Language of Thought

views updated

LANGUAGE OF THOUGHT

Simply stated, the language-of-thought thesis (LOT) holds that thinking (i.e., cognition) is carried out in a languagelike medium, where the thoughts that constitute thinking are themselves sentencelike states of the thinker. Since the demise of philosophical behaviorism in the early 1960s the LOT thesis has enjoyed considerable support as a central tenet of a more encompassing representationalist theory of mind (RTM). Proponents of RTM, led by Jerry Fodor, have mounted a sustained defense of LOT.

RTM offers an account of propositional attitudes—beliefs, desires, doubts, and so on—according to which propositional attitudes relate the possessor of the attitude to a mental representation (cf. Fodor 1981). Mental representations have both semantic and physically realized formal properties: They are semantically evaluable (e.g., as being true or false, as being about or referring to certain entities or properties); they stand in inferential relations to other mental representations; and, like words, pictures, and other representations, they also have certain formal properties (e.g., shape, size, etc.) in virtue of being physical, presumably neural, entities. Mental representations, and hence propositional attitudes, have their causal roles in thinking and behavior in virtue of their formal properties. Propositional attitudes inherit semantic properties from the mental representations that are one of their relata. RTM is silent as to what kind or sort of representation these mental representations are (cf. Fodor 1987, pp. 136–138).

LOT supplements RTM with a specific proposal or hypothesis about the character of mental representations: Like sentences of a language, they are structured entities, and their structures provide the basis for the particular semantic and causal properties that propositional attitudes exhibit. More specifically, they are syntactically structured entities, composed of atomic constituents (concepts) that refer to or denote things and properties in the world. The semantic properties of a mental representation, including both truth conditions and inferential relations, are determined by the representation's syntactic structure together with the semantic properties of its atomic constituents. Mental representations, in other words, have a combinatorial semantics. The causal properties of a representation are similarly determined by the representation's syntactic structure together with the formal properties of its atomic constituents.

Three sorts of arguments have been advanced in support of LOT. The first makes much of the apparent semantic parallels between thoughts and sentences. Both beliefs and declarative sentences, for example, are typically meaningful, truth valued, and intentional (in the sense of being about something). Both stand in various inferential relations to other beliefs and assertions. One obvious explanation of these parallels is that thought has a languagelike character, individual thoughts a sentencelike structure. A second sort of argument focuses on the productivity and systematicity of thought. Thought, like language, is productive in the sense that there are indefinitely many, indefinitely complex thoughts. Whatever can be said can also be thought. Thought, like language, is also systematic in the sense that you can think one thought (e.g., that the child bit the monkey) if and only if one can also think certain other systematically related thoughts (that the monkey bit the child). Again, one obvious explanation is that thought has a languagelike character, individual thoughts a sentencelike structure. A third sort of argument claims that much cognitive scientific theorizing seems committed to LOT. Specifically, our best theories of rational choice, perception, and learning seem committed to the claim, not simply that cognition is a matter of the creation and manipulation of mental representations, but also that these representations are sentential in character. It is claimed, for example, that our best theories of learning are a species of hypothesis testing. But such a procedure, it is argued, presupposes the existence of a language, that is, a language of thought in which the hypothesis being tested is formulated.

Proponents of LOT readily concede that these arguments are not decisive. Each is an instance of inference to the best explanation, and as such each is vulnerable to refutation by some alternative explanation that does not appeal to a language of thought.

Critics of LOT find the foregoing sorts of arguments unpersuasive for any of a number of reasons. Either they believe that there are equally good explanations that do not appeal to a language of thought, or they deny the phenomena that LOT is said to explain, or they hold that the proposed explanations either rest on false presuppositions or are so sketchy and incomplete as not to merit the name, or they believe that these explanations have entailments so implausible as to impugn the explanatory premise that there exists a language of thought. Thus, for example, the argument from learning discussed above apparently entails that to learn a language one must already know a language. Many critics find in this entailment a reductio of LOT. Proponents such as Fodor, by contrast, have courageously embraced this entailment, arguing that all concepts, including, for example, our concept of a Boeing 747, are innate. Whatever the specific merits and defects of the arguments and counterarguments, it seems fair to say that the existence of a language of thought remains an open empirical question.

Bibliography

defending lot

Field, H. "Mental Representations." Erkenntnis 13 (1978): 9–61.

Fodor, J. The Language of Thought. New York: Crowell, 1975.

Fodor, J. Representations. Cambridge, MA: MIT Press, 1981.

Fodor, J. "Why There Still Has to Be a Language of Thought." In Psychosemantics. Cambridge, MA: MIT Press, 1987.

criticizing lot

Churchland, P. M., and P. S. Churchland. "Stalking the Wild Epistemic Engine." Noûs 17 (1983): 5–18.

Dennett, D. C. "Brain Writing and Mind Reading." In Brainstorms. Montgomery, VT: Bradford Books, 1978.

Dennett, D. C. "A Cure for the Common Code." In Brainstorms. Montgomery, VT: Bradford Books, 1978.

Loar, B. "Must Beliefs Be Sentences?" In Proceedings of the PSA, 1982, edited by P. Asquith and T. Nickles. East Lansing, MI: Philosophy of Science Association, 1983.

Schiffer, S. Remnants of Meaning. Cambridge, MA: MIT Press, 1987.