Conventions of language
If we say that understanding a sentence is knowing what it means, then it's natural to think that someone who knows what the sentence S means knows that S means M, where M is some specification of the meaning.
Thus, on the Fregean view, to know the meaning of a sentence is to know its sense, which is to know the conditions under which it would be true. So you know what the German sentence “Es regnet” means if you know that the sentence would be true just in case it was raining. So you have a belief that captures your knowledge of the meaning of the sentence: you believe that “Es regnet” is true just in case it's raining.I used a German sentence here not out of homage to Frege, but because if I'd used an English one, it might have seemed vacuous. I'd have said that you know what “It's raining” means if you knew that that very sentence was true just in case it was raining. But the fact is that this wouldn't be vacuous at all. Of course, the only way I can specify the truth conditions of “It's raining” in English is to use English words that mean the same as “It's raining.” And you will need to understand English to understand that specification. But knowing that a sentence S would be true just in case it was raining is something that you can know without knowing English. Frege knew this about the German sentence “Es regnet.” What that makes plain, I think, is that your specification of the meaning of the sentence in your own mind can't be in English. Let's suppose, then, you have an internal language: the language of thought (as the title of a book by the American philosopher of language Jerry Fodor called it).
So your knowledge of the truth conditions of S must be specified in the translation of S into the language of thought.
Let's call that translation “internal S.” It can't be that your understanding of the language of thought consists in knowing the truth-conditions of internal S, for these would have to be translated further into some other language, and we would be on our way to an infinite regress. What makes it true that internal S allows you to specify the truth conditions of S is that the right connection obtains between internal S and the states of the world that obtain when it is true. That is, the relationships between our mental states and things in the world that give the states their contents—that make, for example, our beliefs have truth conditions—must be different in kind from the relations between language and mental states that give sentence their truth conditions.This is, in fact, just what you would expect. If I currently have a belief that has the content it’s raining, then, following on our discussions in Chapter 1, we would expect this to be a consequence of the functional role of the belief: the fact that it's the sort of state that's produced in me when I look out of the window and see rain pouring from the heavens and that makes me take my umbrella to avoid getting wet. On the other hand, what makes the sentence “It's raining” have the content it does is, presumably, the fact that this is a convention of the English language. If the conventions of English had been like those of German, for example, then the right sentence to express that content would be “Es regnet” instead.
If we remember Frege's discovery of the primacy of the sentence, we shall want to ask, first, what the convention is that gives sentences their truth conditions and their truth values. The natural answer is that the convention that gives sentences truth conditions is the convention that you should use declarative sentences to say what is true. The philosopher H. P. Grice has proposed that what this amounts to is that we all expect someone who understands a sentence S to use S to get other people to believe that S is true.
So now we see why there's an intimate connection between the contents—the truth conditions—of beliefs and of sentences. A sentence is a conventional means of trying to get other people to have a particular belief: the belief with the same truth conditions as the sentence. Starting with this basic idea, you can go on to look at sentences that are not declarative, among them orders and questions.In using declarative sentences we make assertions. In using imperative sentences, we give orders. In each case we are producing a complete meaningful utterance, we are performing what linguists call a “speech act,” though the particular types of speech act differ in their particular functions. Despite these differences, however, we can use the ideas of Frege's semantic theory—the ideas that we used to explain the utterance of declarative sentences in the speech act of assertion—to explain the contents of other speech acts as well. For every one of the central speech acts—assertion, questioning, and ordering—Frege's idea of the truth condition can be used.
We say that the truth conditions of a declarative sentence hold if the sentence is true. In assertion we try to get others to believe that the truth conditions of the sentence we assert hold; in ordering we try to get someone to make the truth conditions hold; in questioning we try to get someone to tell us what truth conditions hold. In Chapter 5, I shall look in more detail at orders in the context of discussing the philosophical theory about moral language that is called prescriptivism.
So Grice's theory tells us what it is to understand sentences as we use them in these many speech acts. To know the meaning of a sentence is to understand how it is used in speech acts. But if we combine it with Frege's discovery of the primacy of the sentence, Grice's theory can also tell us what it is to understand the meanings of words.
To understand the meaning of a word is to know how it contributes to determining the meanings of sentences. So to understand a word, W, is to know how it contributes to fixing what speech acts you can carry out with sentences that contain W.Notice that Frege's theory and Grice's are thus not inconsistent with each other. In fact, they are really complementary. Frege's theory says we have to know the sense of a word to understand it, and that knowing the sense of a word just is knowing how it determines the sense of a sentence. To know the sense of a sentence is to know what it would be for it to be true. But that is precisely what you have to know on Grice's theory. For if you know the truth conditions of a sentence you know which belief people using it are trying to communicate: namely, the belief with the same truth conditions. You could say that Frege tells us what the meanings of sentences are— namely, truth conditions—and Grice tells us what the truth conditions are for.
Thus, on Grice's theory, someone understands “It is raining” if she both
a) uses those words to try to get people to believe that the truth conditions of “It is raining” hold and
b) expects people to use those words to try to get her (and others) to believe those truth conditions hold also.
And, of course, to believe that the truth conditions of “It is raining” hold is just to believe that it is raining. As far as orders are concerned, you understand the command “Peel me a grape!” if you both
a) use those words to try to get people to make that sentence's truth conditions hold and also
b) expect people to use them to try to get you (and others) to make them hold.
To make those truth conditions hold, of course, is just to peel the speaker a grape.
3.13