years of canine-human partnership left traces in both our genomes.

We have not identified genes for virtue, but there are genes for certain dysfunctions in social behavior. Again, there is a useful parallel to language instincts: we have not identified genes for language, but there are genes for certain language deficiencies. Experiments with chimpanzees and human children show that helpfulness is congenital.

So it is entirely possible that humans domesticated themselves. In so doing, we bred the sole moral species on the planet. It is another matter that we did not shed aggressiveness in the process. But that has not worked entirely in the case of domesticating canines, either.