Automatic and strategic effects in human imitation.

Imitation is a very important ability that allows an agent to acquire efficiently a wide range of movements by observing a demonstrator performing them. The tendency to imitate has long been noted in normal and abnormal human behaviour, and in the last decade there has also been an increase in the number studies of imitation in children, monkeys and great apes. In this chapter we will review some main issues related to imitation, including whether imitation is innate or is the outcome of a learning process, whether it is a uniquely human or is shared with other primates, as well as the most prominent accounts that have been put forward to explain how it works. We will then summarize the results from different sets of experiments (Tessari and Rumiati 2004; Rumiati et al. 2005; Tessari et al. 2006; Tessari et al. in press) which we have interpreted within a dual-route model. Normally, individuals use the sublexical, direct route to imitate novel, meaningless (ML) gestures and the lexical-semantic, indirect route (and possibly the sublexical route) to reproduce meaningful (MF) actions that are already in one’s repertoire. We will show how a reduction in cognitive resources, caused either by experimental manipulations with healthy participants (Tessari and Rumiati 2004) or by brain damage (Tessari et al. in press), may end up affecting the route selection in imitation. In particular, when MF and ML actions are presented intermingled, participants selected the direct route because it is suitable for imitating both stimulus types (Tessari and Rumiati 2004). Moreover, in the new Experiments 1-2 we also demonstrated that the sublexical route is selected even when either ML (Experiment 1) or both ML and MF actions (Experiment 2) were presented for more time. Finally we will illustrate the cerebral correlates of imitation of different action types, derived from the lesion analysis of patients with a selective stimulus specific apraxia (Tessari et al. in press), and from a PET study with healthy participants performing an imitation task (Rumiati et al. 2005). Based on our results, we propose that imitation of MF and ML actions is supported by common as well by dedicated brain regions.