Who's the eker this time?

« previous post | next post »

Frank Rich, "Kiss This War Goodbye", NYT 7/31/2010, writing about the Pentagon Papers:

Though the identity of The Times’s source wouldn’t eke out for several days, we knew the whistle-blower had to be Daniel Ellsberg, an intense research fellow at M.I.T. and former Robert McNamara acolyte who’d become an antiwar activist around Boston. [emphasis added]

It's clear that this is a mistake, with eke out having been substituted for leak out. The question is, what sort of mistake is it?

It might be a classical malapropism, which would mean that Mr. Rich has mis-learned or mis-remembered eke out, just as the fictional Mrs. Malaprop substituted allegory for alligator and illiterate for obliterate.

Or it might be an eggcorn, which is like a malapropism except that the substitution makes sense — often more general sense than the original did, as in old-timers' disease for Alzheimer's disease.

Both of these explanations seem unlikely to me.  Rich is a highly educated professional writer who is not a likely source of such mistakes (though all of us have a few similar oddities in our mental lexicons).  "Leak out" is such a simple and common usage that it's unlikely to cause either sort of misunderstanding. And it also seems relatively unlikely for someone to mis-learn "eke out" as a fancy way to say "leak out".

In the COCA corpus, "eke out" is reasonably common — occuring at a rate of about 0.7 per million words in magazines and newspapers — and typical examples seem unlikely to generate a "leak out" hypothesis:

… analysts expect it to eke out profit growth of only 3% a year.
The people who live along the desert shores eke out a tenuous living by fishing
Their well-meaning doctors starved them in an attempt to eke out a few more months of life.
The challenge for scientists is to eke out as much information about ancient life as they can …

Furthermore, a search of the NYT archive turns up eight earlier uses of "eke out" in columns or reviews by Frank Rich, all of them consistent with the OED's gloss "to supplement, supply the deficiencies of anything (const. with); esp. to make (resources, materials, articles of consumption, etc.) last the required time by additions, by partial use of a substitute, or by economy", from the otherwise obsolete verb eke meaning "To increase, add to, lengthen".

A more plausible theory is that "eke out" is the kind of performance error that's come to be called a "Fay/Cutler malapropism", after D.A. Fay and A. Cutler, "Malapropisms and the structure of the mental lexicon", Ling. Inq. 8(3): 505–520, 1977. These are cases where the speaker or writer knows the right word, and means to say or write the right word, but some internal wires get crossed and the wrong one comes out.

This instance of "eke out" could also be a Cupertino. Maybe Rich omitted the 'l', producing the string "eak out", which a helpful spellchecker plausibly changed to "eke out".

And there's a new sort of possibility. Maybe Rich is one of those people who are taking advantage of the fact that dictation software has now gotten quite good, as his colleague David Pogue pointed out a few days ago ("Reliable Dictation, Down to a 'T'", NYT 7/28/2010):

The accuracy is so good that you no longer have to begin by reading a four-minute training text, as in years past. I installed the software on my PC, skipped the training, and dictated one of my old columns, 1,300 words. It achieved 100 percent accuracy, even correctly nailing toughies like “LinkedIn,” “Twitterific,” “freebies” and “twentysomethings.” (It made one error, but I’m letting it off the hook for not recognizing the Web site name Bebo.)

Could Dragon Dictate recognize "wouldn't eke out for several days" in place of "wouldn't leak out for several days"?  Maybe so. And training it on Frank Rich's previous output might make this a bit more likely — his eight previous uses of "eke out" beat his five uses of "leak out".

Whatever the cause, here's the obligatory screenshot of the effect:

[Update — it occurs to me that the hyperlink on the phrase "eke out for several days" was almost certainly added by an editor. Perhaps at that stage the initial 'l' was lost, creating the context for an "eak" to "eke" Cupertino.

"Attributional abduction" strikes again.]

[Update #2, 8/2/2010 — After two days, the online version still reads "wouldn't eke out for several days". So either Frank Rich and his editor think this usage is OK, or no one has pointed the error out to either of them. I'm not sure which I think would be more surprising.]</font>


  1. Heck said,

    August 1, 2010 @ 10:07 am

    Could it be a simple spell-check error? Maybe the "L" was left off "leak" the resulting "eak out" was changed to "eke out."

  2. Heck said,

    August 1, 2010 @ 10:09 am

    Oops, somehow I missed your observation of that very thing. Sorry.

  3. MattF said,

    August 1, 2010 @ 10:38 am

    It's hard to see how even minimal copy editing could have missed this error, so I doubt that Rich is responsible for it. My guess is that, somehow, 'eke out' replaced 'leak out' in the editing process. Maybe the editor had just finished a crossword puzzle– 'eke out' is classic crosswordese.

  4. Lars Mathiesen said,

    August 1, 2010 @ 10:59 am

    Actually, I always thought "eke out a living by xxx" was the same as "glean a living by xxx" — even though I recognize the origin of the verb.

    I think your COCA examples document the same meaning drift — more "getting (small) results with hard work" than "supplementing already existing but unsatisfactory resources with hard work."

    If people had the OED sense in mind, I'd expect "analysts only expect it to eke out its profit growth to 3% a year," "The people who live along the desert shores eke out their tenuous living by fishing," and so on.

    Whether that has anything to do with what Rich wrote, is another question — it's in the wrong voice. "Nobody would glean the identity of the Times' source for several days" is fine — and then "eke out" sounds much more like the thing an investigative journalist would do. But to get from there to what appeared in the paper, you'd have to posit further mistakes. The spelling checker and dictation software theories are simpler.

  5. Brian Weatherson said,

    August 1, 2010 @ 11:15 am

    I'd be surprised if the links are really added by editors. Rich's columns are much more link-heavy than his colleagues, usually by an order of magnitude. Unless Rich has his own editor, who is particularly link-keen, the best explanation of this seems to be that Rich adds his own links.

    At the very least, it seems that someone who works on Rich columns, but not on other editorial page columns, adds the links, and I always assumed that much be Rich.

  6. Dan Lufkin said,

    August 1, 2010 @ 11:20 am

    I just tried dictating the text of Frank Rich's article in the same version of Dragon that David Pogue reported on (saying "leak" instead of "eke"). I can report that Dragon got it 100%, except for failing to capitalize the definite article of the Times.

    As a matter of fact, I'm dictating this submission this very moment. Yesterday, as a test, I dictated 985 words in 31 minutes, with 2 mistakes. One of the mistakes was "rowboat" for "robot." (Got it right this time.) Incidentally, there is internal evidence in the Dragon program data that it somehow characterizes my accent as "Inland US (Great Lakes area)". This isn't 100% accurate, but I'd still like to know how Dragon does it.

  7. Dan Lufkin said,

    August 1, 2010 @ 11:46 am

    There's one more oddity in the Rich piece that deserves comment. In the 4th paragraph from the end Rich writes:

    And they are starting to focus on the morbid reality, highlighted in the logs, of the de facto money-laundering scheme that siphons American taxpayers’ money through the Pakistan government to the Taliban, who then disperse it to kill Americans.

    On the assumption that the mot juste here would be disburse, I ran it through Dragon (which, incidentally, recognized mot juste). The 1st time I dictated it, saying disburse, disperse indeed appeared, but the next try produced disburse. This is why I always have Dragon read my dictated text aloud back to me before I send it on.

  8. mollymooly said,

    August 1, 2010 @ 12:32 pm

    The fact that the error occurs at the beginning of a hyperlink might somehow be relevant. Adding complexity to any task increases the opportunity for something to go wrong. Some clever person might be able to test the relative frequency of errors in hyperlinked and unhyperlinked portions of a text, and of similar text lacking any hyperlinks.

  9. Jan Freeman said,

    August 1, 2010 @ 1:54 pm

    At my paper, links are indeed added by the (website) editors, and a hyperlink error would show up only in the online version — has anyone looked at the paper copy? I'm out of NYT range this weekend and can't check.

  10. John Cowan said,

    August 1, 2010 @ 3:04 pm

    Eke out is itself a bit of a malapropism as used nowadays. Originally it meant 'supplement', as in Grant Allen's crack that the people of the Scilly Isles "eke out a precarious livelihood [often given as "meager living"] by taking in one another's washing." By misinterpretation of this saying or something like it, to the effect that the precarious livelihood was provided exclusively by doing each other's washing, eke out has come to be understood as 'gain with great difficulty'.

    So what's one more semantic shift?

  11. Kip said,

    August 1, 2010 @ 3:10 pm

    Maybe it's because I was always a much better student at math and science than languages and arts, but I don't see a problem with this usage of eke. Maybe I've been the victim of an eggcorn or something. I thought "eke" could mean something like "happen with very slow progress." Like someone that ekes through life–they are doing the minimum necessary, barely progressing making progress. I read "the identity of The Times’s source wouldn’t eke out for several days" to mean that the identity would slowly progress internally at the Times before finally being made public.

  12. Dan Lufkin said,

    August 1, 2010 @ 4:54 pm

    I suppose that everyone in class already knows that the meaning of eke as "additional, supplementary" is preserved in nickname (an eke-name). I dunno about "slow progress".

  13. fev said,

    August 1, 2010 @ 7:52 pm

    @Jan, it's "eke out" in the version of the national edn that we get in Motown. Anyone know if it was caught in the final?

    Cupertino issues aside, I have a hard time imagining a copy editor wandering through the august columns of the Week in Review and messing with the sacred prose at random.

  14. Xmun said,

    August 1, 2010 @ 8:29 pm

    Surely Evelyn Waugh stated the truth about "eke out" in his letter to Nancy Mitford of 8 July 1953:

    "Darling Nancy,
    You do not understand the meaning of the word 'eke'. It means to make something last longer by adding something else to it. eg. eke out butter with margarine. When you write "I might as well eke out the month in London as anywhere else", you commit a gross vulgarism. But I am glad you are coming to England to polish up your English. . . ."

  15. Xmun said,

    August 1, 2010 @ 8:42 pm

    PS Please read those double quotation marks surrounding EW's Nancy Mitford quotation as single. (I have rapped my fingers.)

  16. Nancy Friedman said,

    August 1, 2010 @ 11:48 pm

    It's been a good week for eke. "Eke out" also in Sunday's NY Times in an article about the actress Patricia Clarkson on page 5 of Arts and Leisure. This usage also struck my ear as slightly off, though not as wrong as Rich's:

    "Since her 1998 breakout role as a German actress in druggy decline in 'High Art,' Patricia Clarkson has eked out an uncommon niche: she's a character actress with a name you know."

    I'd never seen "eke out" with "niche," but a Google search turns up more than 2,900 hits for "eke out a niche." ("Carve out a niche" gets more than 13,000 hits.)

  17. marie-lucie said,

    August 2, 2010 @ 12:52 am

    The word "eke" with its two e's makes a convenient filler in crossword puzzles, where it is found with interesting definitions: "supplement" is very rare, but "make, earn, scrape, scratch, squeeze" and a few others of the same semantic range appear regularly. I don't remember anyone actually saying it, but it appears in writing too with those "modern" meanings, as in "to eke out of a tight spot".

  18. Xmun said,

    August 2, 2010 @ 1:20 am

    "It's been a good week for eke." Eek!

  19. Eiollen Cordelia said,

    August 2, 2010 @ 2:46 am

    Just ask Frank: frrich@nytimes.com

  20. Eiollen Cordelia said,

    August 2, 2010 @ 9:31 am

    2 Hypotheses:

    1) Voice recognition software artifact, not caught by NY Times eds. because it looked right on the page. (Even skilled book copy editors sometimes fail to catch this kind of error occasionally.)…OR…

    2) Freudian slip resulting from anxiety of even Times star columnists about salary negotiations.

    (hat tip to friend in NJ)

  21. Afi Konander said,

    August 2, 2010 @ 9:32 am

    2 Hypotheses:

    1) Voice recognition software artifact, not caught by NY Times eds. because it looked right on the page. (Even skilled book copy editors sometimes fail to catch this kind of error occasionally.)…OR…

    2) Freudian slip resulting from anxiety of even Times star columnists about salary negotiations.

    (hat tip to friend in NJ)

    [(myl) Dear Afi, or Eiollen, or however you prefer to be nymed,

    Please maintain a consistent pseudonymity, at least for copies of the same comment. We don't insist that you reveal your True Name, but picking a different false one for every contribution makes even the pretense of conversation difficult.]

  22. Afi Konander said,

    August 2, 2010 @ 10:20 am

    I think the most likely culprit is 'spell-check' or some automatic spelling correction program. You have no idea how many weird word substitutions my iPhone types in for me. If he typed in 'eaked' for 'leaked'. the spell-check would have rejected 'eaked' and assumed he meant 'eked', context be damned.

  23. fev said,

    August 2, 2010 @ 10:44 am

    Rich also has a piece in the current NYRB that uses "soft-peddle." I wouldn't bet too heavily against writer error in the Sunday column.

    [(myl) Bringing back a classic! A screenshot from the NYRB article:

    As each day goes by without an edit in online version of Rich's NYT column — it's still eke, I just checked — your suggestion becomes more and more plausible.]

  24. Dan Lufkin said,

    August 2, 2010 @ 1:18 pm

    Aw, give Rich a break, you guys! I think he does quite well with the language, considering the handicap of attending Harvard and working on the Crimson.

  25. Xmun said,

    August 2, 2010 @ 5:15 pm

    Quite by chance, I had to look up the word "rallonge" in my French dictionary just now. The definition:

    rallonge, s.f. Lengthening-piece, eking-piece; _rallonge de table_, leaf.

  26. Jonathan said,

    August 2, 2010 @ 9:54 pm

    Surely if the intended meaning were something happening slowly/with difficulty, it be passive "be eked out".

  27. Aki Konan said,

    August 3, 2010 @ 12:36 am

    Re: [Update #2, 8/2/2010 — After [X] days, the online version still reads "wouldn't eke out for several days". So either Frank Rich and his editor[s] think this usage is OK, or no one has pointed the error out to either of them. I'm not sure which I think would be more surprising.]

    This writer, moi, Afi Konan, did point it out to Frank in a private email and he received it. Whether his editors care to correct the online version is another thing altogether. But Frank knows of the mistake now.

  28. Aki Konan said,

    August 3, 2010 @ 1:07 am

    Note: I, like some of you here, often point out typos and "atomic typos" to
    online site editors and the original authors as well, and the
    correction is not always entered on-screen later. Why? I recently corrected
    a major booboo in the Christian Science Monitor and they editors
    corrected it. But a recent article in the Chronicle of Higher
    Education in DC carried a major blooper and I told the author and he
    said "Oops" in a reply to me but the editors of the site still have
    not corrected it, 4 weeks later. So my guess is the Times m-eke-take
    will stay up for all eternity. It's a question of manpower,
    womanpower, newsroom staff and an attitude online of we just don't
    care. Sigh. But Frank knows. What can he do? His hands are tied. The
    New York Times is not one big family. It is many separate families all
    with their own turf and they don't all work in the same building and
    their left hands often don't know what they right hands are doing.
    Welcome to cyberspace.

  29. marie-lucie said,

    August 3, 2010 @ 9:35 am

    Xmun: <i.rallonge also means "extension cord".

  30. Abe Thurtell said,

    August 3, 2010 @ 2:54 pm

    My dad, a retired journalist and lover of intentional, playful malapropisms, occasionally tells a story of a time when he was working at a small-town newspaper, and had to literally stop the presses. He had intended to use the phrase "red herring", but, as he usually does in conversation, had instead written "red heron". It made it past all the levels of editing and he only noticed it as the paper was going to press.

    You might expect a higher level of editing from the Times, but it still doesn't surprise me that this sort of thing happens occasionally; there are a lot of editors, but there's also a lot of copy to edit. Though the fact that it hasn't been changed is, as mentioned in the post, somewhat more surprising.

RSS feed for comments on this post