Cnoindented metalanguage

« previous post | next post »

Another example of e-publishing string-replacement gone wrong, from the Kindle edition of Ed McBain's Blood Relatives, originally published in 1975:

The problem here, of course, is that someone turned "hang" into "noindent" in the middle of the words change and changed. This is different from the case of kindle → nook, and the case of its → it's, in that the substitution involves leakage into the book's text of an attempt to correct its text-formatting metalanguage, presumably changing (or, I should say, "cnoindenting") one way of indicating a "hanging paragraph" into another, or perhaps cnoindenting a hanging paragraph into a block paragraph.

It's therefore oddly appropriate that I had already marked this work for a Language Log post, on account of a clever replaying of the Cretan Liar paradox, which also involves a subtle transgression of the language/metalanguage boundary:

Carella didn’t know any private eyes. He knew a lot of cops, though, and hardly any of them behaved the way television cops did. […]  Television cops were dangerous. They made real-life cops feel like heroes instead of hard-working slobs. Carella did not feel like a hero when he got back from the Criminal Courts Building that afternoon. He had left the down-town area at 11:45, and it was now almost 12:30, and he still hadn’t had lunch, and the first thing he saw on his desk when he walked into the squadroom was a memo from the Police Commissioner. The memo may not have disturbed Carella had he not just been thinking about life imitating art imitating life and so on. But it disturbed him now. It very definitely disturbed him. This is what the memo read:



The memo was signed by the Police Commissioner, Alfred James Dougherty. There was only one trouble with the memo. In signing it, the commissioner, or his secretary, or one of his aides had used a rubber stamp.

Carella looked at the memo and at the rubber stamp signature.

The commissioner had clearly ordered that effective this date rubber stamp signatures could not be used on any official correspondence. The memo also stated that any orders or instructions signed with a rubber stamp signature were to be ignored.

Carella’s perplexity was monumental.

He sat at his desk and read the memo again, and then he read it a third time, and tried to decide what he should do about it. His deductive reasoning went something like this:

(1) The commissioner’s memo had been signed with a rubber stamp.
(2) Therefore, the commissioner’s memo was to be ignored.
(3) If the memo was to be ignored, then the use of a rubber stamp signature on official correspondence was still permitted.
(4) And if the rubber stamp signature was still permitted, then any orders or instructions signed with such a signature were not to be ignored.
(5) Therefore, the commissioner’s memo was not to be ignored.
6) But if the commissioner’s memo was not to be ignored, then it outlawed all rubber stamp signatures, and since the memo had been signed with a rubber stamp, it clearly was to be ignored.
(7) Therefore, the commissioner’s memo was to be ignored and was also not to be ignored.

Carella blinked, and looked up at the clock. Only two minutes had passed since the commissioner started causing him heartburn. He decided to go out to lunch.


  1. S Johnson said,

    June 7, 2012 @ 10:44 pm

    The solution is that the memo is not effective until signed. After it was rubber-stamped, no further rubber-stamp signatures are permitted.

  2. NW said,

    June 8, 2012 @ 3:33 am

    Unless the order was issued on the stroke of midnight, there must have been a time that day before it was issued that valid rubber-stamped orders could have been issued, so 'effective this date' must refer to only that part of this date after that order came into effect. At the moment of stamping down, stamping down was still a valid method of issuing orders, and the order would come into effect with the completion of the stamping down. I think that would satisfy me.

  3. Carl said,

    June 8, 2012 @ 7:03 am

    @Nick and S Johnson: the issue is that point 2 says that all rubber stamped orders at to be ignored – the character takes that to apply retroactively as well. So even if it was valid at the time the order was issued to rubber stamp it, after it was issued it would immediately be ignored because it was rubber stamped. Except that, if it is ignored, then the stamp is OK, and the order becomes valid.

    This is a variation of the liar paradox "this sentence is false" in which the truth values seem to oscillate under analysis. Actually it's a pretty clever variation, I might use it in a class.

  4. languagehat said,

    June 8, 2012 @ 7:51 am

    The memo may not have disturbed Carella had he not just been thinking about life imitating art imitating life and so on. But it disturbed him now. It very definitely disturbed him.

    The use of "may" for "might" in a contrary-to-fact statement disturbs me. It very definitely disturbs me. I'm not saying it's wrong, mind you; that would violate my firmly held descriptivist principles. If everyone says it, it can’t be wrong. Since everyone these days says and writes it, it's not wrong. I simply dislike it intensely.

    [(myl) I wonder whether this pattern has increased in frequency recently. It certain has been around for a while — and indeed it strikes me as somewhat antique in tone:

    [Obstetric Gazette, 1883] This accident may not have occurred if the operation had been done at the patient's house and if she had remained quiet for a certain period following. This then is an avoidable accident.

    [Poet and Merchant, 1877] Who knows, dear daughter, what a misfortune may not have awaited thee if thou hadst wedded Daniel Isserlein ?


  5. C Thornett said,

    June 8, 2012 @ 8:44 am

    'May' (as per languagehat), also jars because the passage is a narrative past, and so one would normally expect past forms of those modal verbs which have them.

    Does anyone know, or can they conjecture, why some writers use present 'may' rather than past 'might' in passages set in the past or reported speech, aside from any counterfactual considerations?

  6. Ran Ari-Gur said,

    June 8, 2012 @ 9:16 am

    @languagehat: Same here. It really jumps out at me.

    Not quite as much as "cnoindent", though.

  7. rootlesscosmo said,

    June 8, 2012 @ 9:30 am

    Could "cnoindented" for "changed" be an example of a cuperstannniumo?

  8. Gene Callahan said,

    June 8, 2012 @ 9:45 am

    Just after I was admitted to LSE to study philosophy, I received an e-mail from the IT department that with the subject "Test" and a body that read "Please ignore this e-mail."

    I figured it was a test to see if philosophy students could handle paradoxes.

  9. rootlesscosmo said,

    June 8, 2012 @ 9:48 am

    Wouldn't you know I'd misspell "stannium…" Sorry.

  10. Ran Ari-Gur said,

    June 8, 2012 @ 9:51 am

    Also @languagehat: I don't think it's true that "everyone these days says and writes it". I find it wrong, not because of internalized prescriptivism, and not because of abstract reasoning about how the words should work — for me "may" and "might" are largely interchangeable, and this is one of the few situations where for some reason only one is possible — but simply because it sounds strange and wrong to me. It's not part of the English that I grew up speaking. (I endorse your descriptive urge: obviously it is part of the English that many people do speak, and we can't say that their English is therefore wrong. But it's absolutely nothing like the "try and" construction described in your link, which actually is used by everyone, and sounds wrong only to people who've trained themselves to believe it's wrong.)

  11. Robert Coren said,

    June 8, 2012 @ 10:21 am

    This use of "may" bothered me, too, but "This is what the memo read" bothered me more. "This is what the memo said", "This is how the memo read", "The memo read as follows" — all fine, although I probably wouldn't use the second one. I find myself having difficulty articulating why "what the memo read" seems wrong to me, but it does.

    The "cnoindente" incident confirms my view, already inspired by earlier posts on the subject, that anyone who uses, or authorizes the use of, global substitution without human intervention to make sure nothing stupid has resulted from it should be summarily fired, and required to undergo remedial training before they're allowed to do anything involving published text again.

  12. Jerry Friedman said,

    June 8, 2012 @ 10:28 am

    GKP's Language Log post about may with past-time reference. I too dislike this use of may.

  13. Lee said,

    June 8, 2012 @ 10:31 am

    I'm disturbed by "only one trouble with the memo." I can have trouble, and I can have troubles, but in my version of English, I can't have one trouble (or, for that matter, two or three or more troubles). So now I'm very definitely disturbed. And troubled.

  14. Dan T. said,

    June 8, 2012 @ 10:57 am

    I believe I once got an office memo by e-mail which said not to open any e-mail attachments (I think it came in the wake of some virus/worm issues), but the memo was sent as a M$ Word attachment instead of in the normal message body.

  15. Daniel Barkalow said,

    June 8, 2012 @ 11:43 am

    At least he didn't start worrying about how he was going to figure out which of the many standing orders and instructions that he'd been following long after discarding the original documents he would now need to start ignoring. But he neglected to consider the possibility that, while he shouldn't take a rubber-stamped document to have the force of an order, he could use it as a clue. Either there is a valid order instructing him to ignore rubber-stamped orders, or there is an invalid rubber-stamped order; either way, rubber-stamped orders are untrustworthy, and it doesn't matter whether he knows by command or by demonstration.

  16. BZ said,

    June 8, 2012 @ 11:46 am

    Dan T, Well, then, you already opened that attachment. The instruction could not apply to your past actions. So what's the problem?

  17. Ralph Hickok said,

    June 8, 2012 @ 12:57 pm

    This reminds of an old joke about a junior Army officer who had mistakenly initialed document. He was told by his superior to erase his initials and initial the erasure.

  18. D.O. said,

    June 8, 2012 @ 1:20 pm

    I also noticed "may" and "read", though not "troubled". But gentlemen, if you are going to be bothered or troubled by such things, can you read any English text with serenity?

  19. Toma said,

    June 8, 2012 @ 1:35 pm

    This statement is false.

  20. H Klang said,

    June 8, 2012 @ 3:02 pm

    rootlesscosmo said,

    Could "cnoindented" for "changed" be an example of a cuperstannniumo?

    Nniearly onni the nniail…

  21. The Ridger said,

    June 8, 2012 @ 5:22 pm

    I read nearly all the 87th Precinct novels back in college and they always had something that snapped me out of the narrative. I came to expect it from McBain (just as I did from Dell Shannon, whose characters would say, for instance, "That's the hell of a thing"). After I learned he was Evan Hunter, I decided it had be a deliberate thing, because I didn't have the same experience in The Blackboard Jungle or Every Little Crook and Nanny…

  22. Dan said,

    June 8, 2012 @ 5:34 pm

    I'm curious how this ends up happening in a publishing environment. Usually the markup language woul d have some sort of escape character that would be included in the search-and-replace. I.e., in LaTeX one would presumably look for "\hang" and replace with "\noindent".

    Interestingly, if it was LaTeX and the change was done by a UNIX neophyte, I could see this happening with sed and bash. If you're not careful with your quotes and escapes:

    $sed -e 's/\\hang/\\noindent/g'

    gives the desired result, while:

    $sed -e s/\hang/\noindent/g

    changes both "changed" and "\hang".

    However, I have no idea how much LaTeX/UNIX is used in real publishing.

  23. John Lawler said,

    June 8, 2012 @ 6:07 pm

    Not quite the same kind of error as global string changing, but equally meta-, a friend sent me this quotation about German spelling:

    "Vor der Reform der deutschen Rechtschreibung von 1996 wurde die übliche Bezeichnung für Haartrockner genauso wie der Markenname „Fön“ geschrieben, sodass das Wort vom Föhnwind, der sich schon vorher mit h schrieb, unterschieden werden konnte. 1996 wurde diese Schreibweise in „Föhn“ geändert; die Wortmarke „Fön“ ist davon nicht betroffen."

    And then Google Translate obligingly spilled the use/mention distinction all over everything:

    "Before the reform of German orthography of 1996, the usual term for hairdryer was the same as the brand name "hairdryer" written so that the word of the Foehn, who wrote in earlier with h, could be distinguished. In 1996, this spelling in the "hairdryer" revised, and the word mark "hairdryer" is not affected."

  24. Faldone said,

    June 8, 2012 @ 6:27 pm

    @Ralph Hickok:

    This reminds of an old joke about a junior Army officer who had mistakenly initialed document. He was told by his superior to erase his initials and initial the erasure.

    Not necessarily a joke. I did this one time. I had crossed out something in a document and initialed it. Turns out I shouldn't have crossed it out, so I crossed out my initials and initialed that crossing out. And I was only a second class petty officer.

  25. languagehat said,

    June 9, 2012 @ 6:36 pm

    I wonder whether this pattern has increased in frequency recently. It certain has been around for a while — and indeed it strikes me as somewhat antique in tone

    You astonish me. I understand the recency illusion, but I assure you the pattern has increased dramatically in frequency. There may be a couple of musty 19th-century examples, but it is ubiquitous on the radio and in newspapers these days; can you really not have noticed it? I am, however, heartened by my fellow fossils in this thread who cling to the antiquated use of "might have."

  26. Pete Schult said,

    June 10, 2012 @ 11:01 am

    Clearly, noindentin's too good for that editor.

  27. Robert Coren said,

    June 10, 2012 @ 10:09 pm

    D. O.: Rarely.

RSS feed for comments on this post