Does the language I speak have any normative force on how I go about saying what I mean? Is there such a thing as the set of words I should use to say what I mean? Maybe there is no one set of the right words, but we could make a weaker claim: for anything you want to say, there are some words that cannot be used to say that. Not everything can be said in any way.
But what right do we have to make even this weaker claim? Isn’t it true that new uses for language are being found all the time, and that even nonstandard uses of words still communicate? Given the interpretive agility that mature speakers of a language exhibit, perhaps we should say that, though there are ways of saying what you mean that are the most common (in your area or in your community) and therefore constitute the easiest way to say what you mean, you could say what you mean in any number of other ways.
David Wiggins disagrees:
…even to an audience that knows no better, you cannot utter the sentence ‘the sonatas are just the anecdote to a gloomy summer’s day’ and count as having said that the sonatas are an antidote to such a day, or utter the sentence ‘a wind machine is used to emulate breathing at the start and finish of the work,’ and count as having said that a wind machine is used there to imitate or simulate breathing. That will only become possible when the day dawns on which so many people use these words in this way, and they use them thus so deliberately, and they are so multiply and densely defended and reinforced in these choices of words, that that is what these words mean in the language. (“Languages as Social Objects”, p. 523)
Here Wiggins argues that you do not mean what you say or say what you mean when you utter a malapropism. We may be able to see what a speaker does mean when she makes such a mistake, but that does not mean that ‘the sonatas are just the anecdote’ is a way to say what you mean when you mean that the sonatas are just the antidote. Some people talk in a way upon which Fowler would frown, but we can still puzzle out what they’re trying to say. Only, we should not let this fact convince us that trying to say amounts to saying.
Why is this not just dogmatic insistence that one use of language (ours, conveniently) is the right one? Wiggins grants that speaking in the right way is not a matter of measuring up to some timeless Right Way that comes attached to a language.
…it has long been common ground between all parties that there is no anhistorical, unconditioned standard that a speaker needs to have observed if he or she is to count in the way he or she wishes to count as having said that so and so. It is not a question of how, starting from scratch, the speaker is to speak in order to perform a certain linguistic act. It is a question of how the speaker is to proceed given what has been done so far, what has worked so far and what has been enshrined so far in the language. Sometimes this is a question of what will work, or what the speaker or writer can cause to work. […] But mostly it is a question of what does count by the (however essentially contestable) standard of the language itself. (ibid., p. 522)
But the lack of a timeless standard for saying what you mean can seem to level the playing field in such a way that anything can be used to say anything. How do we avoid this conclusion? We’re stuck oscillating between rigidity and lawlessness, between linguistic puritanism and a chaotic libertinism.
Wiggins’s way out begins with asking what language is for. To a mature speaker of a language, the language exists as a tool, as a tool that makes itself available to be used to accomplish the speaker’s purposes. So a speaker’s intention when she chooses her words is “not the intention to uphold a regularity but the intention to exploit an existing resource” (ibid., p. 513). We use the word “antidote” (not “anecdote”) to speak of antidotes not out of sheer concern for conformity, but because we know that we can do what we want to do by using that word. Wiggins spells out this feature of language by pointing out that our primary purpose in speaking is to go on record as saying or having said that such-and-such (where saying or having said that accomplishes some purpose). Going on record is a kind of public event, and it is useful to have publicly accessible ways of making that event happen. Saying what you mean is supposed to be open-access, and the only way for it to be that way is for there to be a public way of doing so as well as a public way of failing to do so.
For all this, we can still understand a person when she misspeaks. We know what she is trying to say when she says that the sonata would be the perfect anecdote. Is this ability a counterexample to Wiggins’s view? Going beyond Wiggins, perhaps it is right to suggest that we possess this ability not in spite of the normative force of our language, but because of it. Picking the right tool is hard, given the number of tools in our collective linguistic toolbox; no one gets it right all the time. One way to adjust to the possibility of failure is to . When another person reaches for the wrong tool, we politely insert the right tool into their hand. After all, we ourselves have experienced disconnection between what we say and what we mean; we know what the failure of malapropism is like from the inside. So we know what is going on when another person misspeaks because we know what is going on when we ourselves misspeak. We can see what a speaker really means because we know what it is to make this kind of linguistic mistake, and we know what it is to correct such mistakes (and be corrected ourselves).