State-dependent behavioral theory and the evolution of play

Abstract

We show how state-dependent behavioral theory, as implemented by stochastic dynamic programming can enhance our understanding of play, assuming that play is an evolutionary adaptation rather than a non-functional consequence of other processes. We first explain in generic terms the components of a state-dependent behavioral model. We then develop a state-dependent life history model on the assumption that social play is adaptive and thus focus on play’s contribution toward the development of skill and how such development affects an individual’s future lifetime expected reproductive success (i.e. fitness that accounts for both survival and reproduction). Although our model is not intended to capture a specific situation, a number of commonly observed phenomena emerge from the model, including that most social play is dyadic and that animals select partners that are of similar age or skill. We discuss ways of increasing the fidelity of the model by allowing injury due to play, the role of energetic reserves, handicapping, and polyadic play.