Modularity and attention

Abstract

There is abundant evidence that the visual information-processing systems of higher animals are modular systems-systems consisting of relatively isolated networks of neurons that take the visual world apart (for example, Felleman & Van Essen, 1991, for an overview). Within contemporary psychology this modularity is often seen as a problem that requires a solution. The problem is: How does the brain put together all these different activities to produce a unified picture, so that, for example, for any object the right colour is associated with the right shape? The solution is that focal attention integrates the initially separable features in unitary objects.

In this contribution the opposite point of view is defended-not modularity as a problem solved by attention, but attention as a problem solved by modularity. The starting-point is that contemporary experimental psychology has to explain observed behaviour, not visual perception as a subjective experience. In these explanations the theoretical construct “focal attention” forms a major theoretical or conceptual problem. That problem can be solved in terms of modularity. It is argued that the modular character of the information-processing system is at the basis of the solutions nature has offered to solve two major selection problems in vision: the selection of an action and the selection of the object to act upon.