Language Log is written by professional linguists. Even the utterly silly bits (and lord knows I've done a few) are written by professional linguists taking time off from being the serious scholars they normally are. And linguists are often concerned to point out that many of the things ordinary folks imagine to be grammatical errors are in fact perfectly grammatical and acceptable. But linguists are not like indulgent parents lowering the bar so they can congratulate every child for jumping over it. There's definitely such a thing as a syntactic error, even in your native language, even as judged by descriptive linguists. Here is just one example, from a reader's letter that I noticed in a copy of the Philadelphia Inquirer (Friday, December 31, 2004, p. A260) that was lying around near the water cooler in one of the corridors at our headquarters, Language Log Plaza:
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?
"Letters may be edited for clarity, length and accuracy," it says on the letters page. But where are the copy editors when you need them? There are two present-tense verbs here, both inflected for plural agreement. One of the two is wrongly inflected.
The first is the auxiliary verb do. It has the form do because its subject is the plural noun phrase some teachers, parents and religious leaders. That is correct.
The second is a form of be. (Remember, I use bold italics when naming a lexeme.) It's just before the adjective inadequate. It has the form are. But that's a mistake. Perhaps the writer became momentarily confused and thought its subject was the plural noun phrase their religious observances. But it's not. The subject is a gerund-participial clause, celebrating their religious observances in home and church. Clauses count as singular. In fact you can make up a sentence in which either singular or plural agreement is possible with the same string of words as subject:
[1] Moving pianos is dangerous.
[2] Moving pianos are dangerous.
The first has singular agreement (is), so moving pianos has to be read as a clause: it's about the thing you're doing when you move a piano. The second has plural agreement (are), so moving pianos has to be read as a plural noun phrase: it's about loose pianos rolling around. This is possible because moving can function as an attributive modifier of the head noun pianos.
But celebrating their religious observances in home and church is a clause with celebrating as its verb, not a noun phrase with observances as its head. The word celebrating cannot possibly be an attributive modifier with observances, because their, a genitive pronoun, follows it; that cannot be anything but a determiner, and determiners precede attributive adjectives. (Genitive pronouns cannot serve as attributive modifiers: phrases like *the my house or *an our cat are utterly ungrammatical.)
So the subeditors on the Inquirer's letters page dropped the ball here. There's no possible way that sentence is grammatical.
So we linguists don't dismiss as puristic prescriptive nonsense all claims of some attested sentence being ungrammatical. Our aim, at least, is only to dismiss the ones that are puristic prescriptive nonsense.
And I'll tell you something else. It seems clear to me that nearly all strings of English words you can construct are ungrammatical. Try writing down any random sequence of words (a fully grammatical one if you want to bias things against my claim), either with repetitions or without, it doesn't matter. With a very few peculiar exceptions, for any string of words you will find that almost every one of the orders in which those words can be arranged will be ungrammatical — exponentially many more are ungrammatical than are grammatical.
Posted by Geoffrey K. Pullum at January 20, 2005 12:48 PM