Early in the Philosophical Investigations Wittgenstein describes a simple language game:
Let us imagine a language for which the description given by Augustine is right. The language is meant to serve for communication between a builder A and an assistant B. A is building with building-stones: there are blocks, pillars, slabs and beams. B has to pass the stones, and that in the order in which A needs them. For this purpose they use a language consisting of the words ``block,'' ``pillar,'' ``slab,'' ``beam.'' A calls them out;--B brings the stone which he has learnt to bring at such-and-such a call..."
We can say, then, that in the builder's language the terms ``block,'' ``pillar,'' ``slab,'' and ``beam'' stand for kinds of building stones, and that these kinds are the meanings of the words. This is what we would say looking at the language from the point of view of Augustine's picture, and, for this language game, Augustine's picture is correct.
The important point is that it is only because the words have a use in a web of activities -- the orders given by the builder and the executions of those orders by the assistant -- that the words have these meanings.
When the builder says ``Slab,'' he performs a certain type of act, with certain success conditions. The success conditions are that the assistant passes a slab to him. The assistant knows the meanings of the words, if he performs the right actions when the words are given. One point Wittgenstein makes is that it is these actions are not images that may or may not appear in the assistant's mind that determine whether or not he has learned the language and knows the meanings of the words.
The point of the four words, then, seems to be to provide the builder with four different actions. Each of these types of actions will have success conditions: the assistant passes a slab, the assistant passes a block, etc. The actions are the same, except for the word the builder uses, and the success conditions are the same, except for the type of stone the assistant is to pass. Our remark that ``Slab'' stands for or means ``slab'' and so forth just seem to amount to describing the effect the utterance of each of these words has on the success conditions of the acts of which that utterance is a part.
In remark 6 Wittgenstein provides a nice metaphor for this point:
``I set the brake up by connecting up rod and lever.'' -- Yes, given the whole of the rest of the mechanism. Only in conjunction with that is it a brake-lever, and separated from its support it is not even a lever; it may be anything, or nothing.
In the case of the builder's language game, what is established is that the assistant will pass a building stone to the builder. If there were only one kind of building stone, that would be that. But since there are four kinds of stones, a question still remains once it is established that the builder needs a stone. The language, the choice given by the four words, provides a way of resolving this question. The role of the words has to be understood as the resolving of these questions. To say that ``slab'' stands for the slab and so forth is not incorrect, but it needs to be understood in the context of the incremental role the words make to the language game.
Now consider a more or less formal analysis of the builder's language game. For all we have said so far, it seems it might go like this:
(A) Analysis of Builder's Language Game
Analysis (A) describes the builders language game as consisting of utterances of nouns. There is no mention of sentences. The nouns in BL do not occur as parts of sentences, and their meanings do not derive from the meanings of sentences, or their use from the use of sentences.
There is a direct connection here between words and extra-linguistic reality. The words are connected to the intentions and goals of the builder. The semantic facts listed in (A3) are not basic facts, and they are not based on connections between words and images. They are facts based on the role that individual words play in the articulation of a family of types of actions that are connected with the intentions and goals of the speaker.
The argument at the end of the last section is simply fallacious. To say that the semantic facts about words derive from the semantic facts about the commands of which they are a part is not to say that these facts derive from semantic facts about sentences of which they are a part. Words can play a role in the articulation of commands without being parts of sentences, and so can be connected with human goals and intentions without the mediation of sentences.
The idea that sentences are semantically basic would be brought to bear on the builders' language game as follows. The builder has the desire that the assistant pass him a slab, for example. The content of this desire is what gets communicated to the assistant. This is shown by the fact that the assistant, when he understands, then sets out to do something, and he can do either the right thing or the wrong thing. The utterance has compliance conditions, and the compliance conditions accord with the satisfaction conditions of the desire that motivated the utterance.
This seems to show that to understand the utterance, the assistant needs to understand grasp the content, that I pass him a slab. This is what he needs to make true in order to comply with the command, and to understand the command is to understand what one needs to do to comply with it. It seems then that the content of the builder's utterance is not just slab, the thing, but the proposition, that the assistant bring a slab to the builder. And this shows that ``Slab'' really isn't just a noun, standing for a kind of object, but a one word sentence, expressing a desire.
I think this argument confuses two quite different things, however. It is quite right that the content of the builder's command is the proposition that the assistant pass him a slab. But what is the contribution of the word ``Slab'' to the fact that the command has that interpretation? It is clearly the kind of object that is to be fetched.