The Economist's article on the Cumbrian shooting rampage opens with this nicely styled and balanced sentence:
"It's like watching something from America," said one resident of Whitehaven, a gentle Georgian town on the north-western English coast. [The Economist 5 June 2010 p.33]
The subject of said has been postposed. This improves intelligibility because the subject is rather long (it has an attached supplement, the noun phrase a gentle Georgian town on the north-western English coast).
Now compare the following glaringly inept piece of style from a recent issue of The New Yorker:
"Galleries and magazines send him things, and he doesn't even open them," Zhao Zhao, a younger artist who works as one of Ai's assistants, said. [The New Yorker 24 May 2010 p.56]
Grossly and unnecessarily clumsy, and hard to process. What on earth is wrong with them?
Let me make it clear that I am not saying their sentence is ungrammatical. I am saying that it is an example of very poorly chosen style with respect to constituent order. And it appears that it does not represent an isolated slip. As Chris Potts noted on Language Log way back in 2003 ("A ban on quotative inversion?"), The New Yorker apparently has a house-style prohibition on (if I may use the technical terms employed in The Cambridge Grammar of the English Language) subject postposing in a parenthetical report frame for directly reported speech, even when the quoted speech is preposed.
They ban "said NP" even when the subject NP is long and complex. In fact they ban it even when the subject contains additional parenthetical interruptions and thus cries out to have a place at the end of the clause. Chris cites this sentence:
"He used to have this great, dignified passion to him," Christopher Hitchens, who, until his own political change of heart, defended Chomsky, says. (Larissa MacFarquhar, "The devil's accountant", The New Yorker, March 31, 2003, p.67, column 2.)
The subject together with the verb says make up what CGEL calls the reporting frame, and here it is parenthetical. But the subject (the phrase Christopher Hitchens, who, until his own political change of heart, defended Chomsky) itself contains a supplement (the relative clause who, until his own political change of heart, defended Chomsky, which is of the supplementary type), and that too contains a supplement (the parenthetically comma-marked temporal adjunct until his own political change of heart). So we have a supplement inside a supplement inside a supplement, and still the New Yorker's fierce and unyielding house style code will not allow the subject to be postposed, to yield what could have been a perfectly acceptable sentence:
"He used to have this great, dignified passion to him," says Christopher Hitchens, who, until his own political change of heart, defended Chomsky.
How on earth can The New Yorker struggle on, printing horrible stylistic botches solely in order to deny the reader the comfort of a nicely balanced sentence with a postposed subject in a reporting frame? Why do they treat constructions as simple and natural as "Good gracious", said the bishop as if they represented some kind of gross grammatical turpitude?
What sort of writer do they admire? Is Charles Dickens all right? I chose a Dickens novel at random (A Tale of Two Cities), and rapidly found "Wo-ho!" said the coachman in the 6th paragraph of Chapter II. Would G. K. Chesterton be considered capable of writing grammatical English prose? I found "The two clergymen," said the waiter… in the first chapter of The Innocence of Father Brown.
It is irritating to waste even this much time confirming something so obvious: there is no such thing as a fluent native speaker and reader of Standard English who rejects such sentences. Someone at The New Yorker is stone crazy.
It's remarkable that editors can be found who have the syntactic sensitivity to spot the opportunities for applying this ban and carry on the work of enforcing it. Most people would not be able to follow an instruction like "Keep to strict subject-verb order in all reporting frames with verbs of saying, even when parenthetical, no matter how long the subject may be." It isn't that easy to keep an eye open for specific syntactic features so you can change them; people tend to read for content and literary effect. But The New Yorker has someone who (i) can identify these syntactic situations flawlessly, and (ii) is astonishingly obedient, and (iii) works for a boss who is stone crazy.
The New Yorker does appear to have staff members who read Language Log. When I mocked their search engine for its ridiculous error message "I'm sorry I couldn't find that for which you were looking", and remarked that I couldn't imagine of what they were thinking, they changed it within about a week. (They still wouldn't risk a stranded preposition, though; they changed it to "Sorry, there are no results matching that search.") But they have never changed the edict about order of subject and verb in reporting frames. Or so it would seem from the case I quoted above (let me point out that I have not yet undertaken a systematic search of their archives, so I am talking about a purely anecdotal observation here).
If some mole in the New Yorker offices would like to explain more, and they dare not comment below, they can reach me at my Gmail account (the login name is my surname), and I promise to protect their anonymity. I realize that at The New Yorker the ghost of E. B. White still walks abroad, and that to strand a preposition or split an infinitive or postpose a subject (or use singular they as I am doing here) would put someone in fear of losing not just their job but their immortal soul. Still, they can share with me privately whatever they know about the reasoning behind the ban, and help me understand why they are being required to favor appallingly clumsy sentence construction over uncontroversially acceptable syntax. Chris Potts and I, and the Language Log readership, would love to know why.