Playing with anaphora

Mike Keefe’s Denver Post editorial cartoon of November 15:

Out of context, “I’m going to China to visit it” is puzzling, since the referent for the pronoun it would appear to be China, so that Obama would seem to be saying that he’s going to China to visit China, which is sensical but fatuous. But the preceding sentence supplies a better referent for the pronoun it, namely the American economy.

This little text still requires some interpretive work, since you have to extend the meaning of visit some to understand how someone could visit the American economy. That’s not very hard to do, and then you see that the text conveys that the American economy is in China. Clever.

Ellipsis on an island

Malcolm Gladwell, “Offensive Play”, New Yorker 10/19/09, p. 52, quoting a football player:

(1) “They cleared me for practice that Thursday. I probably shouldn’t have. I don’t know what damage I did from that, because my head was really hurting.”

“I probably shouldn’t have ___” contains an instance of Verb Phrase Ellipsis (VPE); the underlines mark the location of the elliptical material. VPE is a type of anaphora, zero anaphora in particular, so we need to find a referent for the missing VP.

The way VPE normally works is that the referent is supplied by an overt VP in the linguistic context that serves as an antecedent for the anaphor, as in this real-life example, where the antecedent is bold-faced.

I lost weight with Jenny Craig, and you can ___ too.

(that is, you can lose weight with Jenny Craig too.)

But sometimes the referent has to be dug out from non-VP material. Some people find such examples unacceptable — they are often at least hard to process — and there’s a considerable literature about some of them, under the heading “anaphoric islands”; see the Language Log discussion here.

(1) is such a case, where the elliptical material is something like “practiced” or “gone to practice” and the referent has to be dug out “from within” the noun practice, which is derived from the verb practice.

Some further examples (some of them intentionally jokey) from my collection:

(2) Many cases go unrecorded, and those that are ___ rarely make it to court. [referent from within the adjective unrecorded]

(3) Me: That’s a gift.
Wife: And he is ___. He’s very gifted. [referent from within the noun gift; note wife's repair.]

(4) Constants aren’t ___ and variables don’t ___. [referents from within the nouns constants and variables]

(5) Friendly fire isn’t ___. [referent from within the NP friendly fire]

(6) one of those see-through blouses you don’t even want to ___! [referent from within the adjective see-through]

(7) A Writer Who Doesn’t ___ [referent from within the noun writer]

Just In: NYT Violates PAP!

I haven’t posted for some time on the Possessive Antecedent Proscription (PAP), a fictitious principle of English usage/grammar that bars possessive-marked nouns as antecedents for personal pronouns, as in this NYT front-page headline on October 9:

Astor’s Son Is Convicted of Stealing From Her

Astor’s (referring to Brooke Astor, “the legendary New York society matriarch”) is the possessive-marked noun, and it serves as antecedent for the personal pronoun her (functioning here as the object of the preposition from).

Most readers will fail to see any problem with the NYT headline, but a few will recoil from it. These will be people who have been explicitly taught the PAP, in school or in a usage handbook or style sheet. I have yet to come across anyone who tacitly induced the PAP from their linguistic experience. In fact, in my experience everyone who espouses the PAP violates it (with no apparent awareness of having mis-stepped) on occasion. (I used to track such violations down, but it was a time-consuming and unrewarding occupation, and I’ve given it up.)

[Before I go on, let me say that if you want to cleave to the PAP, that's fine. If you do so, no one will criticize you for your bit of harmless nuttiness (sort of like not stepping on the cracks in sidewalks). Just don't go around slapping people down for not sharing your avoidance of possessive antecedents.]

The PAP seems to be mostly a product of three bad ideas:

(1) the idea that pronouns are simply replacements for repeated nouns (“That’s why they’re called pronouns, dummy!”);

(2) the idea that possessive-marked nouns are adjectives (because they modify — in some sense of modify — nouns), so of course — see (1) — they can’t serve as antecedents for pronouns; and

(3) the idea that if a linguistic element can in some way contribute to difficulty in understanding, ambiguity, unclarity, or awkwardness, then it should always be barred.

Idea (3) is breath-takingly silly, though it’s trotted out again and again as ammunition against some usage or construction a writer doesn’t like; the writer cites some examples, involving the item in question, where one of these defects arises. Taken at face value, (3) would prohibit speech and writing completely.

The other two ideas spring from more technical misapprehensions, both with long histories in the Western intellectual tradition. But in both cases, they are just hypotheses, however venerable, and most modern linguists reject them, for good reason.

For a more extended discussion of the PAP, see the material here (please check this before you comment on this posting), and note these Language Log postings:

GP, 10/5/03: Menand’s acumen deserts him: (link)

AZ, 10/8/03: Louis Menand’s pronouns: (link)

AZ, 10/21/03: Grammaticality, anaphora, and all that: (link)

AZ, 10/23/03: In search of the fimpossant: (link)

AZ, 2/20/06: Collateral damage: (link)

AZ, 5/22/08: More theory trumping practice: (link)

The implicated event of pizza-eating

I have a large and ever-growing collection of notes to myself on linguistic topics. This morning I came across one of these notes, a slip of paper with the following example on it:

(1) They came by bearing pizza, after which we watched The Music Man.

(The note doesn’t identify the source of the sentence or the date when I collected it.)

On the most straightforward reading, this sentence has a summative relative clause, “after which we watched The Music Man“, in which the relativizer which refers to the event of some people’s coming by bearing pizza; that is, (1) asserts describes two events, an arrival-with-pizza event and a movie-watching event, occurring in that order. In still other words, (1) is paraphraseable as

(2) They came by bearing pizza, after which event we watched The Music Man.

(In fact, some usage writers insist that (1) is unacceptable, because which has no noun antecedent in the sentence — so that (1) is “vague” — and that something like the clunky (2) must be used instead. See the summatives posting linked to above.)

And now for a subtlety. Although (1) describes only two events, most readers will understand (1) as implicating a third event, of pizza-eating, intervening between the other two, and the author of (1) surely intended this implicature. A nice little case of how sentences can end up conveying more than they literally mean. (The sentence is true if no event of pizza-eating occurred on the occasion in question.)