INTRODUCTION

This book's approach to cognitive science has focused on what I have called mental architectures, which are ways of carrying forward the basic principle that cognition is information processing. A mental architecture incorporates both a model of the overall organization of the mind and an account of how information is actually processed in the different components of the architecture. The emphasis in Part III was on different ways of looking at information processing. We examined both the computer-inspired physical symbol hypothesis and the neurally inspired artificial neural networks approach. In Part IV we turn our attention to the overall organization of the mind.

The concept of modularity is one of the basic concepts in theoretical cognitive science, originally proposed by the philosopher Jerry Fodor. Fodor's principle is that many information-processing tasks are carried out by specialized sub-systems (modules) that work quickly and automatically, drawing only upon a proprietary database of information. Those parts of cognition not carried out by specialized modules he describes as central processing. As we see in Chapter 10, Fodor is very pessimistic about cognitive science's prospects for understanding central processing. The massive modularity hypothesis (also considered in Chapter 10) offers one way of dealing with Fodor's concerns. According to the massive modularity hypothesis, there is no such thing as central processing. All cognition is modular and carried out by specialized sub-systems.