Is the mind really modular?

When Fodor titled his (1983) book the Modularity of Mind, he overstated his position. His actual view is that the mind divides into systems some of which are modular and others of which are not. The book would have been more aptly, if less provocatively, called The Modularity of Low-Level Peripheral Systems. High-level perception and cognitive systems are non-modular on Fodor’s theory. In recent years, modularity has found more zealous defenders, who claim that the entire mind divides into highly specialized modules. This view has been especially popular among Evolutionary Psychologists. They claim that the mind is massively modular (Cosmides and Tooby, 1994; Sperber, 1994; Pinker, 1997; see also Samuels, 1998). Like a Swiss Army Knife, the mind is an assembly of specialized tools, each of which has been designed for some particular purpose. My goal here is to raise doubts about both peripheral modularity and massive modularity. To do that, I will rely on the criteria for modularity laid out by Fodor (1983). I will argue that neither input systems, nor central systems are modular on any of these criteria. Some defenders of modularity have dropped parts of Fodor’s definition and defined modularity with reference to a more restricted list of features. Carruthers (this volume) makes such a move. My arguments against modularity threaten these accounts as well. My claim is not just that Fodor’s criteria are not jointly satisfied by subsystems within the mind, but they are rarely satisfied individually. When we draw boundaries around subsystems that satisfy any one of Fodor’s criteria for modularity, we find, at best, scattered islands of modularity. If modules exist, they are few and far between. The kinds of systems that have been labeled modular by defenders of both peripheral and massive modularity probably don’t qualify. Thus, modularity is not a very useful construct in doing mental cartography.