Formal evolutionary modeling and the problems of political science.

다국어 초록 (Multilingual Abstract)

Formal theory in political science is an approach to studying political phenomena through the language of logic and mathematics. Is a science of politics possible? Formal political theory has lead to many advances in the discipline; however, it is facing an increasing number of challenges. I argue that some of the most serious challenges to formal theory can be addressed with a synthesis of game theory, computational modeling, and the theory of evolution. Ultimately, I claim that the science of politics is, in fact, possible if political theory is based upon a rigorous foundation of mathematics and if it is also systematically connected to the life sciences.
I start by examining common challenges to formal political theory, including the assumption of hyper-rationality and a related problem of incorrect predictions about human behavior. This examination suggests that the fundamental problem that formal theorists face is not methodological but rather substantive--an imperfect model of a man.
To emphasize the importance of a good underlying such model, I present the case of prospect theory--a formal scientific analysis of human behavior devoid of sound evolutionary basis (in social sciences).
To discover a good model of a man, formal theorists may have to turn to the theory of evolution. However, the solution will be more substantive than methodological--as is the problem. Evolutionary theory is a formal theory and it is a natural step forward for formal theorists since the logic of evolution can be expressed mathematically. I show that formal evolutionary modeling--evolutionary game theory, models of adaptive learning, and evolutionary computation (computer simulation)--can be useful for addressing some of the most challenging problems of political science without abandoning the rigor of logic and mathematics.
I apply formal evolutionary models to three different issues: endogenous agenda-setting, cooperation and altruistic punishment, heroism and intergroup violence. The models I develop are designed to provide realistic empirically testable predictions consistent with the view of human behavior now emerging from the life sciences.