The last issue of the Journal of Institutional Economics features an interesting and stimulating set of articles on how to account of institutions in game theory (note that all articles are currently ungated). In the main article “Institutions, rules, and equilibria: a unified theory”, the philosophers F. Hindriks and F. Guala attempt to unify three different accounts of institutions: the equilibrium account, the rule account and Searle’s constitutive rule account. They argue that the solution concept of correlated equilibrium is the key concept to make such a unification. The later retains the notion that institutions can only persist if they correspond to an equilibrium, but at the same time it emphasizes that institutions can be understood as correlating devices based on the humans’ ability to symbolically represent rules (as a sidenote, I make a similar point in this forthcoming paper as well as in this working paper [a significantly different version of the latter is currently under submission]). The authors also argue that Searle’s constitutive rules are reducible to regulative rules (I have presented the argument here).
Several short articles by Vernon Smith, Robert Sugden, Ken Binmore, Masahiko Aoki, John Searle and Geoffrey Hodgson reflect on Hindriks and Guala’s paper. They are all interesting but I would essentially recommend Sugden’s paper because it tackles a key issue in the philosophy of science (i.e. whether or not scientific concept should reflect “common-sense ontology”) and Searle’s response. I find the latter essentially misguided (it is not clear whether Searle’s understand game-theoretic concepts and it makes the surprising claim that “if… then” (regulative) rules have no deontic component) but it still makes some interesting points regarding the fact that some institutions such as the promise-keeping one exist (and create obligations) though they are not always followed.