Page 94 - Encyclopedia of Philosophy of Language

Language and Mind
thing about our mental lives must have a publicly accessible semantics: anythingwhichfixesin what way it says things, or a thing, to be—any facts relevant to determining whether it says things to be this way or that, or whether this or that would count as being as they are said to be—must be facts a multitudemight, in principle, recognize to hold, and on which they might base their judgments. Crudely put, if words (about X's mental life) say X to be F, there must be a publicly observable, or specifiable, state of affairs of which one could truly say, 'This is (what we call) (X's) being F.' (This point is as much about thought as about language. To some, it has smacked of behavi- orism. Whether it is that depends on what one allows to fall within the range of the observable. For Witt- genstein, it certainly was not that.)
The general problem about our mental lives is this. There is a familiar system of concepts under which aspects of our mental lives fall. For each of us, there is also our own stream of consciousness—all those experiences we are aware of by, or in, having them. What is the relation between the two? Different things in different cases, no doubt. The private language discussion, viewed as aimed at this problem, shows one thing the relation could not be. To fit one of those concepts could not be for an element, or several, in one's stream to be thus and so, where what counts as being thus and so is only fixed given other facts about that stream which only its subject could see to hold.
4. Language and a Background of Shared Reactions
The private language discussion does not only address the problem of mental life. It has a longer reach. Note that Sect. 243 is preceded by a brief discussion of language and thought in general. The point is that intelligible language or thought rests on a certain background of agreement in judgment. (Elsewhere Wittgenstein refers to systems of natural reactions.) The kind of dependence involved is illustrated in Sect. 142: "The procedure of putting a lump of cheese on a balance and fixing the price by the turn of the scale would lose its point if it frequently happened for such lumps to suddenly grow or shrink for no obvious reason.' Similarly, the procedure of saying something to be thus and so—to fit, or not, a given concept, or to be, or not, as it is said to be in given words— would lose its point if there were no regularity in our reactions in taking things to be that way or not; if we could not rely on what informed and reasonable peo- ple would do when called upon to judge such matters. Were 'reasonable' people as mercurial in their reac- tions in, say, taking things to be red or not, as Wittgen- stein's cheeses, we would lose a range of facts as to where things are reasonably taken to be red, and with that, the point of speaking of things as red or not. (Saying that a cheese weighs a kilo is not saying that it will not grow or shrink. Nor is calling something red saying that most people would say so.)
Given philosophy's penchant for making one thing out to be another, this idea that an activity makes sense only against some backgrounds—so that only against a certain background does it make sense to suppose given words to say anything at all—has sug- gested to some that the upshot of the private language discussion is a 'community view' of how (public) language is possible: for words W to have been spoken correctly, so for them to have said things to be as they are, is for their speaker to be in step with what the rest of his linguistic community would say. He says that his car is red; so would they. Speaking correctly
just means not falling out of line with the community. However, that is the wrong way to see a background as functioning. It serves its purpose only if it remains in the background.
5. Formulating a Private Language Argument
The question, then, is whether some broader and more original point about language is made by the private language discussion. One way to press this question is by pressing on features of private language that have seemed to play a role thus far. For example, in the Cartesian model, private words applied to types or classes of private 'objects'—the elements of one per- son's stream. Must there be private objects for private language? Consider the made-up word gronch. It applies, let us suppose, to some vases, doorknobs, drapes, and turtles, but not to others. Only I, though, am able to discern what something's being gronch requires. In principle, only I am ever in a position to make fully informed and authoritative judgments on such matters. Is that private language? Or suppose a 'private' language did not belong to just one person, but, say, to 10. Might that still be private language? Textual evidence suggests a 'Yes' answer both times (see, e.g., Sect. 207 and Sect. 237).
To press further, ask what a private language argu- ment might be. Here is the first of two suggestions. The private linguist, Pol, examines some thing or situation, and judges that 'F ' is true of it, F being some private term. (It matters not in the least whether the examined thing is private.) But how does Pol know he has not made a mistake? Perhaps he misremembers what is involved in being F, or he isjust bad at distinguishing Fs from Gs. If what it is to be F is anything one could remember, then this seems a possibility. But if it is, Pol cannot check that he has not made such an error, except in ways that let just the same sort of doubt creep in again. So, since there are always doubts he cannot settle, it seems that he cannot ever know whether F is true of a thing or not. But if he cannot know this, no one can, F being private. Language no one could ever know to apply to anything is no real language at all. So there can be no private language. (What is important here is the unavailability to others of at least some of the facts which determine how F is
72