Language Log

« previous post | next post »

In my last posting, I noted that reference work after reference work condemns summative uses of pronouns (He snores, which bothers me. He snores, and that bothers me.) on theoretical grounds: a pronoun must have a noun or pronoun as an antecedent (this is a theoretical assumption), but these summative pronouns lack such antecedents – they refer to a situation or proposition alluded to in the preceding discourse context – and so they are unacceptably vague and should be avoided, though summative non-pronominal NPs (He snores, a fact that bothers me. He snores, and that fact bothers me.) are judged to be acceptable in the same contexts.

Examining the practice of good writers would show that they don’t avoid summative pronouns in general (though of course in many contexts alternative constructions would be preferable). But the advice above has theory trumping practice. Then in looking through the sections of Laurie Rozakis’s Complete Idiot’s Guide to Grammar and Style (2nd ed.) that deal with pronouns, I came across an old acquaintance, the Possessive Antecedent Proscription (the PAP), and, by golly, Rozakis (p. 94) advocates it, and gives an entirely theoretical justification for it.

Her version of the PAP:

Don’t use a pronoun to refer to a noun’s possessive form (the form that shows ownership). You can’t use a noun’s form as the antecedent to a pronoun, unless the pronoun is also in the possessive case.

… Confusing: The proctologist’s discovery brought him fame.

(Because the pronoun him is not possessive, it cannot be used to refer to the possessive proctologist’s.)

Clear: The proctologist became famous because of his discovery.

(There’s a lot here to complain about, but let it pass.) If you haven’t seen discussions of the PAP before, you’re probably baffled by Rozakis’s claim that the proctologist sentence in its possessive version is “confusing”. How? Why? Rozakis explains:

When the possessive quality is added to the noun, that noun becomes an adjective and is no longer suitable to be an antecedent.

Ah, a theoretical hypothesis, that possessives are adjectives (plus, again, the hypothesis that the antecedent of a pronoun must be a noun or pronoun)! Your judgments of acceptability and the practice of good writers count for nothing in the face of a theoretical assumption.

We’ve been here before on Language Log, five years ago, in a series of postings here, here, here, and here. And there’s an extended discussion in a detailed handout for an American Dialect Society paper of mine, available here. The very short version: everything Rozakis says is either confused or simply wrong.

The worst thing you can say about possessive antecedents for pronouns is that sometimes they are hard to process (for reasons explained in the references above) and so sound awkward. But simple cases like Mary’s father adores her are impeccable, and even more complex examples can be found all over the place.

Including in the writing of people who espouse the PAP (examples in the references above), who invariably, in my experience, use possessive antecedents on occasion, when they aren’t focused on matters of grammar. Even Rozakis. Here's one example, from a section on unclear pronoun reference (p. 90), which is illustrated entirely with examples that should present no interpretive problem at all for a well-disposed reader using real-world information and discourse organization:

My mother wants to have the dog's tail operated on again, and if it doesn't heal this time, she'll have to be put away. (Are we sending Mom or the pooch to the happy hunting ground?)

The point is that Rozakis is entertaining the possibility that the dog's is the antecedent of she (in violation of the PAP), and in fact this is the only reasonable interpretation. No doubt other violations can be found in the book. There might be some summative pronouns in there as well.

One final remark: when the handbooks appeal to theoretical assumptions (as in the summative pronoun and PAP cases), these assumptions are simply asserted as truth, with no attempt at justifying them as claims about language. When I question people about such assumptions, they either reply that they're self-evident truths or refer to the authority of the grammatical tradition. But the assumptions themselves require defense. English linguists provide arguments for the claim that possessives in English function syntactically as determiners, parallel to the articles, demonstrative determiners, and the like, and not as adjectivals. (Things are different in other languages.) If you have a different proposal, show us your evidence.

May 22, 2008 @ 6:57 pm · Filed by Arnold Zwicky under Prescriptivist poppycock

Permalink