Wittgenstein introduces a series of different language games, which are imagined languages that are complete, yet much simpler than our own. Language game one consists of A, a builder, shouting orders to B, his assistant. There are four orders in the language: "cube!" "brick!" "slab!" and "column!" After hearing the order, B brings the object A called for. Children can learn this language by means of demonstrative teaching, in which an elder points to each object as he names it.
Does "brick!" in this language have the same meaning as "brick" in our language? We have varied uses for "brick," but we can also shout "brick!" as an order, as they do in the first language game. But in our language, this order is not quite the same as "bring me a brick" and other such sentences, whereas in game one there is no difference. Even if we mean exactly what A means when we shout "brick!", the languages are still different. The different meaning "brick!" has in our language is a result of the system of language surrounding it, not in the mental state of the person speaking.
In language games two to five, Wittgenstein adds numbers, proper names, "this," "there," and the question "how many?" That these elements of language are totally different is clear from the different ways we learn them. Demonstrative teaching can point to the number five (five slabs, five pebbles, five fingers, etc.) just as it can point to a slab. The difference between kinds of words is not in the act of pointing, nor in a mental state, but in how these words are incorporated into the language.
In game nine, A calls "slab! column! brick!" meaning he wants B to fetch those objects in that order. In game ten, he calls "column second! slab first! brick third!" Wittgenstein shows that words and word order can play the same role.
A sentence or proposition is a complete sign. A word is a component of a sentence. Wittgenstein contrasts words with patterns, saying patterns such as pictures or gestures represent the object they denote, whereas words bear no outward resemblance to what they denote. It is not always clear what is a word and what is a pattern.
Suppose A shows B a sample and tells him to fetch a piece of cloth of the same color. There is a wide range of things B could do, from simply walking up to a shelf and grabbing a piece of cloth, to conscientiously placing the sample against each piece of cloth until he finds one whose color is indistinguishable from the sample. There is not one essential or characteristic act of comparing or recognizing, but a whole number of activities that share certain similarities.
We may feel uncomfortable accepting the idea that a man can simply walk in and take a piece of cloth off the shelf without any further act of comparison. How can he recognize that as the correct piece of cloth? No mental state that we can imagine will help make this process any less mysterious.
The Brown Book draws heavily on the idea of language games. Wittgenstein's use of language games in the Brown Book is significantly different from his use of language games in the Blue Book. In the Blue Book, language games are seen as a means of looking at the elements of our language in a stripped-down form. In the Brown Book, Wittgenstein recognizes that the use of the words in simpler language games is different from their use in ordinary language, and that these language games ought to be thought of as languages complete in themselves.
Wittgenstein uses language games to highlight and dispel certain prejudices we have concerning language. It is not that we cling to rigorously honed but false theories of language, it just that we are inclined to think about language in a certain way. At the opening to the Brown Book, Wittgenstein alludes to St. Augustine's description of learning language (an expanded version of which will also be the opening of the Philosophical Investigations) as only accounting for solid, material objects. By this, he does not mean to suggest that Augustine has developed a complex theory of language that leaves out everything but common nouns, but rather that Augustine has succumbed to the temptation to think about language in a particular way—in this case, to think about language as referring only to solid objects. While a more complex theory of language will refine what Augustine has said, it might unconsciously retain some of his basic assumptions.
The first language game is true to Augustine's account: it gives us a language in which every word denotes an object. Wittgenstein uses this example to dispel a further myth that says the meanings of words are fixed by a mental act. First, this example shows us that the mental life of A and B does not even enter into a description of how the language is learned, let alone used. Second, we see that the equivalent words in our more complex language differ from these words not because of our mental state, but because of the system of language surrounding those words. If we wish to argue that "brick" in our language means something different than "brick!" in this language game, that difference has to do with the context of the word, and not with something inherent in the word itself. That is, while "brick!" may simply be the name of an object in the language game, we now see that our word "brick" cannot be adequately defined as simply the name of an object. If it could be, we would be living in the world of the first language game. In order to understand "brick," we have to understand how it is used, how it fits into a larger language.
Part of the problem with saying that words are names of objects is that such a theory creates the illusion that we learn language one word at a time. To say that "brick" is simply the name of an object presupposes that we already understand that the word is a noun, that this is how nouns can be used in our language, and that they can fit in certain ways into certain kinds of sentences. Augustine's account of learning language describes something like language game one, and not ordinary language, because he tells us only how individual words are learned and says nothing about learning the grammatical structure of language.
The second language game brings out another fundamental flaw in Augustine's picture: he assumes that words play roughly similar roles. In examining how number words are learned and then used, Wittgenstein shows us that they are of an entirely different kind than object words. Color words are completely separate from number words and object words. Because of the fact that they are all words, we might conclude that they are all signify the same kind of thing. We can also learn them all by means of ostensive definition. By pointing, we can say "that's a slab" or "that's five" or "that's red." There is no difference in how we point, or even necessarily in what we think while pointing. The difference is in how we subsequently use the words.
Having shown us that different words do different things, Wittgenstein shows that gestures and word order can often do for us what words do. Not do different kinds of words share nothing in common, but there is nothing about words that makes them distinctive from other forms of communication. Wittgenstein uses these and other examples to break down our idea that there is an "essence" to language. In the language games, words are used in all sorts of different ways to achieve all sorts of different ends. If we try to say that all words share something in common, or that all words are learned in the same way, we are simply generalizing about a particular case that has caught our attention.