Do economic games tell us something about real behaviours?

The last issue of Current Anthropology reports a research conducted by Polly Wiessner among the Ju/'hoan Bushmen of the Kalahari. Like a recent article by Gurven (see our previous post), this research calls into question the relevance of economic games. Here is the abstract:

Experimental games – the dictator game and the ultimatum game – were played out among the Ju/'hoan Bushmen of the Kalahari. Subsequently, the experimenter tracked what the players did with the money earned in the games to see how it was used in "games of everyday life." Players were stingy and did not punish in experimental games and were generous and did punish in "games of life." The fact that the conditions of anonymity of the games removed cultural institutions and emotions governing sharing and reciprocity lead Ju/'hoansi to reassess risks and benefits and play more selfishly. The findings underline the importance of cultural institutions such as sharing, reciprocity, and social sanctions (costly punishment) to provide the structure for other-regarding behavior to be expressed and to be rendered beneficial for the participants.

Note that a less "cultural" view is possible. Indeed, Wiessner emphasizes the role of culture and institutions (such as Xaro, a formal gift exchange partnerships). However, from an evolutionary and cognitive point of view, one could note that people lack the most basic cues to behave morally: They interact with an unknown person, they don't know whether they are going to interact with her again, if she is trustworthy, they do not know anything about her situation (is she hungry? ill? wealthy?), etc. In other words, from an evolutionary and cognitive point of view, the participants' moral sense may think that it is not at all a situation when one has a duty to share.




1 Comment

  • comment-avatar
    Nicolas Baumard 18 February 2009 (12:30)

    I agree with Pascal that people have hard time understanding the instructions. However, I would be less severe with Wiessner. Indeed, people do understand the instructions to some extent since they give less in anonymous contexts than non anonymous contexts (which suggests that they have understood that their reputation can be preserved in the anonymous condition). My point was actually a bit different: even if you trust the experimenter that the game is about real money, even if you think that your reputation is not at stake and so on, you still have the problem that moral duties rarely occur in the abstract. Indeed, I have more duties (for instance to lend them some money) toward my close friends than toward some remote acquaintance. By removing the identity of the other player, you remove a lot of relevant aspects of the situation. Consequently, people seem to be less moral than they are in reality. It may not have someting to do with culture and all with knowing the social identity of the other player. Note that this could explain the following result obtained by Wiessner: [i]The low offers of the Ju/’hoansi and low rates of punishment are matched only by the Hadza, Machiguenga, and Quicha in the study of 15 small-scale societies (Henrich et al. 2004a; Henrich and Smith 2004b; Patton 2004; Marlowe 2004).[/i] It could be the case that the smaller the society is, the smaller the duties are toward strangers. Indeed, in our modern societies where we are used to meet strangers everyday, to interact by phone and e-mail with strangers, and to make new friend very easily, we have good reasons to think that we have some basic duties toward strangers. In other words, the moral community is bigger. On the contrary, in a society in which a stranger is often an enemy, it is not clear that one have duty to help him. We are led to the conclusion that the abstract settings of economic games, albeit non ecological, are more ecological for westerners than for the inhabitants of small scale societies. Contra Wiessner, this difference does not come from different cultural institutions, but from different expectations about everyday interactions.