Commenting on this:

Very interesting read. To me experimental economics, economic anthropology and also some other sub-sub-disciplines are really converging in terms questions and methods. Converging on something that used to be social psychology. Take contemporary experimental economics: they ask classical psychological questions, they do experiments as social psychologist do, the only difference is that the incentivise everything. From that perspective John Henrich’s psychology research is very much the same as, say, the research of Ben d’Exelle who is an economist. To me both of these guys work in the same field which is artificially split across different departments.

As to the psychological rejoinders of game-theoretical models: I do not take them too seriously. Game theory models are in great majority analytical models not predictive models: they are created to understand some simple, basic mechanisms of interdependence not really to predict (predicting as in weather forcasts). We know that they “fail descriptively” (as Kahnemann have said here ). I like to think about game theory models in social sciences like Newton laws in contemporary physics: everybody knows that they are essentially incorrect (as Einstein showed), nevertheless they serve as a good approximation and a starting point for analyzing the real thing (quoting Gintis from here).

Classical game theory is extended in various directions to give better predictions. I see two essential directions. The first one is “psychological” in the sense that the theory is extended by looking “inside” the decision maker: c.f. prospect theory, alternative utility models etc. The second one is more “sociological”: games are not played in isolation, the presence of third-parties matter. Put it differently: the games are embedded. Embedded not only within social structures, as Fabio wrote referring to Raub&Buskens, but also temporarily, i.e. played repetitively. Consequently things like norms, expectations, reputation etc. matter too.