Game theory without theory: Interactive choice in pigeons, humans and machines
[Display omitted] •A novel methodology is described for the experimental analysis of cooperation, competition and other social behaviors.•Pigeons in separate chambers are coupled via the computer programs that control their contingencies of reinforcement.•This technique can be extended to study Pris...
Gespeichert in:
Veröffentlicht in: | Learning and motivation 2022-02, Vol.77, p.101784, Article 101784 |
---|---|
1. Verfasser: | |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | [Display omitted]
•A novel methodology is described for the experimental analysis of cooperation, competition and other social behaviors.•Pigeons in separate chambers are coupled via the computer programs that control their contingencies of reinforcement.•This technique can be extended to study Prisoner’s Dilemma and many other games of Game Theory.•Two-player choice converges at the game’s Nash Equilibrium in games that have one and to near-equality in those that do not.•Pigeon and human results completely explained by the principle of positive reinforcement and Nash’s equilibrium concept.
A unique approach to the experimental analysis of interactive choice behavior is described. Pairs of subjects were coupled in such a manner that each subject’s choice determined either the probability of reinforcement or the points awarded to the other subject’s choice. In this manner, several of game theory’s familiar games, like Prisoner’s Dilemma, were created. With pigeons as subjects, with humans playing against pigeons or humans playing each other (either in the laboratory or on the Internet), and even with a simple computer simulation, the results were clear. In every instance, players’ choices gravitated to the game’s pure strategy Nash equilibrium in games that have them and to a near equality of responding in the game Rock-Paper-Scissors, which does not. The principle of positive reinforcement entirely accounts for the results. There is no need for appeal to complex mathematical models or higher-order concepts of strategy, optimality, rational expectations and the like. |
---|---|
ISSN: | 0023-9690 1095-9122 |
DOI: | 10.1016/j.lmot.2022.101784 |