Interacting cognitive subsystems: a framework for considering the relationships between performance and knowledge representations

Abstract

The paper has two principal aims. The first is to explore the role played in expertise by information represented in different mental codes. The second is to forge some links between research in the more traditional problem-solving domains and research in the domain of human-computer interaction. Barnard's (1985,1987) model of human information processing, Interacting Cognitive Subsystems (ICS), is referred to extensively in order to do this. The ICS framework distinguishes several different subsystems, each specialised for processing information represented in particular mental codes. The paper explores the relationships between tasks and different types of mental representations. Examples from traditional problem-solving domains and from the human-computer interaction domain are analysed in terms of their requirements for different classes of mental representations in order to show how the approach provides useful insights into the development of skilled behaviour.