Regular Article
A Note on the Implications of Common Knowledge of Rationality

https://doi.org/10.1006/game.1994.1006Get rights and content

Abstract

I show that common knowledge of rationality is not, in general, equivalent to the limit as n → ∞ of order n mutual knowledge of rationality. In particular, one can construct games in which an action is consistent with order n mutual knowledge of rationality for every finite n but is not consistent with common knowledge of rationality. Hence finite order mutual knowledge may be very poor as an approximation of common knowledge. Journal of Economic Literature Classification Numbers: 021, 026

References (0)

Cited by (35)

  • Iterated bounded dominance

    2023, Economics Letters
  • Learning to agree over large state spaces

    2022, Journal of Mathematical Economics
    Citation Excerpt :

    An analogous question has been studied in game-theoretical contexts. Specifically, Lipman (1994) shows that in some games it is possible to attain common knowledge of rationality only through a transfinite process of elimination of never-best replies. Finally, we do not claim that transfinite dialogues are necessary in every possible situation covered by our model.

  • Common belief in rationality in psychological games: Belief-dependent utility and the limits of strategic reasoning

    2022, Journal of Mathematical Economics
    Citation Excerpt :

    Theorem 4.3, Example 4.1, and Theorem 5.5 show a striking similarity between strategic reasoning in psychological games and strategic reasoning in traditional games with infinite choice and/or player sets: Firstly, in both classes of games, it is not guaranteed that common belief in rationality can be satisfied, even if arbitrary finite orders of belief in rationality are satisfiable (see Dufwenberg and Stegeman, 2002, Example 2). Secondly, selecting the choices consistent with common belief in rationality might necessitate an elimination procedure of arbitrary transfinite length, rather than a countably infinite one (see Lipman, 1994). A notable implication of these remarks is that the gap between finite psychological games as in Definition 2.1 and psychological games with infinite choice- and player-sets becomes very small:

View all citing articles on Scopus
View full text