Last December, more than 1,000 customers at a Starbucks drive-thru window in Newington, Conn., did something irrational. They paid for the order of the person behind them in line. The money-saving move would be to break the chain: Gratefully acknowledge that the person ahead of you in line paid for your order, then drive off with your wallet firmly in your pocket, cackling madly like the Joker.
Why are so many people obeying the unwritten rules of the pay-it-forward movement? (By the way, Pay It Forward Day this year is April 24—no joke.) The usual explanation is that fairness is a product of altruism, the tendency to help other people at a cost to ourselves.
But new research by a pair of philosophers suggests that fairness may not spring from altruism after all. “Our study shows that the connection is not so straightforward—fairness may have darker evolutionary roots,” Patrick Forber of Tufts University and Rory Smead of Northeastern University write in the latest issue of the Proceedings of the Royal Society B: Biological Sciences. I read the paper, called “The evolution of fairness through spite” (behind a firewall), after seeing it mentioned today in an article by Natalie Angier in the New York Times called “Spite Is Good. Spite Works.”
The Boston-area philosophers created computer simulations of four types of people and had the avatars interact in a well-known test of game theory known as the Ultimatum Game. In the game, there is a pot of money to be split between two people. The first person offers a share of the money to the second person that can range from 100 percent down to zero. The second person either agrees to the offered split or rejects it. If an offer is rejected, neither player gets anything. If both players were strictly rational, the person suggesting the split would offer the other person just one penny and that person would accept it, because a penny is better than nothing. But some people are so outraged by lowball offers that they reject them, hurting both sides. You can see how this game generates interesting behaviors.
Forber and Smead created four idealized types. I’ve taken the liberty of giving them nicknames. The rational Homo Economicus, as noted above, makes unfair offers and accepts any offer. The Spiteful Person makes unfair offers but rejects unfair offers by others. The Laid-Back Person makes fair offers (say, half of the pot or more) while accepting any offer, no matter how stingy. The Judge makes fair offers and rejects unfair offers. The Judge hates unfairness and kills deals—leaving herself with nothing—to punish people who make unfair offers. This is altruistic behavior.
There have been thousands of Ultimatum Game experiments. Most of them have found that the spiteful strategy doesn’t work when people are paired in the game with others who are like them. But Forber and Smead put a twist in the experiment by preferentially matching people with different types. In that simulation, Spiteful Person quickly chases away Homo Economicus and the Judge. The one person with whom Spiteful Person coexists successfully is Laid-Back Person.
What you see then is something that you might call partial fairness: Laid-Back Person makes fair offers to Spiteful People and other Laid-Back People. Alas, Spiteful Person doesn’t reciprocate. He continues to make unfair offers. Like the scorpion, it’s in his nature. But at least a little fairness is generated. To the philosophers, the most interesting part of their model is that it generates some fairness without any altruism, since the only altruistic player, the Judge, is pushed out of the game. (There is one other equilibrium consisting of Laid-Back Person and the Judge, but it’s not as stable.)
I spoke with Forber today about the authors’ rather dark vision. “It’s not fairness in the sense of justice and equity for all,” he said. “But it does generate some fairness in the population.”
I asked him if he thought this model, in which the two surviving types in the population are Spiteful Person and Laid-Back Person, accurately reflects modern society. “That’s an excellent question,” he said. “It’s a very idealized experiment. That said, these models can represent biological evolution and learning. There are general lessons. It’s a second possible route to a kind of fair play. It’s a route that many people thought didn’t exist.”