Fodor, Jerry A. (1935–)

views updated


Jerry Fodor is the most significant philosopher of mind in the last fifty years. A student of Hilary Putnam, he joined him, Noam Chomsky, and others at MIT in the early 1960s and became the philosopher most responsible for the "cognitive revolution" that replaced the behaviorism that had dominated much of philosophy and psychology since the 1920s, replacing it with a computational approach derived from the work of Alan Turing. In this way he hoped to provide a basis for a naturalist and realist account of mental processes that rendered them amenable to scientific study. Indeed, he is one of the few philosophers who has combined philosophical and empirical psychological research, publishing work in both domains, and developing at least two theories that have become highly influential in each: a computational/representational theory of thought processes ("CRTT") and a "modularity" theory of perception and linguistic processing.

CRTT is an effort to salvage what Fodor (1975) regards as essential to the familiar "belief/desire," or "(propositional) attitude" psychology with which people routinely explain each other's behavior, as when one explains someone's crossing a road in terms of a desire to meet a colleague. As the name emphasizes, the theory has two parts. According to the computational part, each attitude involves a computationally specifiable relation to a syntactically specifiable representation in a "language of thought" entokened in the agent's brain. For example, judgment might be the output of perceptual and reasoning systems that serves as the input to decision making. For the "representational" part, Fodor (1998) argues at length against popular "prototype," "conceptual role" and "holistic" theories of content, and defends instead an "atomistic," "informational," "asymmetric dependency" theory according to which, (i) ceteris paribus, tokenings of symbols causally co-vary with phenomena that they thereby mean; and (ii) tokenings caused by phenomena they don't mean depend upon (i), but (i) doesn't depend upon them. For example, "horse" means horse if (i) it's a ceteris paribus law that "horse" tokens are caused by horses, and (ii) nonhorses (e.g., distant cows) causing "horse" tokens depends on horses doing so, but not vice versa (1991). Thus, Jones's judgment that horses fly might consist in a sentence, "Horses fly," playing the aforementioned judging role in her brain, where "Horses" and "flies" are each asymmetrically dependent upon the respective phenomena in the world. In this way Fodor hopes to defend intentional realism, in contrast to the widespread eliminativism about the mental, and mere instrumentalism about psychology, that renders psychological ascription a matter of mere "interpretation," such as one finds in the work of Willard Quine, Donald Davidson, Daniel Dennett, and Paul and Patricia Churchland. (Fodor's [1983, 1998] account of representation also leads him to claim that virtually all concepts expressed by single morphemes in natural language are innate, rejecting the empiricism also associated with these figures.)

CRTT is a species of functionalism, or the view, due originally to Putnam, that mental states are to be individuated by their causal relations, for example, to inputs, outputs, external phenomena, and each other, in ways analogous to the individuation of a program in a computer (1968). Since different physical phenomena can realize these relations, functionalism naturally gives rise to cross-classificatory layers of explanation: one level of causal relations may be "multiply realized" by different mechanisms at lower levels (1968, 1975). Specifically, the intentional level of a cognitive psychology may be implemented at a lower level by various computational syntactic processes, which in turn may be implemented by different physical mechanismsbrains in the case of people, transistors in the case of machines. For this reason, intentional psychology enjoys a considerable "autonomy" from levels of explanation closer to the brain, for example, neurophysiology. However, although the laws and explanations at the intentional level are not reducible to laws and explanations at the lower levels, Fodor presumes that they "supervene " on them.

One of Fodor's (1986) main arguments for CRTT is that it promises to account for the sensitivity of human beings to indefinitely many properties that are not "transducible" by sense organs, in particular, to arbitrary nonphysical and/or nonlocal properties, such as being a morpheme or a noun phrase, a crumpled shirt, a grieving widow, or a collapsing star. These sensitivities are particularly impressive given that they seem to be (i) productive and (ii) systematic (1987): that is, people seem capable of discriminating stimuli of indefinite logical complexity, such as being a crumpled shirt that was worn by the thief who stole the cat that chased the rat ; and anyone capable of thinking one logical form is capable of thinking logical permutations of it: for example, one can think John loves Mary if and only if one can think Mary loves John (1987). Fodor (1968, 1987, 1988) argues that non-CRTT accounts, such as behaviorism, Gibsonianism, and purely connectionist accounts are either vacuous or empirically inadequate for this task. What one needs is a system that can exploit internal processes of logical combination, inference and hypothesis confirmation, which presuppose at the least the resources of a CRTT.

However, Fodor (1983) has also been a critic of the "New Look" theories of perception, such as one finds in the work of Jerome Bruner, Thomas Kuhn, and Nelson Goodman, which emphasize how people's background expectations color their perceptions. Against this view, Fodor calls attention to the fact that the very perceptual illusions that New Look theorists prize actually tell against their case: for many of these illusions do not disappear even when one knows better, suggesting, along lines developed by Zenon Pylyshyn, that visual perception occurs in a "cognitively impenetrable module," that is "informationally encapsulated" from the "central" system whereby we reason and fix our beliefs. Fodor argues for a similar view of linguistic perception.

By contrast, the central system is "Quinian" (i.e., computed over the totality of beliefs, as when people settle on a theory that is, for example, simplest and most conservative overall) and "isotropic" (every belief is potentially relevant to the confirmation of every other, as when radio waves confirm the age of the universe). This leads Fodor (1983, 1999) to somewhat pessimistic conclusions regarding the tractability of central reasoning to a Turing-style CRTT, which depends upon exploiting local syntactic features of representations. Although CRTT is necessary for an adequate theory of mind, it may not be sufficient.

See also Behaviorism; Chomsky, Noam; Computationalism; Content, Mental; Davidson, Donald; Dennett, Daniel C.; Functionalism; Goodman, Nelson; Kuhn, Thomas; Language of Thought; Mental Representation; Putnam, Hilary; Quine, Willard Van Orman; Turing, Alan.


works by fodor

For a full bibliography of Fodor's work up until 1991, see B. Loewer and G. Rey, Meaning in Mind: Fodor and his Critics (Oxford: Blackwell, 1991), which also contains a long introduction to his work and critical essays by a number of prominent philosophers and cognitive scientists.

Psychological Explanation. New York: Random House, 1968.

The Language of Thought. New York: Thomas Y. Crowell Co., 1975.

"How Direct is Visual Perception? Some Reflections on Gibson's 'Ecological Approach.'" With Z. Pylyshyn. Cognition 9 (1981): 13996.

Representations: Essays on the Foundations of Cognitive Science. Cambridge, MA: MIT Press, 1981.

The Modularity of Mind: An Essay on Faculty Psychology. Cambridge, MA: MIT Press, 1983.

"Why Paramecia Don't Have Mental Representations." In Midwest Studies in Philosophy, vol. 10, edited by P. French, T Uehling, Jr., and H. Wettstein. Minneapolis: University of Minnesotta Press, 1986.

Psychosemantics. Cambridge, MA: MIT Press, 1987.

"Connectionism and Cognitive Architecture." With Z Pylyshyn. Cognition 28 (12) (1988): 371.

A Theory of Content and other Essays. Cambridge, MA: MIT Press, 1990,

Holism: A Shopper's Guide. With E. Lepore. Oxford: Blackwell, 1992.

Concepts: Where Cognitive Psychology Went Wrong, Oxford: Oxford University Press, 1998.

The Mind Doesn't Work That Way: The Scope and Limits of Computational Psychology. Cambridge, MA: MIT Press, 2000.

Georges Rey (2005)