jennifer_jacquet's picture
Associate Professor of Environmental Studies, NYU; Author, Is Shame Necessary?

Tit For Tat

Selfishness can sometimes seem like the best strategy. It is the rational response to the prisoner's dilemma, for instance, where each individual in a pair can either cooperate or defect, leading to four potential outcomes. No matter what the other person does, selfish behavior always yields greater return. But if both players defect, both do worse than if they had cooperated. Yet when political scientist Robert Axelrod and his colleagues ran hundreds of rounds of the prisoner's dilemma on a computer, the repetition of the game led to a different result.

Experts from a wide range of disciplines submitted 76 different game strategies for Axelrod to try—some of them very elaborate. At the end of 200 rounds, one strategy was far more successful than the others. It was also the simplest. Tit For Tat, a scheme where the player cooperates on the first move and thereafter does what was done on the previous move, was the winner. The importance of cooperation to evolution was detected by humans, but simulated and verified with machines.

This elegant explanation was then documented in living egoists with an elegant experiment. Evolutionary biologist Manfred Milinski noticed Tit For Tat behavior in his subjects, three-spined sticklebacks (Gasterosteus aculeatus). When he watched a pair of these fish approach a predator, he observed four options: they could swim side by side, one could take the lead while the other followed closely behind (or vice versa), or they could both retreat. These four scenarios satisfied the four inequalities that define the prisoner's dilemma. 

For the experiment, Milinski wanted to use pairs of sticklebacks, but they are impossible to train. So he placed in the tank a single stickleback and a set of mirrors that would act like two different types of companions. In the first treatment, a parallel mirror was used to simulate a cooperative companion that swam alongside the subject stickleback. In the second treatment, an oblique mirror system set at a 32-degree angle simulated a defecting partner—that is, as the stickleback approached the cichlid, the companion appeared to fall increasingly and uncooperatively behind. Depending on the mirror, the stickleback felt he was sharing the risk equally or increasingly going it alone.

When the sticklebacks were partnered with a defector, they preferred the safer half of the tank furthest away from the predator. But in the trials with the cooperating mirror, the sticklebacks were two times more likely to venture into the half of the tank closest to the cichlid. The sticklebacks were more adventurous if they had a sidekick. In nature, cooperative behavior would translate to more food, more space, and therefore greater individual reproductive success. Contrary to predictions that selfish behavior or retreat was optimal, Milinski's observation that sticklebacks most often approached the predator together was in line with Axelrod's conclusion that Tit For Tat was the optimal evolutionary strategy.

Milinski's evidence, published in 1987 in the journal Nature, was the first to demonstrate that cooperation based on reciprocity definitely evolved among egoists, albeit small ones. A large body of research now shows that many biological systems, especially human societies, are organized around various cooperative strategies; the scientific methods continue to become more and more sophisticated, but the original experiments and Tit For Tat strategy are beautifully simple.