A three part investigation is proposed to clarify the discrimination of action units in the behavior stream. In the first phase, behavioral event sequences that regularly occur in the experience of normal pre-school age children will be identified and filmed. These sequences will be analyzed by means of a movement notation in order to identify their surface, or movement structure. Based on some previous developmental findings, it is hypothesized that sequences whose surface structure provides salient movement cues which co-occur with action unit boundaries will produce the least ambiguous and most elaborated cognitive representations. In the second phase of the research, novel action sequences will be produced which possess or do not possess this hypothesized optimal structure, and their encoding studied over age. It is anticipated that younger children will benefit the most from redundant surface structures in the behavior stream. In the third phase, sequences from the first two parts will be transformed so as to enhance or eliminate supportive surface structure. It is hypothesized that such manipulations will have the greatest impact on young children who lack both a generalized event representation for the specific class of event, and a set of perceptual categories differentiated from unrelated generalized event representations down to the level of individual actions. Results could have wide implications for theories of social cognition, and could have wide application to the way in which children learn from observation.
