Do chimpanzees care about fairness? The jury’s out

A stranger gets a pot of money and offers you a share of it. If you accept the offer, both of you walk away with your proposed shares. If you reject, you both leave with nothing.

This is the ultimatum game—a classic psychological experiment used to study fair play. If both players behave completely selfishly, the proposer might offer as little as possible, while the responder should accept any offer as long as it’s not zero. That way, both of them walk away with something. In practice, responders typically reject any offers less than 20 percent. They care enough about fairness to do themselves out of money in order to punish unfair partners. Meanwhile, proposers from industrialised countries, who are wary of social norms and the potential for punishment, tend to offer between 40 and 50 percent of the pot.

But what about chimpanzees? Do our closest relatives also share our attitudes to equality? Darby Proctor from Georgia State University thinks so. She modified the ultimatum game so that chimps and human children could play it in the same way. In both cases, when proposers needed to cooperate with responders, they became more likely to offer an equal split, rather than trying to hog the rewards for themselves. The conclusion: “humans and chimpanzees show similar preferences regarding reward division, suggesting a long evolutionary history to the human sense of fairness.”

But her study has prompted stern criticism from scientists who have also tested chimps at the ultimatum game and found the opposite. In their studies, our fellow apes did not make fair offers and accepted anything as long as it was greater than zero.

This debate reflects a growing divide between scientists who study chimpanzees, but it is more than an academic spat. It speaks to a fundamental question about our evolution: is our sense of fairness a uniquely human trait, or one that we share with our closest kin?

The experiments

Keith Jensen, Josep Call and Michael Tomasello were the first to play the ultimatum game with chimps. In their experiments, two chimps sat in adjacent cages, facing a contraption with two sliding trays. Each tray contained one dish on the proposer’s side and another on the responder’s side, and the two dishes carried varying numbers of raisins. The proposer made an offer by using a rope to pull one of the two trays half-way over. The responder could accept this offer by pulling the tray the rest of the way, allowing them both to eat their respective raisins. Alternatively, they could reject the offer by doing nothing, leaving both animals unfed.

The team found that the proposers were likely to choose the tray that gave them the most raisins, and the responders tended to accept any offer, no matter how unbalanced. They wrote that “in this context, one of humans’ closest living relatives… does not share the human sensitivity to fairness.” And in a second study, the team found that our other close relatives—the bonobos—behaved in a similar way.

But Proctor was unimpressed by the team’s set-up. She felt that the “complex mechanical apparatus” was unlike anything that humans use when we play ultimatum games, and may have been too complicated for the chimps to understand. And while humans usually play for money, which is exchanged for other rewards, the chimps were playing for food, which is immediately rewarding.

In her version of the ultimatum game, two chimps play for tokens that are exchanged for six bananas laid out in front of them. A human experimenter offers two tokens to the proposer chimp, one signifying an equal banana split and the other signifying an unequal 5:1 divide. The proposer picks one token and passes it to a responder in an adjoining cage. They can drop the token to reject the offer, or pass it back to the experimenter to accept.

Proctor played the game with three pairs of chimps, one of which swapped roles as proposer and responder. She found that two of the proposers chose the equal-split token more often than expected by chance. And all four of them chose that token more often in the ultimatum game than in a straight preference test, when their choice dictated their reward irrespective of what a responder did. Proctor concluded that when chimps need to cooperate to get a reward—that is, when the proposer depends on the responder—they change their behaviour to favour the fairer option. Young children, aged two to seven, behaved in the same way.

So what does that mean?

The obvious criticism is that Proctor’s study only included six chimps, and only the two animals who played both roles offered the equal token more often than expected by chance. Proctor admits that the numbers were small, but says that these were the only chimps that passed her rigorous pre-tests and clearly understood the nature of the game.

But Jensen disputes Proctor’s claim. He is glad that another team tried to replicate his results, since “one can only conclude so much from one or two studies,” but says that Proctor’s experiment was no ultimatum game. “The most important aspect of the ultimatum game is not what the proposer does, but how the responder reacts,” he says. The proposer’s offers are strategic rather than a sign of fairness—they’re a reaction to what the responder might do. It’s the responder’s ability to reject unequal offers that drives fairness in the game.

And among Proctor’s chimps, no responder ever refused an offer, even the many unfair ones. That’s even less rejection than in Jensen’s study. “Not rejecting unfair offers is puzzling if chimps are really playing the ultimatum game,” says Call. “I see that as a fatal flaw,” adds Jensen. At best, it confirms his original experiment by showing that the responders are insensitive to unfairness and only motivated by getting bananas. At worst, it shows that they didn’t understand the task.

Proctor counters that she did extensive tests to be as sure as possible that the chimps understood what the tokens meant. She admits that she did not explicitly train the chimps that they could refuse offers, but says, “This actually makes our results more striking. Without experiencing a refusal, proposers changed their behaviour to be more equitable. They may be responding to the potential for refusals as do adult humans.”

Jensen doesn’t buy it. “There isn’t the tiniest shred of evidence that proposers understood that responder could reject their offers, and no demonstration that responders understood anything of the possible consequences of their choices,” he says.

David Rand, a psychologist from Harvard University who has used the ultimatum game in human studies, agrees with Jensen’s criticisms. While Proctor’s set-up does look like an ultimatum game, “it looks like maybe the chimps didn’t understand the game structure,” he says.

Jensen thinks that this confusion arose because Proctor’s task is not as simple as she claims. Unlike human ultimatum games, where players interact with each other, Proctor’s chimps spent as much time exchanging tokens with humans. “Passing a token is just an intermediate step to getting food from experimenters, something they are highly trained to do,” says Jensen. He doubts that this set-up, which involved tokens exchanging hands three times, is truly simpler than his tray-pulling machine.

The verdict

As Proctor notes, there are many reasons to suspect that chimps care about equality. They help one another, share food, and cooperate extensively to hunt, fight, patrol, defend, and more. But it’s difficult to interpret wild anecdotal behaviour, which is why experiments are valuable.

None of the existing studies is perfect. In all the chimp ultimatum games, the animals could only reject offers passively, by not pulling a tray or not handing over a token; in human games, rejection is an active choice. In the chimp games, the animals could see each other, and played multiple rounds with the same partners; in human games, partners usually play single rounds anonymously to stop social dynamics and reputations from clouding the results.

Given these shared weaknesses, Proctor’s team is right that Jensen’s studies don’t prove that chimps are insensitive to fairness even though they support that hypothesis. After all, absence of evidence is not evidence for absence. But equally, the problems in Proctor’s study prevent it from confirming that chimps are sensitive to fairness. Until more research is done, we’re at an impasse.

Reference:  Proctor, Williamson, de Waal & Brosnan. 2013. Chimpanzees play the ultimatum game. PNAS http://dx.doi.org/10.1073/pnas.1220806110

Note: This study was “contributed” to PNAS by co-author Frans de Waal, a publishing route where members of the National Academy of Sciences can nominate their own peer-reviewers. I try to avoid papers that use this track but did most of the reporting before I noticed, so here’s the piece anyway.