January 26, 2005

"Everything is correct" versus "nothing is relevant"

On January 23 a user identified as Zink made some comments on ceejbot's blog about the Language Log post Nearly all strings of words are ungrammatical. They struck me as really interesting:

There's a funny bit in there where they try to at once claim to be "descriptivist, not prescriptivist" while at the same time decrying the word "are" in

Why do some teachers, parents and religious leaders feel that celebrating their religious observances in home and church are inadequate and deem it necessary to bring those practices into the public schools?

Sorry kids, you can't be an apple and an orange, and if you're a descriptivist, and someone honestly makes a sentence, that's an honest sentence in the language that actually is.

By "they" Zink means the Language Log staff (the post was actually mine; none of my colleagues need take responsibility for the views expressed here). What's so interesting is that it is quite clear Zink cannot see any possibility of a position other than two extremes: on the left, that all honest efforts at uttering sentences are ipso facto correct; and on the right, that rules of grammar have an authority that derives from something independent of what any users of the language actually do.

But there had better be a third position, because these two extreme ones are both utterly insane.

I actually devoted my presentation at the December 2004 Modern Language Association meeting to a detailed attempt at getting the relevant distinctions straight, after thinking them through with a great deal of help from a philosopher of linguistics, Barbara Scholz. The concepts are by no means easy to get a grip on. But let's make a start.

First, I didn't "decry" the form are in the quoted example from a letter published in the Philadelphia Inquirer). (Decrying is strong disapproval, open condemnation with intent to discredit; check your Webster.) It needs no strong public condemnation; it doesn't offend me. I merely said it was wrongly inflected. And I explained in painstaking detail why it couldn't satisfy the normal principles of English. Now, what are these things I'm calling the normal principles? Where do they come from?

Barbara Scholz and I have taken to using the term correctness conditions for whatever are the actual conditions on your expressions that make them the expressions of your language — and likewise for anyone else's language. If you typically say I ain't got no hammer to explain that you don't have a hammer, then the correctness conditions for your dialect probably include a condition classifying ain't as a negative auxiliary, and a condition specifying that indefinite noun phrases in negated clauses take negative determiners, and a condition specifying that the subject precedes the predicate, and so on. The expressions of your language are the ones that comply with all the correctness conditions that are the relevant ones for you.

Which conditions are the relevant ones for you is an empirical question. Descriptive linguists try to lay out a statement of what the conditions are for particular languages. And it is very important to note that the linguist can go wrong. A linguist can make a mistake in formulating correctness conditions. How would anyone know? Through a back and forth comparison between what the condition statements entail and what patterns are regularly observed in the use of the language by qualified speakers under conditions when they can be taken to be using their language without many errors (e.g., when they are sober, not too tired, not suffering from brain damage, have had a chance to review and edit what they said or wrote, etc.).

Sometimes, though, one can formulate the relevant correctness condition exactly right, and then observe a sentence in the Philadelphia Inquirer that does not comply with it. This is because people do make mistakes in their own language, and some mistakes even get past newspaper copy editors.

But by saying that, I'm not endorsing any right of descriptive linguists to be considered correct in their statements regardless of what people say! There's no contradiction here (though Zink thinks he sees one). One could imagine that there might be people who actually have different correctness conditions, so that the quoted sentence was grammatical for them. There could be people for whom tensed verbs agree with the nearest noun phrase to the left, for example. For such people, this would be grammatical (I mark it with ‘[*]’ to remind you that it's not grammatical in Standard English):

[*]Celebrating religious observances in home and church are inadequate.

In fact they would even find this grammatical (with the meaning "Celebrating birthdays is silly"):

[*]Celebrating birthdays are silly.

If they really did (one could check by interviewing them or recording them for a while), and if the letter in the Inquirer was written by one of them, I'd change my mind. I've made an empirical claim: I think the person who wrote the letter speaks the same language that I do, and would regard all three of the examples given so far as ungrammatical. I think the person just made a slip while writing, failing to keep in mind that they were writing a sentence in which be inadequate had a clause as its subject, and inflecting be as if it had observances as its subject, through a moment of inattention.

Zink thinks that if you're a descriptive linguist and "someone honestly makes a sentence, that's an honest sentence in the language that actually is." But this is not about honesty. It's about whether an occurring utterance matches the correctness conditions (whatever they may be) for the speaker who uttered it. Either speakers or linguists can be wrong. Speakers will sometimes speak or write in a way that exhibits errors (errors that they themselves would agree, if asked later, were just slip-ups); and linguists will sometimes state correctness conditions in a way that incorporates errors in what is claimed about the language (errors that they themselves would agree, if asked later, were just mistaken hypotheses about the language). I claimed that I'm right about the correctness conditions on verb agreement in Standard English, and that the person who wrote the letter I quoted made a slip-up. That's not a contradiction — no one is attempting to be both an apple and an orange.

And none of the foregoing has anything to do with prescriptive claims about grammar, which are a whole different story. Prescriptivists claim that there are certain rules which have authority over us even if they are not respected as correctness conditions in the ordinary usage of anybody. You can tell them, "All writers of English sometimes use pronouns that have genitive noun phrase determiners as antecedents; Shakespeare did; Churchill did; Queen Elizabeth does; you did in your last book, a dozen times" (see here and here for early Language Log posts on this); and they just say, "Well then, I must try even harder, because regardless of what anyone says or writes, the prohibition against genitive antecedents is valid and ought to be respected by all of us." To prescriptivists of this sort, there is just nothing you can say, because they do not acknowledge any circumstances under which they might conceivably find that they are wrong about the language. If they believe infinitives shouldn't be split, it won't matter if you can show that every user of English on the planet has used split infinitives, they'll still say that nonetheless it's just wrong. That's the opposite insanity to "anything that occurs is correct": it says "nothing that occurs is relevant". Both positions are completely nuts. But there is a rather more subtle position in the middle that isn't. That is the interesting and conceptually rather difficult truth that Zink does not perceive.

[You'll see that there's now lots more discussion available courtesy of ceejbot. There you can have the pleasure of seeing me described as "an abyssmal [sic] dunce" (for not believing that which is limited to supplementary relative clauses in Standard English). You'll read that I'm "a liar"; "smugly superior"; "muddled"; and someone who "thinks his judgement counts more than everyone else's". The strange thing about this kind of commentary is that while I stress (above) that it is entirely possible for a linguist to be wrong about what the correctness conditions on a language (even their own language) really are, the people calling me smug, stupid, and mendacious have no doubts whatsoever. They seem utterly convinced of their rectitude, as they angrily attribute to me the exact opposite of what I said. For example, you'll see that Scholz and I are directly accused (by a user called Nick) of holding that "correct" means "what happens". Our actual view is that we firmly and explicitly deny that, though we also resist the opposite lunacy, the position that what happens has no relevance to the determination of what's correct. As Mark pointed out to me when he first referred me to ceejbot, it's not just the the existence of ignorant authoritarian prescriptivism in this culture that needs an explanation, it's also the level of anger that accompanies its expression.]

Posted by Geoffrey K. Pullum at January 26, 2005 12:49 PM