ABSTRACT
This book has explored a number of areas of human cognition and has illustrated the contributions made by both cognitive neuropsychology and connectionist modelling in each area. This section attempts an assessment of how useful these approaches are. Traditionally cognitive psychologists have relied on observation and experiment to try to understand the mechanisms of cognitive processes. Neuropsychology and computational modelling have provided additional research tools and enable researchers to use converging evidence from different sources to build, test, modify, and confirm or disconfirm their theories. The new approaches have benefited cognitive theorising in two particularly valuable ways. First, they have added a lot of fine-grain detail to the prevailing broad-brush accounts of cognitive processes. Second, they have added many important constraints to these models by showing what is, or is not, possible; what can be done and what cannot be done. Neuropsychological studies of damaged brains can reveal which modules are essential for normal function and which are not. Computational models can challenge established cognitive theories by showing that it is possible to simulate human performance using a different system. However, the concept of simulation rests on the idea that brains and machines are in some ways analogous: the rationale for cognitive neuropsychology rests on the idea that intact brains and damaged brains are analogous.
