How can cooperation emerge in a world of selfish individuals ruled by a Darwinian competition for survival?

This is the question that Martin Nowak, Professor of Mathematics and Biology at Harvard University, discusses in Super Cooperators: Altruism, Evolution, and Why We Need Each Other to Succeed.

Nowak and his collaborators have published a series of articles in major scientific journals that give partial answers to this question. The book provides a gentle overview of the technical results, with frequent comments about the implications for politics and economics. For example, Nowak repeatedly mentions climate change as an example of something requiring cooperation among humans.

The hope is that if we understand, mathematically, how cooperation emerges, we can better design policies and structures to promote cooperation and deter selfishness.

I propose that “natural cooperation” be included as a fundamental principle to bolster those laid down by Darwin. Cooperation can draw living matter upwards to higher levels of organization… Cooperation makes evolution constructive and open-ended.

Super Cooperators: Altruism, Evolution, and Why We Need Each Other to Succeed

The book has a few simple mathematical formulas, but the educated layman should be able to understand the gist of the arguments, thanks to generous use of example, analogy and simplification. Indeed, the book’s readability benefits from the aid of Roger Highfield, an author of popular science books, who helped Nowak with the writing.

Darwinian evolution is based on competition for survival, for resources, and for mates. Winners reproduce, losers leave few offspring. Due to mutations, individuals vary in their fitness. Over many generations, fitter (configurations of) genes proliferate, while weaker ones disappear.

In fact, fitness is defined in terms of ability to reproduce, so the fact that fitter individuals reproduce is something of a tautology.

Similarly to evolution, in an economy, people often act selfishly, trying to get paid as much as possible for what they sell, whether goods or their services, and trying to pay as little as possible for what they buy.

It would appear that cooperation is difficult to explain in a pure, evolution-based model or in a selfish profit-based economy. You’d expect that selfishness would always win out. But it’s clear that cooperation is common, both among non-human animals and among humans.

The basic reason is that, in the long run being nice pays off, for you or for your children, kin, or neighbors.

In the context of this book, cooperation basically means: an individual is willing to sacrifice some short-term benefit in exchange for a longer-term reward, either for itself or for related individuals (e.g., children or kin or members of the same group). In other words, cooperation is a form of reciprocity, or reciprocal altruism. This sense of cooperation isn’t as pristine or as self-sacrificing as some religious traditions’ ideals of pure selfless love. But even Christianity relies on a promise of reward and punishment in the afterlife to motivate moral behavior.

Albert Einstein once said, “If people are good only because they fear punishment, and hope for reward, then we are a sorry lot indeed.”

Unfortunately, from the point of view of biology, all we seem to have is punishment and reward, where reward means reproductive fitness: produce descendants who survive and who likewise reproduce. (It is not sufficient to have children: if your kids are too weak to survive, or if they don’t reproduce, your reproductive fitness isn’t really high.)

Yet the bearer of fitness (the entity getting the reward or punishment and that gets to reproduce) isn’t necessarily the individual of a species. Richard Dawkins famously suggests that the unit of competition and survival may be the gene: animals exist to promote the interests of their genes, not the other way around. Moreover, genes, as well as gene networks, span individuals and species.

There are also theories which say the unit of the reward is the group: related kin, or cooperating subgroups, or (at a higher level) cooperating species who live in symbiosis with one another.

Indeed, group-based reciprocity seems to be the essence of cooperation.

We are all in it together.

We are interdependent.

Nowak thinks cooperation, and not just competition, is a fundamental force in evolution.

I have argued that evolution “needs” cooperation if she is to construct new levels of organization, driving genes to collaborate in chromosomes, chromosomes to collaborate in genomes, genomes to collaborate in cells, cells to collaborate in more complex cells, complex cells to collaborate in bodies, and bodies to collaborate in societies.

A set of genes working together is an example of cooperation. And in the primordial soup, sets of cooperating chemical reactions led to the origins of life.

Within biology, there have been attempts to explain cooperation in terms of kin selection (in which an individual is willing to sacrifice itself to aid close relatives who share many genes with it). The social insects are prime examples of cooperators; the worker ants who build and defend the nest are closely related to the queen.

A related notion is group selection (aka multi-level selection), according to which groups which are more fit (e.g., due to being better cooperators) out-compete groups which are less fit.

The idea of group selection seems intuitively correct, and Darwin was aware of the role of cooperation in evolution and of the apparent presence of group selection, both in biology and in culture (where ideas or what are now called “memes” reproduce).

But there are heated disagreements among professional biologists about whether the phenomenon of group selection really occurs and about the extent to which it occurs. Richard Dawkins has famously ridiculed both the idea and the biologists who support it. Nowak seems to be among the latter group.

Examples of cooperation among humans include: lending a cup of sugar to a neighbor, taking the bus instead of driving the car, paying taxes instead of cheating, contributing to the donation plate, bringing in your neighbors’ and garbage bins from the curb, as well as more dramatic examples such as risking your life to safe someone who has fallen onto train tracks. Most parents would instinctively risk their lives to save the lives of their (small) children.

In the mathematical and computer models of cooperation, various individuals interact with other individuals, either in a well-mixed pool; in a network of connections such as on social networks; in various sets of interests groups; or on a grid. Whenever you interact with another individual, each of you decides whether to cooperate or whether you will defect (be selfish). You are rewarded or punished accordingly.

Mathematically, cooperation is formalized in the form of such a two person game. The standard game of this sort is is called The Prisoner’s Dilemma.  It models the situation where two prisoners who have been arrested by the police and are being interrogated separately. Each prisoner gets to choose, independently, whether to cooperate (keep his mouth shut and deny the crime) or defect (accuse his partner of the crime). If they both cooperate they each get only one year in prison on a lesser charge, because the police have insufficient evidence. If they both defect, they each get two years in prison. If one person cooperates with his partner and the other person defects, then the first person (the cooperator) gets three years in prison and the second person (the defector) gets off free.

From the point of view of each prisoner, it seems the smartest thing to do is defect.

Suppose the other person cooperates and stays mum. Then you should defect, because you get off free.

On the other hand, suppose the other person defects and accuses you of the crime, then you better defect too. For if you cooperate with your partner, you get three years in prison, whereas if you defect you get just two years in prison.

What could prevent defection is loyalty, or the knowledge that in the future, after you’re both out of prison, the other person could punish you. Likewise, in a future similar situation, where cooperation might help he will remember your betrayal.

The tragedy of the commons is a similar scenario.

In the more general game, where rewards and punishments can take the form of money or some other outcome, there are likewise four possible outcomes: Cooperate-Cooperate, Cooperate-Defect, Defect-Cooperate, and Defect-Defect. Each outcome has a (possibly different) payoff for each of you. If you both cooperate, you both get the same reward R for cooperating. If one person cooperates but the other person defects, the first person is punished (S for Sucker) bad but the other person wins a big reward (T for Temptation) If you both defect, you’re both punished slightly (P). Depending on the relative values of P,R, S, and T, and on the structure of interactions — specifically, whether you can learn about the reputation of the person you’re interacting with — cooperation may or may not emerge.The standard Prisoner’s Dilemma game has

T > R > P > S.

Yet cooperation can emerge. This result is non-intuitive, because given the inequalities above, the values P, R, S, and T guarantee that in the short-term the smartest thing to do is to defect. Here’s why. Your opponent is either going to cooperate or defect (and you won’t know which he does til after you make your move).

Assume he cooperates. Then you can win big by defecting. Here’s why. If you cooperate, you get only R. But if you defect, you get T and T>R. So, it seems you should defect.

Likewise assume he defects. Then you better defect too, because if you cooperate, then you’ll get only S, but if you defect you’ll get P, and P>S.

So in either case, the best thing to do, in the short run, is to defect.

But in a community of people playing the game repeatedly, there are benefits from cooperation. A group of cooperating individuals will have a higher fitness (reward) than a group of turncoat defectors, because R>P.

If the last time I interacted with you, you cooperated, and if I remember that, I can try cooperating again, in the hopes that you will reciprocate.

So in the presence of repeated interactions, and memory, cooperation can emerge.

Cooperators are rewarded with help from other cooperators. Defectors are punished by future defection. If cooperators gain a benefit as a group that is unavailable to defectors, then cooperation can flourish. But cooperation is always susceptible to exploitation by defectors: a population of trusting cooperators can be taken advantage of by a few defectors.  Such invasions by defectors are visible in computer simulations.

Cancerous cells can be modeled as defectors.  So can tax dodgers and alleged welfare moms who drive Mercedes.

Using the formalization of Prisoner’s Dilemma, Nowak was able to prove mathematical theorems, and run computer simulations, that show under what conditions cooperation can flourish.

He showed that cooperation emerges if you meet the other person often enough in the future and can remember the previous interactions, so you can punish or reward him. It also helps if people have a reputation that is is public knowledge or that is shared between individuals (indirect reciprocity). Furthermore, it helps if people are organized into small groups; this allows cooperators to shield themselves from being taken of advantage of by nasty defectors; large groups are difficult to police. Finally, it helps if it’s possible to move between groups, to escape defectors.

Even if we can explain cooperation biologically, in terms of kin selection, or group selection, there is still a problem: how inclusive is the in-group?  Does it include people of a different race or nationality? How about individuals of a different species?

As indicated above, the biologically inspired notion of cooperation is somewhat unsatisfying, because it still relies on a form of reciprocity, albeit at the group level. If someone chooses not to identify with the group, then why should they cooperate?

Indeed, conservatives are the consummate defectors: individualists who detest and ridicule cooperation and community endeavors, at least by governments. Conservatives detest the United Nations. They detest the International Court of Law.  Conservatives avoid paying taxes,  but typically like spending money on wars, both domestic (e.g., the wasteful and disastrous war on drugs) and foreign.  Conservatives oppose laws and regulatory agencies that deter their antisocial behavior. They under-fund the IRS, encouraging tax cheats. And they under-fund Congressional staff, so that lawmakers are dependent on lobbyists and outside groups for information.

Like parasites, conservatives destroy the body politic, all in the name of “freedom.”

Conservatives and their ideology can be defeated only when enough people wake up to the lies and the half-truths behind their movement,  and when enough people realize that we’d be better off in the long run by cooperating on building a government that works for everyone, that makes sound environmental, health, and safety policies, and that makes sure tax cheats pay their fair share.