Given that many behavior patterns cluster together in sequences that are organized to solve specific problems (e.g., foraging), a fruitful perspective within which to study behaviors is as distinct 'behavior systems'. Unlike many behavior systems that are widespread (e.g., anti-predator behavior, foraging, reproduction), behavior that can be relegated as playful is diverse, involving behavior patterns that are typically present in other behavior systems, sporadic in its phylogenetic distribution and relatively rare, suggesting that play is not a distinct behavior system. Yet the most striking and complex forms of play have the organizational integrity that suggests that it is a behavior system. One model that we develop in this paper, involves three stages of evolutionary transition to account for how the former can evolve into the latter. First, play-like behavior emerges from the incomplete development of other, functional behavior systems in some lineages. Second, in some of those lineages, the behavior patterns typical of particular behavior systems (e.g., foraging) are reorganized, leading to the evolution of specific 'play behavior systems'. Third, some lineages that have independently evolved more than one such play behavior system, coalesce these into a 'super system', allowing some animals to combine behavior patterns from different behavior systems during play. Alternative models are considered, but irrespective of the model, the overall message from this paper is that the conceptual framework of the behavior system approach can provide some new insights into the organization and diversity of play present in the animal kingdom.
