The mind doesn't work that way

Jerry Fodor’s The Mind Doesn’t Work That Way (2000) serves as a sharp, intellectually salty rebuttal to the "New Synthesis" of cognitive science. At the time of its release, figures like Steven Pinker and Henry Plotkin were popularizing the idea that the human mind is a massive collection of evolved, computational modules. Fodor, despite being one of the fathers of the Computational Theory of Mind (CTM), stepped in to tell his colleagues they were getting ahead of themselves.His argument isn't that CTM is wrong, but rather that it is fundamentally limited. He contends that while computation explains how we perceive the world, it fails to explain how we think about it.The Critique of Massive ModularityThe cornerstone of evolutionary psychology is the "Swiss Army knife" model of the mind. This view suggests that the mind is composed of hundreds or thousands of "modules"—specialized mini-computers evolved to handle specific tasks like recognizing faces, detecting cheats, or finding a mate.Fodor argues that this "Massive Modularity" thesis is a mistake. He distinguishes between two types of mental systems: Vertical Systems (The Modules): These are the "input analyzers" like vision, hearing, and language processing. They are informationally encapsulated, meaning they don't have access to everything you know. For instance, knowing an optical illusion is a trick doesn't stop you from seeing it; the visual module is "walled off" from your general knowledge. Horizontal Systems (Central Processes): This is the "real" thinking—the part that weighs evidence, makes plans, and forms beliefs. Fodor argues these systems cannot be modular. To form a belief, the mind must be able to look at any information from any source. If a system is open to everything, it cannot be encapsulated, and if it isn't encapsulated, it isn't a module.The Problem of Global Properties and "The Frame Problem"Fodor’s most devastating critique involves the nature of computation itself. The Computational Theory of Mind posits that thinking is the manipulation of symbols based on their syntax (their shape or form) rather than their semantics (their meaning). A computer doesn't "know" what the symbol "7" means; it just follows rules for how that shape interacts with other shapes.The problem, Fodor argues, is that human thought is sensitive to global properties, which syntax cannot capture. A global property is something that depends on the context of the entire system. For example, the "simplicity" or "conservatism" of a belief isn't a property of the belief itself, but of how that belief fits with everything else you know.This leads to the Frame Problem. In AI, if a robot is told to get a battery out of a room, it must decide which facts are relevant to its mission. Does the color of the walls matter? Does the fact that it's a Tuesday matter? A local, syntactic processor has to check every single thing it knows to see if it’s relevant, leading to "computational paralysis." Humans, however, instantly hone in on what matters. Fodor argues that because CTM is local, it cannot explain how we handle global relevance.Abduction: The Mystery of ReasoningFodor points out that most human thought is abductive, or "inference to the best explanation." This is the process of looking at a set of facts and finding the most likely conclusion. If you wake up and see the grass is wet, you assume it rained. But if you also see a film crew with a hose, you change your mind.The Limits of AdaptationismFodor also attacks the "Evolutionary" part of Evolutionary Psychology. He critiques adaptationism—the idea that every complex mental trait exists because it was specifically selected for by natural selection to solve a problem in the Pleistocene era. "The fact that a phenotypic trait is complex and functional is not a sufficient reason to assume it is an adaptation."