September 24, 2005

The New Yorker is unfazed (though ungrammatical)

In reference to Geoff Pullum's posting of Alex Gregory's blogs-as-pointless-incessant-barking cartoon, Heather Green at BusinessWeek's Blogspotting site has posted some correspondence from Bob Mankoff, identified as "cartoon editor of The New Yorker and president of the Cartoon Bank service, which is owned by the New Yorker". Ms. Green tells us that she "was intrigued that though the New Yorker and their pr seem aware of the copyright violation that's going on with this, they appear unfazed". For my part, I was intrigued to find Business Week apparently in pursuit of possible small-time infringements of other people's intellectual property rights.

I imagine that Mr. Mankoff sees cartoon-blogging as an extension of the traditional practice of posting cartoons on refrigerators, bulletin boards and walls, which has always provided valuable free advertising for the cartoonists and their publishers, while enriching the culture at large. If so, this is a mature and sensible response, which I applaud. Another view of his reaction, however, is that blogs and other online media are simply beneath one's notice.

Mr. Mankoff's response does emit a certain odor of condescension, along with some linguistic oddities:

"While many cartoons on blogs have been submitted and rejected by The New Yorker, this one, seemed to us, to perfectly capture the irony inherent in a communications phenomenon that permits so many to say so little about so much. I think it will become a classic, in every way as emblematic of its time, as this other cartoon, also involving dogs, published in 1993 by Peter Steiner was of its."

This being Language Log, rather than Social Psychology Log, I'll focus on the linguistic oddities, starting with the apparent example of anti-FLoP coordination in the first clause:

... many cartoons ... have been submitted and rejected by The New Yorker ...

Incompletely parallel coordinations of this type are usually unambiguous but also ungrammatical. A New Yorker example that I cited earlier also involved leaving out to in the first element of a verbal coordination, resulting in a sentence that apparently can't be parsed at all in its intended reading:

... for that to happen a person would have to have been exposed at great length, or have eaten raw, infected poultry.

You could construe this sentence as conjoining [have been exposed at great length] or [have eaten raw, infected poultry], but I think it's clear from the context that the author meant to convey that a person would have had to have been exposed at great length to infected poultry, or to have eaten infected poultry raw. You can't parse the sentence that way without adding a missing to.

In contrast, Mr. Mankoff's subordinate clause has two valid parses, one that evokes the image of The New Yorker repeatedly rejecting its own submissions -- [[submitted and rejected] by The New Yorker] -- and another one that means what he meant to say, whose structure is [[submitted] and [rejected by The New Yorker]]. However, this second parse strikes me as so awkwardly asymmetrical as to be unlikely, and I wonder whether instead he meant the first structure with the second meaning, a missing to having been lost in the heat of composition.

There's some other evidence of compositional carelessness in Mankoff's note, such as the puzzling commas around "seemed to us":

While many cartoons on blogs have been submitted and rejected by The New Yorker, this one, seemed to us, to perfectly capture the irony inherent in a communications phenomenon that permits so many to say so little about so much.

Presumably this is a blend of "this one, it seemed to us, perfectly captured..." and "this one seemed to us to perfectly capture..."

Mr. Mankoff is not the first person to (mis-)use {"submitted and rejected by"}, for which Google finds 150 hits, e.g.

Subsequent designs were submitted and rejected by the Commission in November 1929.
Once a petition for certification has been submitted and rejected by the Commission, the same signatures may not be submitted in any subsequent petition to certify a new political party.
IF AN APPLICATION IS SUBMITTED AND REJECTED BY COUNCIL ON THREE (3) OCCASIONS ... IT WILL BE REMOVED FROM ANY FURTHER CONSIDERATION ...
Five separate plans were submitted and rejected by New York City agencies over a twenty- year period.

I suspect that in many of these examples, the author intended to conjoin [submitted and rejected], and just lost track of the to. More than twice as many authors (379) used the pattern {"submitted to and rejected by"} to express similar ideas:

Authors must inform the editor if their manuscript has been submitted to, and rejected by, another journal.
... and also regarding the recent manuscript which was submitted to and rejected by the Journal of Clinical Oncology ...
... the charter amendments which were submitted to and rejected by the electors in 1938 ...
... generally such complaints must provide proof that the complaint has first been submitted to and rejected by the relevant local law enforcement agency ...

And I'll bet that a much larger number wisely decided to frame their thoughts in some other way altogether. As evidence, observe that comparable but fully parallel sequences are much commoner: "considered and rejected by" gets 28,300 Google hits, "addressed and delivered to" gets 10,700, "written and submitted by" gets 7,140,000, and so on.

Mankoff's second sentence also has some curious comma-placement choices, an oddly placed by-phrase, and a strange ending:

I think it will become a classic, in every way as emblematic of its time, as this other cartoon, also involving dogs, published in 1993 by Peter Steiner was of its.

The noun phrase "this other cartoon" has three separate and parallel post-modifiers:

  • also involving dogs
  • published in 1993
  • by Peter Steiner

It's hard not to read this sentence as suggesting that the cartoon was published in 1993 by Peter Steiner, although this is obviously not what Mr. Mankoff meant: the cartoon was created by Mr. Steiner, but published by The New Yorker. Stacking up postmodifiers like this is usually a bad idea, especially when adjacent pairs have as strong an affinity for one another as "published in 1993 by Peter Steiner".

And last but by no means least, we come to the ending "was of its". After clearing out the textual undergrowth, we have:

It [is] ... as emblematic of of its time as this other cartoon ... was of its.

My attempts to find other phrases of this type on the web turned up almost exclusively a set of SEO blackhat pseudo-text sites. (Compare the results of this search, which substitutes "of his" for "of its".)

The reason, I think, is that constructions with its as the object of a preposition, by analogy to phrases with mine, hers, his, and theirs, are simply ungrammatical:

Her legs are longer than mine.
*Looking at the lion, I concluded that my legs are longer than its.

She returned with some relatives of hers.
*The bird returned with some relatives of its.

Bill Gates is as important to his company as Steve Jobs is to his.
*The hyena is as important to its ecosystem as the polar bear is to its.

[As Arnold Zwicky has pointed out to me, the problem seems to arise in other cases where its winds up stressed, so that these are also problematic:

*I wiggled my ears, and then the dog wiggled its.
?My ears are bigger than my cat's, but its are nicer to stroke.

though the second one doesn't seem so bad to me.

Arnold observes that Jorge Hankamer wrote about this in a 1973 paper, "Why there are two than's in English. CLS 9.179-91, and adds

Jorge has the constraint in an especially robust form; he really hates all occurrences of accented "it". For a fair number of other speakers -- I am one -- there's a dispreference rather than a constraint. I find Mankoff's sentence clunky, but not unacceptable.

]

Mankoff's feelings about this are probably like Arnold's, though it's also possible that he expressed a complicated thought a bit carelessly, so that the ungrammaticality of the result was obscured by its complexity.

I guess I need to admit at this point that I've aligned myself -- and not for the first time -- with the cause of rationalist prescriptivism, according to which certain ways of talking or writing are judged to be ill-advised on logical or analytical grounds. But unlike the plural people peeve and many other examples where the appeal to alleged linguistic logic is mistaken and even silly, the questions that I've raised about Mankoff's note seem to me to be justified ones.

For a lucid defense of this point of view in general, see Geoff Pullum's post "'Everything is correct' vs. 'Nothing is relevant'".

[Update: Chris Waigl found more evidence that other writers on the web share Mankoff's willingness to use its as the object of a preposition at the end of a clause:

link) We don't, for example, inevitably compare every three-panel gag-a-day strip to Garfield, even though it is arguably as influential to its form as Far Side was to its.
(link)It bears the same relationship to its genre as Hamlet does to its, and while it doesn't have the same level of language to sustain it, that wouldn't be appropriate in the medium.

Chris also found several historical examples:

(link) This is in accordance with our policy of not intervening unless the European powers are unable to agree and make request for our assistance. Whenever they are able to agree of their own accord it is especially gratifying to its, and such agreements may be sure of our sympathetic support. (Calvin Coolidge, State of the Union Address, *December 8, 1925)
(link)In short, it was he who turned Austria on its axis, and France on its, and brought them to the kissing pitch. (Thomas Carlyle's "History of Friedrich II of Prussia V")
(link)Meanwhile he is the great artisan and laborer by whose aid men are enabled to build a world within a world, or, at least, to smooth down the rough creation which Nature flung to its. (Fire Worship, by Nathaniel Hawthorne )

Chris observes that

Admittedly, the Coolidge cite rather makes your point: I'm quite unsure what the antecedent of "its" is supposed to be.

Yes, I wonder whether there might have been a scribal (or OCR) error in that case: the sentence as presented seems not so much ungrammatical as simply incoherent. I suspect that the "its" should actually have been "us".

Finally, Chris finds some historical examples of stressed its as the object of a verb rather than a preposition.

(link) The water consumed in the three crank engine is 12.93 lb., against 13.0 in the two crank, but the former drives its ship nearly ½ knot per hour faster than the latter does its, and when both ships are driven at the same speed the consumption of coal in the three crank ship is considerably less than in the other. (TRIPLE COMPOUND ENGINES. By Mr. A.E. SEATON. In: Scientific American Supplement, No. 492, June 6, 1885)
(link)It was at this point that her keen attention became fixed on him and never afterwards wavered. If everything had its story, the mistletoe would have its; he must interpret that: and thus he himself unexpectedly had brought about the situation she wished. (Bride of the Mistletoe, by James Lane Allen)

These seem just as problematic to me, but I'm becoming persuaded that this is one of those cases like "such the surprise", where the English language has some corners that not everyone is able to visit.

Apparently reactions to these "accented it" sentences range from mine (they're just as bad as "I gave it to he") through Arnold Zwicky's (they're awkward but basically OK) to some other readers' (they're perfectly fine, there's no problem with them at all). I wonder what the distribution of these judgments are, overall and by age, region and so on. It's amazing how many aspects of the English language remain essentially undocumented.]

Posted by Mark Liberman at September 24, 2005 12:06 PM