The Lesson of SuperCooperators

• Email
• Share
• Print

Life is based on competition — for the food, territory, mates, and other resources that will increase our own chances of survival and reproduction. Yet amid this competition, we often find cooperation. Genes, cells, and organisms join forces and work together. Potential rivals help each other instead of engaging in mutual antagonism. The question from an evolutionary point of view is, why would do we do things that benefit someone else at a cost to ourselves?

Martin Nowak, a professor of mathematics and biology at Harvard University and director of its Program for Evolutionary Dynamics, has long studied why cooperation exists and how it evolved. Over the years, he has used game theory and equations to express the behavior of cells, viruses, insects, and humans, trying to understand to what extent unselfish behavior can be grounded in biology. What emerged, he says, are mathematical proofs that individuals can gain by looking after one another.

In SuperCooperators, a new book written with New Scientist editor Roger Highfield, Nowak explains how computer models and experiments show that “to succeed in life, you need to work together — pursuing the snuggle for existence, if you like — just as much as you strive to win the struggle for existence.” He spoke with Big Questions Online about his work.

You write that cooperation is the most creative force in biology, whether we’re talking about cells, animals, or people. Explain.

The traditional forces of evolution are mutation and selection. Let us consider a population of reproducing individuals. If reproduction occurs with mistakes, we have mutation. If mutants reproduce at different rates, we have selection. Mutation and selection are the classic components of the evolutionary process, and they explain a lot. But if we ask: How do we get complexity? How do we get the emergence of more and more complicated structures? Then I’m arguing you need a third component, and this is cooperation.

Yet your mathematical models show that natural selection, on its own, actually opposes cooperation — non-cooperators will do better than cooperators and wipe them out. So natural selection needs help to favor cooperation — a mechanism that makes sure we get more from cooperating than being selfish and working alone. How do these mechanisms work?

A mechanism is an interaction structure. It describes how the individuals of a population interact to accumulate payoff and how they compete for reproduction. There are five mechanisms. Direct reciprocity is based on repeated encounters between the same two individuals. I help you and you help me. Indirect reciprocity is based on reputation. I help you and somebody helps me. Spatial selection means that clusters of cooperators can prevail. Neighbors help each other. Group selection occurs if there is competition between groups. The members of a group help each other. Kin selection is based on interaction between close genetic relatives. Brothers help each other, for example.

Where there is direct or indirect reciprocity, your calculations show the winning strategies must be “hopeful, generous, and forgiving.” What does this mean?

“Hopeful” means that if I meet a new person, I expect that I can establish a cooperative interaction; my first move is cooperation. “Generous” means that in a repeated interaction with another person, I do not try to win; I am content with slightly less than half of the total payoff. “Forgiving” means that we must have the ability to reestablish cooperation after occasional rounds of defection.

It can be hard to understand how cooperation can be a winning strategy in life based on your mathematical computer games. Can we use the reality show Survivor instead? The contestants are each battling to be the sole winner of a $1 million prize, but the most successful players use an alliance-based strategy, working with a group of fellow competitors against the others. By cooperating, they keep themselves alive in the game longer than the non-cooperating players, or those in weaker alliances. In other words, they earn a bigger payoff.

This scenario has elements of group selection. There’s a quote from Darwin in my book about a competition between two tribes. In one tribe, people are helpful to each other, they give aid, and they risk their lives to save each other. One can imagine that such a tribe would be victorious over another tribe. And that would be natural selection.

But alliance formation is also very important in the context of indirect reciprocity. Whoever forms better alliances has a reputational gain and others want to join such alliances.

At some point, though, the alliances always seem to break down. Someone in the group decides to defect.

Cooperation is always a risky endeavor because it is never completely stable. Cooperation always waxes and wanes. Eventually it breaks down and has to be reborn. There are always oscillations in the evolution of cooperation.

In principle, there is always a temptation to defect, if you think you can walk away with a gain. But then you destroy many possibilities.

If we can never maintain cooperation, what can we do to make it last longer and re-establish quicker after a collapse?

You need to have more possibilities for repeated encounters and you want to have situations where reputation matters — people know about each other and they know who did what to whom and why. Repetition and reputation promote cooperation.

In your opinion, incentives and rewards work better than punishment for promoting public cooperation, yes?

Some people think that costly peer punishment can play a positive role in human society. But my view of this type of punishment is less optimistic. I would say costly punishment is a behavior that people frequently use, but in most cases for selfish reasons — as a force to establish one’s dominance over others. Punishment often leads to escalation of conflict. Punishment leads to revenge and to vendettas. Punishment is that which destroys much human productivity. And in many situations, it would be better not to use punishment but to use reward. One could even go so far as to say that in many situations, not using punishment is a form of cooperation.

If you incentivize people with the threat of punishment, they will do as little as needed to avoid punishment. If you offer the right kind of reward, then they will give their very best.

Based on the idealized communities in your mathematical models, you ultimately conclude that “winners don’t punish” and “nice guys finish first.” Can these strategies lead to success in real life? Do you follow these rules?

These observations are based not only on mathematical models but also on experimental games conducted with people who interact with each other. In my own life, I certainly try to avoid using any form of punishment or escalation of conflict. I am always trying to be a “hopeful, generous, and forgiving” cooperator, but others will tell you I have yet to succeed.