Human behaviour is intrinsically goal-directed. In the laboratory, experimenters usually impose tight constraints on which goal(s) will be chosen to control current behaviour, but in any more complex environment we need an account of how goals are chosen among competing, often unrelated alternatives. As background to the discussions, I summarize some recent findings on component skills in driving. Questions addressed include the relative skills of experts, normals, and novices; individual differences in both skill and accident involvement; and distraction by a secondary task. An account of goal choice in a complex world is then sketched. Perceptual data are combined with internal preferences to guide a competition between alternative momentary goals. A central role is given to the efficiency of goal-weighting, whereby useful sub-goals are selected by joint activation from current and goal states. This idea is linked to work on frontal lobe dysfunction, 'intelligence' and dual tasks. Learning and the development of routine behaviour are also considered. Finally, the findings on driving skills are discussed within this theoretical framework, and some practical implications are suggested. (A) For the covering abstract of this conference see IRRD 834497.
Abstract