Thomas J. H. Morgan: Arizona State University

The evolution of human behavior and cognition is often studied with a combination of theory and experiment—theory is used to explore evolutionary dynamics, and generate predictions that are tested empirically. However, agent psychology is typically highly simplified and so it can remain unclear whether human behavior would produce the same evolutionary dynamics. To address this question, I describe a new method, “experimental evolutionary simulations”, that combines aspects of theoretical and empirical approaches by inserting large numbers of human participants into an evolutionary simulation. I use this approach to provide new data concerning the ability of human social learning to adapt to an unstable environment. Theory has identified different social learning strategies that are highly successful or unsuccessful in a changing environment. Experimental evidence suggests that human behavior is broadly consistent with many of these predictions and so it remains unclear how well a population of human learners would cope with environmental change. Across a series of experimental evolutionary simulations, I find that although human behavior is broadly consistent with theoretical strategies that are successful in a changing environment, this similarity is insufficient to actually buffer human populations against environmental change. I conclude that human psychology is designed for high fidelity copying and not adapting to environmental novelties. More generally I suggest that experimental evolutionary simulations are a valuable complement to existing methods in the evolutionary study of mind and behavior.