One adaptive function that has been suggested for social play in higher primates is the learning, refining, or practicing of communicative skills. Despite the absence of experimental data, a comparison of the structure of aggressive play and communication may shed light on this hypothesis. The great majority of rhesus monkey agonistic interactions are mediated by stereotyped signals. Deprivation experiments have shown that appropriate interpretation and use of such signals requires early social experience. Such experience seemingly must have the following characteristics: (i) The relevant agonistic signals must occur, (ii) These signals must regularly be paired with other stimuli for appropriate responses to signals to be learned. Signals must produce regular responses in other monkeys for appropriate use to be learned. Aggressive play seems to be an unpromising context in which to learn, refine, or practice agonistic signals. These signals are not observed in play, and the signals that do occur are restricted to play and are not regularly paired with unconditioned stimuli.