From NPR’s Morning Edition on the 17th, in the story “Farmers Cautious of Drought-Resistant Seeds”:
Z4.72. Like many Iowa farmers, [Gary] Plunkett’s corn harvest numbers have gyrated …
Some usage critics (like Philip B. Corbett, the NYT‘s associate managing editor for standards, in charge of the paper’s style manual) would reject the initial like-phrase out of hand as a “dangling modifier” — see below — but people not under the sway of an explicit rule about these things tend not to see anything at all notable in examples like this one, which are very common, even in careful writing.
Some background, from my posting on a class of “dangling modifiers” that are, ceteris paribus, acceptable, involving initial as a phrases, as in
Z1.3. As a linguist, what struck me especially about his work on meaning in natural language was his belief that… (Stanley Peters, In Memoriam Jon Barwise, 2000)
The larger topic we’re looking at concerns:
a SPAR (a Subjectless Predicative Adjunct Requiring a referent for the missing subject) that’s non-canonical, in that the adjunct doesn’t obey the Subject Rule (doesn’t pick up its referent from the subject of the main clause); such non-canonical SPARs, or X-SPARS, are popularly known as “dangling modifiers”, a condemnatory label. But some types of X-SPARs are in fact acceptable (except to those who have internalized the teaching that X-SPARs are necessarily ungrammatical)
I’ll call the referent for the missing subject in a SPAR the controller for the SPAR. This is usually, but not necessarily, provided by a NP in the text surrounding the SPAR; sometimes the controller is “in the air”, especially when it’s 1st person:
MW2. Speaking as an old friend, there has been a disturbing tendency in statements emanating from Peking to question the good faith of President Reagan. (Richard M. Nixon (cited by William Safire, NYT Magazine, 19 June 1983))
For O-SPARs (ordinary, canonical, SPARs), the controller is the subject of the main clause, and there’s normally no issue about acceptability. In contrast, X-SPAR examples range from some that are very hard to comprehend to certain classes of them that are, ceteris paribus, acceptable; the contributions to acceptability include structural properties of the SPAR (in as a SPARs, for instance), structural properties of the main clause (in main clauses that have dummy-it subjects, for instance — note the following example),
Z3.337. Wrapped in Atlantic mists and storms three hours’ flying time east of Ottawa, it was only in 1948 that Newfoundland and Labrador voted by a slim margin to relinquish its status as a British colony to become the tenth province in Canada. (The Economist 10/13/07, p. 42)
and, especially, the discourse topicality of the controller (the content and organization of preceding discourse is often very important, as I’ve pointed out many times).
That brings me back to Z4.72, with which this posting began. The relevant structural property here is a main-clause subject with a possessive NP determiner, in this case Plunkett’s (in Plunkett’s corn harvest numbers), where the determiner supplies the controller for the SPAR; Plunkett is like many Iowa farmers in that Plunkett’s corn harvest numbers have gyrated.
A couple more examples, one with a pronoun determiner, one with a non-pronominal determiner:
Z4.26. After raising three children and consulting with thousands of mothers of infants, my suggestion to parents of such children is simple … (Elizabeth Hatherell, Neufeld Institute, Winnepeg, answer to query in New Scientist “The Last Word”, 11/26/11, p. 65)
Z4.28. Like several other states, Washington’s minimal wage is pegged to … (NPR Morning Edition reporter, 1/3/12)
I have a considerable number of examples with possessive NPs supplying controllers, and no doubt could easily collect more. The SPARs in them range over a number of structural types — as a + NP, like + NP, unlike + NP, at + AGE, VPprp, VPpsp, subordinator + VPprp, subordinator + VPpsp (with subordinators after, when, since, if, without, before) — and they are generally acceptable, the main exception being in cases where both the possessive NP and the subject NP it’s a constituent of are plausible candidates to supply the controller of the SPAR, as in the invented example
After raising three children, my / Kim’s parents were exhausted. [cf. Z4.26 above]
In such cases, the Subject Rule wins; only the full subject can supply the controller.
Now to Phil Corbett, who posts regularly on the NYT site with collections of examples he judges to be deficient in grammr, usage, and style. Danglers are a regular topic, and in this domain he’s a positive demon about enforcing the Subject Rule; in particular, he rejects any example with a controller supplied by a possessive NP.
There are two main contributions to this attitude. First, there’s his treatment of the Subject Rule as absolute and inviolable, a rule of English grammar, not just a default stylistic choice; for discussion of the (complex) reasoning that would lead someone to this conclusion, see my “Disregarding context”, here.
Then there’s an assumption that Corbett and a number of other usage writers make that would ensure that the Subject Rule is necessarily true — true by definition — when the subject has a possessive determiner, because it rules out that determiner as a possible supplier of the controller for a SPAR: the assumption (a) that possessives are adjectives (or adjective phrases) and not nouns (or noun phrases), in combination with the assumption (b) that a SPAR is a modifier of a noun (the head noun of the subject) in the main clause.
Now (a) is a confusion: possessives (like other determiners) are modifiers, but there is hardly any way in which they behave syntactically like adjectives (or adjective phrases). This assumption confuses syntactic function (Modifier, of which determiners are one class of expressions serving in that function) with syntactic category (N or NP).
Why should we care whether determiners are adjectives or nouns? Because of other assumptions that hinge on the distinction, for instance (b). The assumption that SPARs modify something, in particular a N(P), in the main clause makes the assignment of possessives to Adj(P) or N(P) crucial. (Assumption (b) is just dogma from a version of “traditional” grammar, not argued for any place I know of, and you can argue that the relevant relationship involves coreference, not modification. Nevertheless, we’re still looking for a N(P) in the main clause.)
Here are two of Corbett’s examples that turn on assumption (a), with his comments:
Z4.35. More than some other large banks, Bank of America’s fate is also heavily intertwined with that of consumers.
[Corbett’s commentary: The thing being compared to other large banks is Bank of America — not “Bank of America’s fate.” Rephrase.]
Z4.39. PSP-I-POSS-3P-OK Always drawn to the sciences, Mr. Kaido’s interest in progressive politics grew in 1970, after he entered the law department at the University of Tokyo, a steppingstone for the country’s elite.
[Corbett’s commentary: This is probably the most common dangler situation. The person described by the modifying phrase, Mr. Kaido, does come immediately after the phrase — but as a possessive modifier, not as a noun. Strictly speaking, in this formulation the introductory phrase is modifying “interest,” which is not what we intended.]
Note that I’m not claiming that a possessive can always supply the controller for a SPAR, only that it can do so ceteris paribus. When the whole subject is a plausible controller, the determiner is unavailable, as in this example discussed by Mark Liberman here:
Z4.51. At one show in particular, ex-Germs/45 Grave drummer Don Bolles went to review the band’s live performance for the L.A. Weekly newspaper and gave the band a favorable review. The following week the band went to Bolles’ apartment in hopes of persuading him to join the group. When asked to join the band, Bolles’ pet rat went into a spastic fit and died. Bolles took this as some sort of strange sign and joined the group cementing his spot as the band’s permanent drummer.
Mark notes a connection between these SPAR phenomena and the “rule” I’ve labeled the Possessive Antecedent Proscription, which treats the following as ungrammatical if Mary’s and her are supposed to be coreferential:
Mary’s father admires her.
(Mark gives links to relevant Language Log postings.) The claim (in very simplified form) is that a possessive can’t serve as the antecedent for a pronoun — because, it is said, a possessive isn’t a N(P). The parallel between possessive controllers for SPARs and possessive antecedents for pronouns isn’t perfect, but you can see that the status of possessive determiners as AdjP or NP plays a central role in both phenomena.
A final note: talking about the Subject Rule and the Possessive Antecedent Proscription can be incredibly hard, because both constraints have been taught (explicitly or implicitly) in schools or other settings and internalized by some minority of speakers, mostly highly educated and articulate speakers. These people are generally incapable of disregarding what they’ve learned; the constraints have become part of their grammars, and they’re inclined to view the graded, contextually sensitive judgments that the rest of us make as abandoning correctness.