Reasoning, Perception and Beliefs in Strategic Settings: Theory, Behavior and Cognition - Home

From Santa Fe Institute Events Wiki

Working Group Navigation

Aims and Scope


While game theory, the theory that analyzes purposive behavior in strategic settings, has had notable successes in predicting behavior in some cases, the theory fails to provide an accurate prediction in a large class of real-life situations as well as experimental settings. The conundrum that still plagues game theory today, 65 years after the publication of Von Neumann and Morgenstern's book that set off the field, is how individuals perceive strategic situations, reason about others, and form beliefs about others' perceptions, reasoning processes and beliefs. That is, the interactive nature of strategic settings makes the problem of predicting behavior in such settings fundamentally more difficult than that of predicting choice behavior in single-person decision problems, where ideas and concepts from psychology have been incorporated into general theories with considerable success.

Adaptations to the theory based on experimental evidence on behavior have led to better predictions in some settings, but have failed to provide a general theory based on intuitive and basic principles. At the same time, a good understanding of how evidence on human cognition can be incorporated into game-theoretic models is still lacking. For instance, can we use evidence on human cognition to develop models of bounded rationality in games? Is it possible to use experimental results on people's Theory of Mind—the mechanism that people use to infer and reason about others' state of mind—to provide better predictions of human behavior?

This working group aims to revisit the fundamental assumptions underlying game theory, in particular questioning the modeling of individuals' representations, reasoning processes and beliefs, with an eye towards empirical evidence of human behavior as well as on cognitive processes. It does so by bringing together game theorists (in particular epistemic game theory and learning), decision theorists, cognitive neuroscientists, psychologists, and computer scientists, to discuss problems of joint interests, and to confront perspectives. Presentations are scheduled in such a way that there will be ample time for discussion among the select group of invitees.


Confirmed Participants