Mechanisms of imitation

This thesis aims to discover the mechanisms of imitation by testing the predictions of three theories. These are Associative Sequence Learning Theory (Heyes and Ray, 2000), Ideomotor Theory (e.g. Prinz, 1997), and Active Intermodal Mapping (e.g. Meltzoff Moore, 1997). Chapter 1 identifies three issues upon which the theories of imitation can be differentiated. The first is concerned with the development of effector-dependent representations through observation. The second and third relate to the role of awareness and experience in imitation. These differences form the basis of the experiments reported in Chapters 2, 3 and 4. Experiments 1 - 3 (Chapter 2) investigated whether effector-dependent representations could be formed through action observation. A series of tests based on the serial reaction task (SRT) were utilised. It was found that with relatively short, simple movement sequences, participants learned the structure of the sequence as effector-dependent motor representations. Sequence knowledge could not be expressed using effectors other than those used by the observed model. Experiments 4 - 6 (Chapter 3) used similar tests to those in Chapter 2 but investigated whether a longer, more complex, movement sequence could be learned implicitly i.e. without concurrent awareness. Two experiments suggested that observation of a movement sequence, but not inanimate stimuli, could support implicit learning. Experiments 7 and 8 (Chapter 4) investigated the role of experience in imitation. It was shown that while responses made to movement stimuli were faster when stimulus and response movements matched, compared to when they were different, the advantage for matching movements disappeared after incompatible training. This result supports an experience-based, rather than innate, view of imitation. The results of the experiments reported in this thesis suggest imitation is experience- based, supports effector-dependent learning by observation, and can operate without awareness. This combination is best described by Associative Sequence Learning Theory.