Humans are distinguished from other animals by their extensive imitation repertoires, but the determinants of this behaviour are still not fully established. The present programme of research explored the relationship between children's ability to name the components of modelled actions and their matching accuracy, in tests that presented them with modelling of empty-handed gestures that terminated on different parts of body. A total of 106 children, aged between 2 and 4 years, participated in seven experiments.
The first two studies, presented in Chapter 5 and Chapter 6, employed correlational group designs to establish which body parts and movements children can accurately name in response to the experimenter's modelling, and produce in response to the experimenter's verbal instructions. Children's responses were more accurate for body parts than for body movements; they responded best to those body parts that commonly feature in naming-and-matching games played with their caregivers; and their patterns of errors were similar to those previously observed in imitation tests. Next, three single-case experiments presented in Chapter 7 systematically tested the effects of tact and listener training of previously unnamed body parts on children's generalised imitation of gestures that terminated on these locations. Children's matching of hand-to-body gestures markedly improved as the result of body part naming training. Next, a single-case experiment presented in Chapter 8, and a group experiment presented in Chapter 9, tested the effects of training the children to name the movements "across" and "to the side" on the accuracy of their matching of hand-to-body gestures that incorporated these movements. This training was effective in reducing the rates of ipsilateral mismatches to contralateral models, which are typically very high in imitation tests.
These results are in agreement with the predictions of the naming account of Horne and Lowe (1996), who consider naming to be the earliest form of self-instructional behaviour. The present results show that the ability to name features of the actions that they see is an important determinant of imitation in children. These findings have wide-reaching implications for the interpretation of children's performances in matching tests, and for the evaluation of behaviour-analytic and cognitive-developmental accounts and theories of imitation.