Heurísticos en la representación del conocimientoestrategias de clasificación de categorías difusas y bien definidas

Abstract

The hypothesis establishing that the system adjusts its descriptions to a linear transformation of the properties that result from the perceptual analysis has not been justified. This is seen with the formation of categories or concepts even when these categories do not fit with a monotonic model of representation. Two experiments analysed the strategies that allow for the description of a domain of knowledge in terms of a binary system of diffuse or well defined categories. We can infer, from subject descriptions, the formation of a non-monotonic system of rules of inference that allow for the distribution of stimuli in complementary classes, according to their contribution to the decisionmaking process. As opposed to the current model of categorization, it is proposed that the non-linear distribution of the stimuli in terms of the response alternatives is determinant of the categorization process.