Abstract

Contemporary Cognitive Psychology has developed a research program that has two assumptions: 1) superior cognitive functions are representations and 2) to know is to compute. Therefore representations are the main theorefical concept. Traditionally two theories argue about the status of imagery in cognitive process. Images are considered mere epiphenomena from the descriptionalistic point of view. By the contrary, the imaginistic theories support images as genuine phenomena to be explain. Different interpretations of empirical tests are claimed. In 1978 J. R. Anderson considered that it is impossible to settle the debate appealing to behavioural data. He proposed a general and formal argument to prove that given any pair is possible to build other pairs with different representations for the same behavioural data. Differences in representations are made amends for different processes. In this paper, consequences of Anderson´s argument are put into question and another strategy for considering imagery is suggested