Anderson’s hypothesis suggests that even if we can fully uncover the microfoundations of behavior—for example, acquire a complete specification of the psychological aspects of behavior or the probability of interaction—we may still not have a simple way to understand their macrolevel implications.
