ON COMPUTATIONAL EVIDENCE FOR DIFFERENT TYPES OF SPATIAL RELATIONS ENCODING - REPLY TO COOK ET-AL (1995)

Computational models in psychology play an increasingly important role in characterizing theoretical distinctions, understanding empirical r esults, and formulating new predictions. However, the proper use of mo dels is subject to debate and interpretation, as Cook, Fruh, and Landi s (1995) have demonstrated in a critique of neural network simulations reported by Kosslyn, Chabris, Marsolek, and Koenig (1992). These simu lation results supported a distinction between two types of spatial re lations encoding. Cook et al. argue that Kosslyn et al.'s models did n ot process ''spatial'' representations and that input-output correlati ons rather than properties of spatial relations encoding processes exp lain the performance of the models. This article provides conceptual a nd analytic rebuttals of those criticisms.