Introduction

A central aim of imitation research is to identify how complex patterns of motor output are generated based on observing a similarly complex pattern of motion produced by a model. How is imitative action informed by perception? During the 1990s a number of findings, and most markedly the discovery of “mirror neurons” by Rizzolatti and coworkers, supported the idea that motor structures are already involved in action perception and not only when reproducing the observed action. On first sight, this idea appears to provide a clear answer to the issue of imitative perception-action mediation: a motor representation of the observed act is formed already during model observation, which should allow its reproduction with high fidelity, at least when the model's action is in the behavioral repertoire of the observer. Also, this idea seems to depart from earlier theorizing about imitation and observational learning, in which typically two temporally distinct stages were implied: the formation of a cognitive (e.g., verbal or iconic) representation during action observation, and the later “translation” of this cognitive representation into action (e.g., Carroll & Bandura, 1990; Keele, 1986; Meltzoff, 1988). These two stages could be a few seconds, or even days, apart. Accordingly, this view may be characterized as late mediation between perception and action, whereas mirror neurons indicate the possibility of early mediation. In this chapter, I investigate the consequences of early mediation accounts for functional theories of human imitative behavior.