Via Geoff Nunberg, there’s a little brouhaha brewing in blogland about a question of grammar. The question is whether this sentence is grammatical.

(T) Toni Morrison’s genius enables her to create novels that arise from and express the injustices African Americans have endured.

How could such a brouhaha have brewed? Because blogland bothers ’bout bothersome rules? Not exactly. One of the questions on the PSATs was whether (T) was grammatical. The answer ETS was looking for was no. Maryland teacher Kevin Keegan argued that the answer was that it was, for a reason that I found rather baffling. Here’s his side of the story (from the Washington Post article that broke the news.)

The word “her,” he posited, was improperly referring to “Toni Morrison’s,” so the answer should have been “A,” signifying a mistake in “her to create.” Many grammar manuals insist that a pronoun such as “her” should refer only to a noun, not, as in the case of the possessive “Toni Morrison’s,” an adjective.

Later in the article it transpires the “many” grammar manuals are in fact two, and they are unnamed. The rule seems preposterous on its face. As Eugene Volokh notes, it would rule that (J) is also ungrammatical. (Eugene also wonders what the rulebooks are that rule (T) et al out.)

(J) John’s legs couldn’t carry him any further

So there are some intuitive counterexamples, which is strange for an alleged syntactic rule involving pronominalisation. We don’t need a schoolbook to tell us that “She likes her” cannot be used to mean that she likes herself.

And it is hard to see the theoretical motivation for the rule. Surely “Toni Morrison” is a constituent of “Toni Morrison’s”, so I don’t see why it can’t be attached to an anaphoric pronoun. The rule can’t be that constituents of longer NPs cannot be linked to anaphoric pronouns, else (S) would be bad.

(S) After the party, Jack and Jill repaired to his apartment for coffee.

So the rule must be very dependent on surface structure. “Toni Morrison” is not a standalone word in (T), so ‘her’ cannot refer back to it. I doubt that where the spaces go, whether we use “Toni Morrison’s genius” or “The genius of Toni Morrison”, cannot make that much of a difference. But maybe that’s just because I don’t really speak English. (And if the fuss is over whether the NP is a standalone word, there’s no reason ‘her’ could not attach to ‘Toni’. Now there are good reasons in terms of the underlying structure of the sentence that ‘her’ could not attach to ‘Toni’. For one thing ‘Toni’ here is not an NP. But once we start looking that deep, we see that there’s no reason ‘her’ couldn’t attach to ‘Toni Morrison’.)

The unintentional humour prize in all this goes to David Skinner in the Weekly Standard, who supports the decides to take the opportunity to denigrate Morrison’s novels, and as Matthew Yglesias notes manages to harm his own reputation as a writer somewhat more than he damages Morrison’s. Skinner (who bears some resemblance to his funnier British namesake) thinks:

Thanks to a vigilant English teacher from Maryland, the sentence has been proven to contain an error of grammar (“her” doesn’t refer back to a proper female subject noun, but to the possessive “Toni Morrison’s genius”).

If a ‘proper female subject noun’ is a proper noun then (M) is bad, which would be odd.

(M) The Prime Minister liked anyone who shared her fondness for 70s death metal bands.

If a ‘proper female subject noun’ is not a proper noun, then I have no idea what Skinner is writing about. And by the time he actually gets to slagging off at Morrison’s novels, which it seems was the point of his article, it’s pretty clear he doesn’t either.

UPDATE: Also via Geoff Nunberg, the rule is supported in Wilson Follett’s 1966 Modern American Usage:

A noun in the possessive case, being functionally an adjective, is seldom a competent antecedent of a pronoun: On F’s arrival from Virginia at La Guardia Airport last night, he denied to reporters that… F would legitimately lead to he; F’s cannot. Reconstruct, then: F, on his arrival, denied… (Of course a possessive noun can be the antecedent of a possessive — i.e., an adjectival–pronoun: F’s denial was made on his arrival.)

This doesn’t make the rule correct, but it does provide some support for the argument that a question like this should not have been used on the PSAT. I think Keegan’s original complaint was right about at least that much. Much thanks to Geoff Nunberg for the link.

It might be worth noting that I got the (alleged)rule wrong in the some of the above examples. My (T) and (M) are OK by Follett’s lights because the anaphor occurs in a possessive phrase. I hadn’t imagined that could make a difference to grammaticality, which just goes to show how unimaginative I am.