Information theory and Saussure’s speech circuit

3. The Code Model Decoded 79

3.3.3. Information theory and Saussure’s speech circuit

As mentioned previously, the fact that Saussure cast his model of communication in electromechanical terminology, as a circuit model, undoubtedly contributed to its later integration with the electrical engineering model offered in information theory. Harris comments: Modern technology is deeply committed to circuit models. Without them, it is no exag- geration to say, technology as the modern world knows it could scarcely exist. … Bearing this in mind, it is not a naïve question to ask why Saussure unlike Locke insisted that speech involved a circuit; and whether as a matter of fact circuit was an appropriate term to choose. Harris 1987 :213 As has been mentioned previously section 3.2.2.2.4 , Saussure himself employs the term code in discussing the langue. Harris translates: Speech … is an individual act of the will and the intelligence, in which one must distin- guish: 1 the combination through which the speaker uses the code provided by the language in order to express his own thought, and 2 the psycho-physical mechanism which enables him to externalise these combinations. Saussure 1983:[30–31]; italics added Accordingly, if a contemporary theoretician employs the term code alone, without employing associated code model terms e.g., encode, decode, message, transmission, that does not necessarily indicate an appeal to the code model. This is particularly so if the point of discussion is Saussure’s theory. For example, Koerner writes: At the core of Saussure’s linguistic theory is the assumption that language is a system of interrelated terms which he called ‘langue’ in contradistinction to ‘parole,’ the individual speech act or speaking in general. This ‘langue’ is the underlying code which ensures that people can speak and understand each other; it has social underpinning and is an operative system embedded in the brain of everyone who has learned a given language. The analysis of this system, Saussure maintains, is the true object of linguistics. Koerner 1994 :3663; italics added When theoreticians employ multiple examples of information theoretic terminology, however, the collective choice of terminology indicates at least a limited appeal to the code model. For example, consider the following quotation from Hartmann and Stork 1972 . Having defined langue and parole, they then paraphrase their definitions via an appeal to information theoretic terminology: Language may be said to have two facets: Langue refers to the system of language which is passed on from one generation to another, e.g. the grammar, syntax and vocabulary, whereas parole refers to all that which a speaker might say or understand. Langue is the social, con- ventional side of language; parole is individual speech. Another way of expressing the difference is to say that langue is the code and parole is the message. Hartmann and Stork 1972 :126; underscore added While readers may rightly disagree with Hartmann and Stork’s paraphrasing langue and parole as code and message, the fact that they elect to use such a characterization demonstrates their willingness to fuse these constituent models. 80 3. The Code Model Decoded One of the more common theoretical areas where Saussure’s speech circuit is integrated with information theory is in discussion of semiotics. In fact, it is so common in this domain that some theoreticians even call the code model account the semiological perspective e.g., Givón 1989 :81. Semiotics is, of course, a direct outgrowth of Saussurean semiology: Saussure’s emphasis on language as ‘a system of arbitrary signs’ and his proposal that linguistics is the central part of an overall science of sign relations or ‘sémiologie’ have led to the development of a field of inquiry more frequently called ‘semiotics’ following C. S. Peirce’s terminology, which deals with sign systems in literature and other forms of art, including music and architecture. Koerner 1994 :3663 For many linguists, most of whom were not well acquainted with the details of Shannon’s theory, Shannon’s model seemed to fit well with the Saussurean view concerning language and the sign-meaning relationship. Languages were considered comparable to Shannon’s codes and the model’s mechanical conversion of a message into an electronic signal seemed to fit well with the classical semiotic notion of signs. While John Lyons does seem to have understood information theory better than many others, he nevertheless offers a classic example demonstrating how these con- stituents are integrated see Lyons 1977 , esp. page 36. In a discussion titled “The Semiotic Point of View” in his Language and Linguistics 1981 , Lyons introduces the semiotic tradition, then casts semiotics in information-theoretic terminology: Semiotics has been variously described: as the science of signs, of symbolic behaviour or of communication-systems. … For present purposes, we will think of semiotics as having to do with communication-systems; … There are certain concepts relevant to the investigation of all communication-systems, human and non-human, natural and artificial. A signal is transmitted from a sender to a receiver or group of receivers along a channel of communication. The signal will have a particular form and will convey a particular meaning or message. The connection between the form of the signal and its meaning is established by what in a rather general sense of the term is commonly referred to in semiotics as the code: the message is encoded by the sender and decoded by the receiver. Looked at from this point of view, natural languages are codes, and they may be compared with other codes in all sorts of ways …. The problem lies in deciding what prop- erties of the codes, or of the communication-systems in which they operate, are significant for the purpose of comparison and what properties are either insignificant or of less importance. Lyons 1981 :17–18 The integration of these constituent models has become commonplace within lin- guistics—so much so that most linguists no longer consider it necessary to reference Shannon 1948 or Shannon and Weaver 1949 , even when reproducing versions of the diagram initially published in Shannon 1948 see again figure 3.6 . Undoubtedly many of the linguists employing the code model are even unaware of where the diagram and analogy originate. Nor are they aware of how Shannon’s original theory differs from their misapplications of that theory via the code model. 3. The Code Model Decoded 81

3.4. Code model presuppositions