Playing Fair Came Late

Decisions, decisions. An Au man of Papua New Guinea decides how much of his stake to allocate to an anonymous second party in the Ultimatum Game.

David P. Tracer

Humans began living in large communities only about 10,000 years ago. Yet most of us tend to treat those around us fairly, even if they're not close friends and relatives. Is this sense of fairness a holdover from the days when we lived in close-knit hunter-gatherer groups, or did it evolve as society evolved?

A team led by anthropologist Joseph Henrich of the University of British Columbia, Vancouver, tested this with members of groups including nomadic African herders, Colombian fishermen, and Missouri wage workers. The idea was to cover communities as varied as hunter-gatherer groups comprising no more than several dozen people and modern communities of several thousand members. Societies were assessed by three measures: size, degree of "market integration" (food purchased rather than caught or grown), and membership in a world religion—a marker for having a moral code that extends beyond kin to people in general.

The subjects were given three tests that revealed their willingness to punish selfishness—even at a cost to the punisher. In one, the Dictator Game, volunteers simply decided how to split an amount of money—equivalent to a local day's wages—with a member of the community who was kept anonymous by being kept out of sight of the subject. In the second, the Ultimatum Game, a subject split money with an anonymous person. The stranger could refuse the offer as unfair, in which case neither party got any money. In the final test, the Third Party Punishment Game, the subject made an offer to an anonymous person, and a third party decided whether the offer was fair. If he decided it wasn't, both he and the initial offerer lost money. In both the second and third games, punishers pay a price because they get more money if they abide by an unfair decision.

Members of large, complex communities had both a keener sense of fairness and a greater willingness to punish unfairness, the team reports in tomorrow's issue of Science. In fact, on the three tests, the money offered by subjects from the largest societies ranged from 25% to 51% higher than the offers by subjects from the smallest subsistence-based groups. With the latter groups "you get results in those experiments that look nothing like what you get with university students," the subjects commonly used by researchers in such experiments, Henrich says. "They're unwilling to punish, and keep much more money for themselves."

The findings argue against the conventional evolutionary psychology hypothesis that ideas of fairness and punishment are just evolutionary holdovers, says anthropologist of Robert Boyd of the University of California, Los Angeles. "If this were just the Pleistocene psychology working itself out, you would not expect to see these systematic differences between societies." In other words, fairness to strangers—who would be regarded with antagonism or at least mistrust in small, close-knit hunter-gatherer groups—is a behavior that has evolved along with other norms in complex societies.

Other experts add a note of caution, however: Administering tests anonymously and with no cultural context to members of small tribal societies yields "uninterpretable" results, say Harvard University evolutionary game theorists Martin Nowak and David Rand. The games are an "artificial situation," says Rand. College students are "used to [such] concepts and hunter-gatherers aren't. Who knows how they're understanding the game?"