Uncle Jazzbeau’s Gallimaufrey

language != communication

Well, I continued plowing on through chapter two in Ellis' Language, Thought, and Logic, and it's getting funnier. In this chapter he lists the three most common initial missteps made by those who theorize about language. Here's number one: "It is the assumption that the purpose of language is communication." So, he got my attention, but what does he mean?

The most common and simple version looks at language in terms of its information content: a piece of language means whatever information it contains, and an adequate analysis formulates as completely as possible what that information is. The variant that uses such words as message and encode is essentially similar. [...] A more philosophical version of essentially the same view speaks of the truth claims made by a piece of language.

[Ellis, p.16.]

I guess the question is whether a piece of language contains anything. Is a word a container? I'd always felt that a sign doesn't contain its meaning anymore than a tool contains its function. In fact, tools and words are often used against their functions or meanings. Pieces of language are interpreted, i.e., meaning is assigned to them by some hermeneutic process on the listener's end of the phonic chain.

Nothing seems more reasonable than the assumption that the purpose of language is communication. But there is a subtle trap here: granted , a particular act of language use may result in communication between two people, but much must have happened before they could get that far.

[Ellis, p.17.]

OK, now he's lost me, but it becomes clearer later on that Ellis feels that language divies up the world of its speakers and gives them a way of making sense of it, and it just happens that we can use that to communicate. I am reminded of Searle's US GI trying to convince Italian soldiers that he is a German officer by reciting a piece of Goethe, but perhaps I'm reading too much into it. Did speech precede language? Did thought precede language? But Ellis continues:

The distinction I am making will become clearer if we go back to the vocabulary used by scholars who think of language as the transmission of information, and therefore use the vocabulary of code, encoding, and message. This language certainly fits well with that of communication and information, the latter term substituting for message without difficultym while coding and encoding seem to describe well enough the process by which the message is conveyed from one person to another.

Nevertheless, as soon as we look closely at languages and codes it becomes clear that they are very different things, so much so that the linguistic situation is completely distorted if we we speak of it in this way. Codes are only devices for disguising pieces of language so that their meaning is not immediately recognizable without removal of the disguise. [...] What is in Morse is only inforSo, imative only because it is English. The surpising fact is the very vocabulary of code and message, used so often as a model of how language works, has nothing whatever to do with language.

[...]

The essential and distinctive feature of a language is not its ability to transmit information—for this would not distinguish it from Morse code—but a logically prior attribute, the process of analysis, evaluation, and organization of experience which must have taken place before communication can occur. It is because a code lacks this stage that it is not a language. To function at all, a code has to latch on to something that has already done what languages do. Codes encode messages; language do not. Codes merely transmit information; languages make information what it is.

[Ellis, p.17f.]

So is grammar a code? More to come later.