Tuesday, November 04, 2014

Chimps Play Fair in the Ultimatum Game

The model of Homo economicus--that human beings are rational maximizers of their selfish interests--has been challenged by ultimatum bargaining experiments.  In the ultimatum game, there is a pool of money (say $100) to be split between two people.  The person designated as the proposer must offer a split to the other person designated as the responder.  If the responder accepts the offer, the money is split as proposed, but if the responder rejects the offer, then neither person receives any of the money.  Homo economicus would propose to take most of the money for himself (say $95) and give little to the other person ($5), and the responder would accept this unfair offer, because even a small share of the money is better than nothing.  In most experiments, this does not happen.  Proposers tend to offer 40% to 50% of the money to the responders, and the responders tend to reject any offer to themselves of less than 20%.  This seems to show a sense of fairness.  Responders are willing to suffer a cost to themselves to express their resentment against unfair offers by punishing the unfair proposers, and proposers are inclined to make fair offers, perhaps anticipating that they will be punished if they make an unfair offer.

There is some cultural variation in the results of these ultimatum game experiments.  In a few tribal societies, proposers tend to make unfair offers that are generally accepted by responders.  One of the factors explaining this is that these societies have little experience with market exchanges, and thus they have not developed the norms of fair exchange that tend to arise in societies where trading in markets is an important part of their economy. 

This seems to provide experimental confirmation for Adam Smith's claim about the influence of commerce on the manners of a people.  "Whenever commerce is introduced into any country, probity and punctuality always accompany it.  These virtues in a rude and barbarous country are almost unknown. . . . A dealer is afraid of losing his character, and is scrupulous in observing every engagement.  When a person makes perhaps 20 contracts in a day, he cannot gain so much by endeavouring to impose on his neighbors, as the very appearance of a cheat would make him lose.  Where people seldom deal with one another, we find that they are somewhat disposed to cheat, because they can gain more by a smart trick that they can lose by the injury which it does their character" (Lectures on Jurisprudence, 538-39).  Commerce fosters a concern for one's reputation for being trustworthy and fair.  Moreover, Smith argues in The Theory of Moral Sentiment, all of our morality can be explained as rooted ultimately in our concern for how we appear to others, for winning their approbation and avoiding their disapprobation.  This concern for our public appearance can be internalized as a conscience in which we imagine how we would appear to a perfectly informed and impartial spectator.  Thus we care not only about being praised but also about being praiseworthy.

If human beings show a sense of fairness in how they play the ultimatum game, then we might wonder whether other animals closely related to us in their evolutionary ancestry show a similar sense of fairness, which would suggest that human morality has deep evolutionary origins.  This question has provoked a debate among primatologists and evolutionary psychologists.  On the one side, Michael Tomasello and Josep Call have concluded that the expression of a sense of fairness in the ultimatum game is unique to human beings, and that chimpanzees show no concern for fairness in their play of the ultimatum game, because they are actually rational maximizers of their selfish interests.  On the other side, Frans de Waal and Sarah Brosnan have concluded that chimpanzees really do show some sense of fairness in their play of the ultimatum game, and this suggests that human morality has evolutionary roots in a primate ancestry shared with chimpanzees.

Until recently, the testing of chimpanzees in simulated ultimatum games failed to show any concern for fairness.  But in recent years, de Waal, Brosnan, and their colleagues have shown that chimpanzees do tend to make fair offers in playing the ultimatum game.  In this short video, you can see how they designed their experiment.  One individual chose between two tokens (functioning as money) that could be exchanged for food.  One token could be exchanged for an equal division of the food with another individual.  The other token would favor the proposer with a larger portion of the food.  Once the proposer had selected a token, the responding individual could either transmit the token to the experimenter, who would divide the food as indicated, or the responder could refuse to transmit the token, and neither individual would receive any food.  In 75% of the trials, the proposer choose the equitable token (a 50/50 split) over the inequitable token.  Although the responder never rejected any offer, the responder did often express irritation when the inequitable token was offered--sometimes by spitting water at the proposer.  Experiments with human children (2-7 years old) showed the same pattern of behavior.

De Waal and Brosnan have also shown "inequity aversion" in experiments with monkeys and chimpanzees.  Two individuals are put near to one another in cages.  They perform some simple tasks to earn tokens that they can exchange for food.  At first, they are given slices of cucumber, and they are happy with that.  But then one of them receives a grape, which is much more desirable than a cucumber.  Now, when the other individual receives a slice of cucumber, he protests by throwing the food away and pounding on his cage.  At the end of his TED talk on "Moral Behavior in Animals," de Waal shows a video of this experiment, and he suggests that the angry monkey's behavior is similar to that of the "occupy Wall Street" protesters.

As de Waal indicates, this research provoked indignant responses from philosophers who insisted that monkeys and apes could not show any sense of fairness, because this must be unique to human beings.  Some of them argued that to show a true sense of fairness, it was not enough for the unfairly treated animal to protest, but the animal benefiting from the unfairness should also protest against this unfair outcome.  Remarkably, there is now some evidence that chimpanzees do show this: the chimpanzees will react negatively not only when they receive the less desirable food but also when they receive the more desirable food (see Brosnan et al. 2010; de Waal 2014)!

And yet, when one actually looks at the data in these experiments, the conclusions that de Waal and Brosnan draw from them might be disputed.  For example, one set of experiments involved 16 adult chimpanzees, 10 males and 6 females (Brosnan et al. 2010).  The chimps were trained to exchange an inedible token that they had been given for a food reward. They could refuse to do this by refusing to accept the token, by refusing to accept the food, by not returning the token, by not eating the food, by sharing the token with a partner, by sharing the food with a partner, by pushing away the token, or by pushing away the food. 

In the inequity test, where the subject received a low value food (carrot), and the partner received a high value food (grape), the percentage of total refusals (either refusing to return the token or refusing to accept the food) was 30% for the males and 8% for the females.  6% refused by either pushing away the token or pushing away the food.  So the great majority of individuals did not refuse the inequitable distribution.  In his public  lectures like his TED talk, de Waal shows a brief film clip of a chimpanzee throwing away the low value food (carrot or slice of cucumber), when the other chimpanzee has received the high value food (grape), and the audience laughs.  But de Waal does not show the film of the many times that chimpanzees in the inequity test accept the inequitable distribution.  So the audience does not see that the "inequity aversion" that de Waal claims is shown only in the minority of cases.  

Brosnan and de Waal have suggested that this sense of fairness did not evolve for the sake of fairness as such but for the sake of beneficial cooperation.  Both chimpanzees and human beings are highly cooperative animals who need the cooperation of individuals who are not their kin.  Evolution will favor the propensity to cooperate only as long as the cooperating individuals receive roughly equal benefits from the cooperation.  Individuals who feel that they are not receiving equal benefits will withdraw from the cooperative activity, and thus those receiving more than their fair share will be punished.

Human beings show this same evolutionary psychology of fairness.  But they are unique in that their cognitive abilities and their capacity for language allow them to develop and enforce rules of fairness that are far more abstract and complex than is possible for chimpanzees.


REFERENCES

Brosnan, Sarah, et al., "Mechanisms Underlying Responses to Inequitable Outcomes in Chimpanzees, Pan troglodytes," Animal Behaviour 79 (2010): 1220-1237.

Brosnan, Sarah, and Frans de Waal, "Evolution of Responses to (Un)fairness," Science 346 (17 October 2014): DOI: 10.1126/science.1251776.

de Waal, Frans, "Natural Normativity: The "Is" and "Ought" of Animal Behavior," Behaviour 151 (2014): 185-204.

Jensen, Keith, Josep Call, and Michael Tomasello, "Chimpanzees Are Rational Maximizers in an Ultimatum Game," Science 318 (5 October 2007): 107-109.

Milinski, Manfred, "Chimps Play Fair in the Ultimatum Game," Proceedings of the National Academy of Science 110 (5 February 2013): 1978-1979.

Proctor, Darby, Rebecca Williamson, Frans de Waal, and Sarah Brosnan, "Chimpanzees Play the Ultimatum Game," Proceedings of the National Academy of Science 110 (5 February 2013): 2070-2075).  Available online.

Other posts on this debate can be found here, here, here., here, and here.

No comments: