ABSTRACT
In Piagetian theory, cognitive development was characterized as a progression through a series of stages that culminated in formal operational logic. For Piaget (1947/1950), logic refl ected properties that are inherent in thought. However, the assumption that human thought is logical has been questioned and alternative models that do not incorporate this assumption have been proposed. These include information processing theories (Anderson, 1983; Newell & Simon, 1972), as well as approaches based on heuristics (Kahneman, Slovic, & Tversky, 1982), rational analysis (Anderson, 1990, 1991) and mental models (Johnson-Laird & Byrne, 1991; Markovits & Barrouillet, 2002). One implication of these developments is that if human reasoning is not logical then criteria based on normative logic will be inappropriate for evaluating thinking. Different criteria for evaluating reasoning in children and adults are required. In the current and earlier versions of Halford’s theory (Halford, 1982, 1987, 1993; Halford & Wilson, 1980; Halford, Wilson, & Phillips, 1998) there is an emphasis on providing an objective basis for quantifying the complexity of cognitive tasks and the resulting demand on cognitive resources. Halford and Wilson (1980) expressed complexity in terms of levels of representational structure as derived from Category Theory. In the most recent refi nement, Relational Complexity theory (Halford et al., 1998), complexity is defi ned in terms of the complexity of the mental models that underlie thinking. Cognitive development is characterized as the ability to construct mental models that involve more complex relations. Thus relational complexity replaces Piaget’s psycho-logic as the index of cognitive development. If relational complexity is to serve an alternative criterion to logic and provide a viable account of cognitive development, then there are a number of requirements that should be met. First, the theory should provide a principled way to analyse cognitive tasks and to quantify their complexity. Second, the theory should be applicable in different content domains. Third, the theory should be capable of making predictions in advance of data. Fourth, predictions derived from the theory should be supported by empirical evidence. Relational Complexity theory makes many predictions. Of most rele-
vance here are the effects of task complexity on performance, the relative sensitivity of tasks with different complexity levels to age-related change, the extent to which tasks at a given complexity level form an equivalence class, and within-person consistency in performance. In this chapter, we illustrate application of the relational complexity approach to a selection of tasks used in cognitive developmental research. Our purpose is to evaluate the extent to which Relational Complexity theory meets the requirements outlined above. In Relational Complexity theory, as in other neo-Piagetian theories, including those of Case (1992), PascualLeone (1970) and Demetriou, Efklides, and Platsidou (1993), cognitive development is underpinned by growth in information processing resources. We will outline our views regarding the nature of this resource, and its links with reasoning. Finally we will demonstrate that our approach is consistent with research about brain function and maturation of the prefrontal cortex. We start by describing the theory.
