The Blue Book opens with the question, "what is the meaning of a word?" When asking such general questions, we often define words by thinking of of solid, material objects, like pencils, chairs, and tables. These words can be defined ostensively, by pointing to the object they denote. We might then be tempted to think that the meaning of these words is the mental act of interpretation that connects the word with the thing it denotes. Wittgenstein asserts that, contrary to conventional wisdom, the meaning of a word is determined by its use in language. If these mental acts that supposedly determine meaning are simply a matter of operating with signs, we could just as well say that thinking and meaning can be done on paper, or with the voice box.
Wittgenstein introduces the idea of a language game, a more primitive form of language that helps us highlight certain features of our own language. We can construct a variety of different language games, each with distinctive features. Philosophers usually shun this approach out of what Wittgenstein calls a "craving for generality": they want to discuss the general features of language rather than its particulars. This craving encourages the misconception that every word has a single, fixed meaning.
We are tempted to think that a spoken word needs interpretation (we need to be told what it means), but that its meaning does not require any interpretation. This is false, just as it is false to think that while we need to follow a particular rule to behave in a particular way, we do not need a further rule to interpret that rule. There is no clear reason why every word must be connected to its meaning in our mind. Words are not inherently related to the things they denote.
Metaphysical claims often try to make general statements about the nature of things, like "everything is in flux." But the words "in flux" can only have meaning when contrasted with their opposite, "stable." If we say that everything is in flux, the term "in flux" loses its meaning. Wittgenstein talks about solipsism, the view that "only I exist." The solipsist encounters a problem in trying to make overall claims about the nature of experience. If she claims "only what I see is really seen," she empties the word "see" of meaning, because she uses the word incorrectly. Such a claim can only amount to an appeal for a new notation, a redefinition of the word "see." It cannot state any metaphysical discoveries about the nature of experience.
The first part of the Brown Book consists of a series of language games. Wittgenstein uses these games to highlight the different forms of expression, and to point out that although we may think all words share something in common because they can all be expressed as written signs, there actually have very little in common. For instance, the words "chair," "one," and "this," are not similar.
Throughout the Brown Book, Wittgenstein examines words like "recognize," "compare," "believe," "read," "understand," and so on, to show that there is no common feature of all the different uses of these words. Rather, there is a family resemblance. Certain uses of a word may share certain features with others, just as members of the same family might have certain features in common. This line of reasoning shows us that these words have no single fixed meaning, but only a number of loosely related uses.
If we accept that individual orders must be interpreted according to some sort of rule, we must also accept that rules themselves need to be interpreted. For instance, if in reading items from a printed table, I need to understand the rule that we correspond different columns by reading left to right, I may also need a rule to tell me how to correspond different columns by reading left to right, and so on. Sometimes I can read off a table without referring to any general rule about how to read tables. That rule needn't always be present in my mind.
There are many uses of the word "can," some of which refer to what someone has done, some of which only refer to a potential for future action. We should not be misled by grammar into thinking that the present tense of "can" denotes a state of the person we are talking about. Similarly, we should not be misled by grammar into thinking of the past and the future as things that have passed or are yet to come, and then puzzle about where the past goes to. Wittgenstein's discussion of "can" also leads to some reflections of reading and on the expression, "I can go on," in both cases showing that there is no distinct process that is present in all uses of these expressions.
Part II of the Brown Book focuses primarily on the idea of seeing something as something else, and on the idea that there must be a feeling of similarity when we use the same word in two different contexts. It makes sense to see a bunch of squiggles as a face, but it does not make sense to see a pencil as a pencil, because there is no real alternative.
We should not think that there is a single, paradigmatic use of a word that all other uses are compared to. If I talk about one vowel being "darker" than another vowel, I needn't be comparing vowels to colors. Similarly, there need not be a paradigmatic "feeling" that is present whenever I mean or believe what I say. The meaning of a word is simply a matter of how we use it, and not a matter of identifying it with other objects or paradigmatic cases. There is not a standard outside of language that language must compare itself to.