<< 前へ │ 次へ >>

# The language of thought (Fodor) #1
Jerry Fodor is a well-known contemporary philosopher concerned mainly with philosophy of the mind and cognitive science. Fodor strongly supports the position of “computationalism” in the area of contemporary cognitive science. Computationalism emerged from the progress of artificial intelligence (AI) in the 1950s and posits that cognitive processes are a form of computation. This view suggests that the mind operates similarly to a computer, processing information through computational means. Fodor philosophically tackles this view of computationalism, and he suggests that we have, in our mind, mental representations of propositional attitudes (beliefs, desires, thoughts, etc.) before translating them into natural languages such as English, Russian, and so on. This position is called the "representational theory of mind" (RTM), and according to Fodor, mental representation is composed of a syntactic structure, just like ordinary language. He terms this innate and universal language “the language of thought.” However, several arguments have been made against Fodor's claim. Another possibility is the approach called "connectionism," which models cognitive processes through neural networks rather than symbolic representations. In this essay, we are concerned with Fodor's language of thought thesis, its criticisms, and the alternative theory, connectionism. 　　　　According to Fodor, computationalism is based on mental represen-tationalism. Thus it should be helpful to consider familiar examples that represent something, such as sentences, photographs, maps and so forth. Mental representation can also be thought of as representing one of these items. In fact, Fodor argues that mental representation is quite analogous to sentences because they both possesses a syntactic structure. He called these mental sentences “the language of thought” (or “Mentalese”). However, it certainly would not be prudent to maintain that the language of thought serves to represent something for the same reason as the natural language, because we are not consciously aware of this innate language. Therefore, Fodor shows some evidence to prove that our thoughts are a kind of language, and that they could manage to represent something, just like the natural language does. 　　　　Fodor suggests that the language of thought should be understood as being composed of and ruled by a syntactic structure. Fodor and Pylyshyn (1988) state that: [R]epresentations of the system have a combinatorial syntax and semantics: structurally complex (molecular) representations are systematically built up out of structurally simple (atomic) constituents, and the semantic content of a molecular representation is a function of the semantic content of its atomic constituents together with its syntactic/formal structure. From this feature of the language of thought, they point out that thought clearly possesses two grammatical features of language: productivity and systematicity. At first, the productivity of language is one of the linguistic features of human beings, with which we can generate an infinite number of sentences from a limited vocabulary. In other words, an ordinary speaker, who properly understands the rules for making sentences and has "enough" (but a limited) vocabulary, can construct an uncountable number of distinct sentences even though he has never heard most of the sentences that he can produce. For example, “There is a big lake of melted gold on the dark side of the moon” (Aydede, 2004). Thus the productivity of thought is understandable in the same way as the natural language: thoughts involve a structural feature of representation analogous with the natural language. Second, the language of though consists of another important feature of language, systematicity. For example, if someone learned the sentence, “John loves Mary,” then nothing could prevent us from expecting him to be able to learn the sentence, "Mary loves John" as well. Both sentences are structurally the same, and the second sentence simply reverses the object and subject of the first sentence. In other words, their constituent parts have been systematically rearranged. According to the above quotation by Fodor and Pylyshyn (1988), sentences are not atomic (the minimum unit), but have a syntactic constituent structure. If someone were a native English speaker, his knowledge of its grammar tells him how to systematically combine words to produce sentences. Moreover, Fodor and Pylyshyn say, “If one views all sentences as atomic, then the systematicity of language production is a mystery, but if he acknowledges that sentences have syntactic constituent structure, systematicity of linguistic capacities is what you automatically get; it is guaranteed. This is the orthodox explanation of linguistic systematicity” (Fodor and Pylyshyn, 1988). For the above reasons, Fodor concludes that the language of thought can be thought of as possessing sufficient character-istics of language; and cognition is the process of transforming the language of thought into the natural language. Therefore, this compatible process between the innate and common languages is well matched to the view of computationalism. --- ## 注 --- ## 参考文献

<< 前へ │ 次へ >> │ 一番上に戻る