Find Enlightenment

What is a Language Game?
by Bryan Register

Date: 1 Aug 98
Forum: University of Texas at Austin
Copyright: Bryan Register

Note: The author may or may not still agree with the views expressed in this paper.

Ordinarily, we try to give a Socratic answer - what is the essence of a language-game - or an Aristotelian answer - what are the genus and differentia of 'language-game'. But to give a Wittgensteinian answer is more difficult. In this paper, I will cast about aimlessly trying for a way to answer the question, and then, having figured out how, see if I can actually answer it.

Wittgenstein distinguishes between words which do and words which do not have an extension. Some words are names for objects, but not all are. At §27, for instance, he asks us if we are willing to say that certain words are names of objects. What would be, for instance, a No!? Clearly there are no No!'s in the world; the word lacks an extension.

But even words which are of the proper kind to have an extension (nouns, verbs, perhaps adjectives and adverbs) do not have rigorous criteria for inclusion within the extension. Wittgenstein justifies this by reference to the word 'game' in §66. Wittgenstein, looking without thinking, cannot find something which all games have in common and concludes that all of the members of the extension of 'game' share only a family resemblance. Just as I share my uncle's chin but my father's eyes, Risk shares the multi-player aspect of Diplomacy but the random element of poker. Likewise, presumably, for language-games. One language-game is like each of two others, which are themselves basically unlike.

If one can't know the criterion for extension-inclusion of a word (because it doesn't have one), what counts as knowing the word? Wittgenstein answers in §6. He is discussing the language-game of §2 (which has four possible moves: Block!, Pillar!, Slab!, and Beam!) when he asks "Don't you understand the call 'Slab!' if you act on it in such-and-such a way?" Moreover, since "...the ostensive teaching helped to bring this about" we can assume that 'such-and-such a way' is the right way. If we understand Slab! if we carry a slab to the correct place, then perhaps one can see that we understand 'explaining a language-game' if we explain a language-game. As you can see, casting about aimlessly tends to get us nowhere; to prove that we understand 'language-game' we must exhibit 'understanding language-game' behavior, which is just what we are trying to figure out how to do.

Let us take a different pass. Let us look at Wittgenstein's use of the word. On the assumption that he is exhibiting 'understanding language-game' behavior, mimicking Wittgenstein will prove that we, too, understand 'language game'. Wittgenstein shows two very interesting instances of understanding 'language-game'. At §7, he says the following: "We can also think of the whole process of using words in (2) as one of those games by means of which children learn their native language. I will call these games 'language-games'..." Moreover, "...and the process of naming the stones [slabs, etc.] and repeating words after someone might also be called language-games." And finally, "I shall also call the whole, consisting of language and the actions into which it is woven, the 'language-game'." Thus, some language-games are the quasi-games by which children learn language, and the language-game as a whole is everything linguistic or involved with the linguistic.

The second display of mastery of 'language-game' appears at §23, where Wittgenstein asks us to "review the multiplicity of language-games" in a long series of examples. One cannot say whether every example is an example of its own kind of language-game, or whether some of them are examples of subkinds of language-games into which more than one of the examples would fit. But one way or another, each of the examples is either a language-game or part of one.

Now, the process of learning a language is a language-game. So is every individual act within language. So is the whole of language and that with which it is involved. Now, all human activity is 'woven' into language. Further, all human artifacts are named, understood, and tied up with activity. Yet further, many places on the earth hve at some time been used for some purpose. (This beach is an invasion-place, this river a trade-route, this mountain a climber's-goal, this scene an artist's-reputation.) Even the sun, moon, and some planets have grounded calendar-use and been the targets of space probes. And of course, all human beings engage in activity which is woven into language (and one might even say that life is an activity). Only a few things are not part of the language-game: places no one has ever gone, stars which are only catalogued but are not part of the zodiac or an important constellation, and so forth. Everything else is part of the language-game.

Being aimless before got us nowhere, but now it seems to have gotten us too far. Surely Wittgenstein would not wish to say that Proxima Centauri is a move in a language-game? Perhaps we have tripped and made a false move in our explication-game.

Indeed we have. There is a difference between the language-game and a language-game. 'The language-game' is a proper noun, the name of a particular thing which just happens to be most everything. This thing is not something which it is of any help to name (and if we must name it, we could at least be traditional and call it 'being'). 'A language-game', however, is some specific form of communicative activity. At §23, Wittgenstein asks " many kinds of sentences are there?... countless kinds... And this multiplicity is not something fixed... but new types of language, new language-games, as we may say, come into existence..." There are countless kinds of sentences . It is this countlessness which is "this multiplicity". But when "this multiplicity" is multiplied, what we have added are "new types of language", or, "as we may say", "new language-games". Thus a language-game is a kind of sentence.

But Wittgenstein wants to expand the concept. Thus in his list of examples, he includes not only giving an order, but obeying it; not only describing an object but drawing it; not only forming an hypothesis but also testing it. His list is overblown. When you give an order, you use words. When I follow it, I use guns. When you describe an object, you use words. When I draw according to your description, I use pencils. When you form an hypothesis, you use words. When I test it, I use scientific apparatus. The resemblance between an hypothesis and its test is a radically different kind of resemblance from the one between asking a question and answering one. The only family resemblance between mixing a chemical solution and asking someone how to say something is the family resemblance that I share with an unrelated stranger of radically divergent ethnic background. We are all in the 'family of man', perhaps, but this tells us very little; all acts could be called 'language-games' but only uselessly so. Let us therefore return to the clear notion with which I ended the last paragraph: a language-game is a kind of sentence.

But what of Wittgenstein's larger language-games, for instance the language-game with four workmen's orders? If all our different orders (different sentences) are part of the one game, then a language-game is either a kind of sentence or something else. Since the four-order language-game is in fact a primitive language, a language-game is either a sentence or an entire language. And what of some of Wittgenstein's other (at least linguistic) examples in §23, such as 'making up a story', which is not some particular sentence but rather making a great many sentences? A language-game is: a sentence, a group of sentences, or a language. (One is tempted through frustration to say that there is in fact a very clear criterion for the use of the word 'language-game': Anything Wittgenstein wants to call a language-game is a language-game.)

Particular sentences and particular groups of sentences can show family resemblance. For instance, a one-line joke and a long joke are both jokes because they are both intended at humor. A single statement and a news story are similar in that they both try to assert facts. But the integrating factor here seems to be that both individual sentences and longer groups of sentences can be similar because they share the same point or purpose.

But it seems silly to say that a language is a language-game. Note that Wittgenstein's examples - the stone game, the colored square game - are examples of languages in which only one kind of activity can be performed. In the stone game, one can give only orders. In the colored square game, one can only describe arrays of squares. Additional rules to make it possible to ask what color the squares are, command that the squared be made a certain color, let someone know what color the squared are, complain about the color of the squares, and so forth, are not present. The fact that Wittgenstein's language-games are so simple leads us to distingish them from languages, which contain the apparatus for multiple kinds of speech acts.

A language-game is then a kind of sentence or group of sentences. Any particular kind is integrated by its linguistic goal. (Note that when we carve off Wittgenstein's errors, we move toward Searle's account of illocutionary acts and their integration in accordance with illocutionary point.)

I have done some mimicry now to exhibit 'understanding language-game' behavior, and I have even done some criticism to prove that it was more than mimicry. But one final test will be to show a kind of language-game not discussed by Wittgenstein - it is easier to carve off half of his examples than to add one of my own. Thus let me mention that getting married is a language-game.

When we get married, we say a few words which create the marriage. These words, called 'vows', create commitment and obligation. (If a marriage is created by making a promise, then a marriage, perhaps, is a promise.) What Wittgenstein says at §64 holds true of marriage vows. He discusses here a different way to communicate the information of language game 48 which gets the same point across - instead of specifying what color a square is, we will now specify what color each half of a rectangle is. Likewise, there are many ways to get married. One can have someone say 'By the power vested in me by the State of Texas...', or 'By the power vested in me by God...' The reason to prefer the one over the other is not linguistic but theological and so as language-games the two are equally good, depending on extra-linguistic purpose. Getting married also shows a more than family resemblance to Wittgenstein's example of giving orders - a promise is a little like an order given to oneself. Besides, making a promise should have been on the list of examples.

Find Enlightenment at