Patrick Perret
This paper explores, through the example of the class inclusion literature, the major hypotheses that have been proposed to account for competence- performance discrepancies in the course of cognitive development. It is shown that these hypotheses focused successively on each of the three parameters stated by Flavell and Wohlwill's initial model. A first set of empirical studies has established that performance varies as a function of task-related variables. A second set, making use of tasks with more complex logical demands, accounts for performance variability by a lack of logical competence that is supposed to emerge at adolescence. From a third perspective, competence is available early and discrepancies result from the inefficient inhibition of competing knowledge structures. After a discussion of the theoretical difficulties met by these alternative explanatory frameworks, we suggest that a « knowledge-levels » approach is necessary to coordinate the various hypotheses.