August 01, 2007

Cousin of eggcorn

Over on the American Dialect Society mailing list, we've been looking at the verb troll, in meanings similar to the verb trawl, and in passing the Christmas carol "Deck the Halls" was mentioned (as irrelevant to the topic), because of its line

Troll the ancient Yuletide carol

Beverly Flanigan reported that she'd always heard the line with trill.  I immediately trotted out the relevant OED2 subentry (which had cites for this use of troll from the 16th century through 1977) and noted that {"trill the ancient Yuletide"} got only one Google webhit, while the version with "troll" got 3,590.  Trill is clearly a reshaping; although there's an OED entry for the singing sense of troll, few people these days are likely to have encountered this sense anywhere except in "Deck the Halls", so that it's not surprising that some people have altered the verb to something that's recognizably musical.

What we have here is a cousin of the eggcorn, a misquotation that improves a line by replacing an archaic or rare word by a phonologically similar word that makes sense in context.

(This is entirely beside the point, but I can't hear any part of "Deck the Halls" without calling up the Pogo version, "Deck Us All With Boston Charlie", in which the counterpart to the troll line above is

Trolley Molly don't love Harold

Now I've probably given many of you an earworm.)

Here's the OED's entry for the transitive musical verb troll:

10. a. trans. To sing (something) in the manner of a round or catch; to sing in a full, rolling voice; to chant merrily or jovially.

... Perh. originally fig. from 6 = to sing in succession, as a round or catch (each line being as it were passed on to the next singer).

The speculation about its source refers to an entry for a now-obsolete sense of troll:

6. trans. To cause to pass from one to another, hand round among the company present; esp. in phrase to troll the bowl.

As far as I'm concerned, sense 10 is virtually obsolete itself.  I'm a bit surprised that more people haven't "fixed" the Christmas carol by shifting to trill.

[Added: or, as Thomas Thurman points out to me, to toll, treating the carol like a bell.  I got 10 webhits for {"toll the ancient Yuletide"}, including this one, from Yahoo Answers, where one helpful poster explains that "it is toll... not troll and it means to tell over and over" and another that "you mean Toll the ancient yuletide carol. it means to say somthing over and over again".  Such responses are very much like the ones you often get from people defending garden-variety eggcorns.  (No hits on {"tell the ancient Yuletide"}, alas.  But Will Fitzgerald tells me that some people have gone one step further and replaced troll by sing; there's a "sing the ancient Yuletide carol" version by the Carpenters.)]

[Whimsical addendum 8/2/07: My correspondents have been wondering if someone has yet taken the step to syntactic reanalysis, as "Troll, the ancient Yuletide carol", like "Olive, the other reindeer" (Andy Hollenbeck) and "Gladly, the cross-eyed bear" (Larry Horn).]

In any case, eggcornesque misquotations are not uncommon, though there seems to be no generally accepted name for this specific type of misquotation.  Two well-known cases:

"Once more unto the breach" altered to "into the breach"

"All that glisters is not gold" altered to "glitters" or "glistens"

"Unto" has a slight edge over "into" in the first quotation (13,000 to 96,900 raw webhits), but "glitters" has definitely won the day over "glisters" in the second (258,000 to 11,500, with "glistens" getting a mere 2,550).

Eggcornesque misquotation can be seen as arising from yet another type of conflict between Faithfulness (in this case, preserve the wording of the original) and Well-Formedness (in this case, make the word choice appropriate to modern English), with Well-Formedness (WF) winning over Faithfulness (Faith) in the misquotation. 

I first talked explicitly on Language Log about the conflict between the two principles Faith and WF in a discussion of the conventions of punctuation ("Dubious question marks"), with a side excursion into spelling conventions, in particular British Labour vs. American Labor.  A conflict arises when material printed according to one set of conventions is quoted in places where a different set of conventions is in force: Faith says to reproduce the original, WF says to convert it to follow the local conventions.

As I said in that posting,

The larger point -- the conflict between faithfulness and well-formedness in linguistic mention -- is a gigantic one.  I originally started a Language Log posting on the topic back during the discussion of taboo words in titles of books and movies, but it quickly bloated up horribly.

Suppose you want to refer to Harry Frankfurt's 2005 book that was on the best-seller lists for many weeks.  (See the Language Log posting here, with links back to earlier Frankfurt-related postings.)  Faith says to cite it as On Bullshit, but depending on who you are and what context you're writing (or talking) in, local conventions of modesty (a species of WF) might tell you to avoid the taboo word in one way or another.

More recently, I posted about Faith confronting WF in the spelling of English plurals.  What is the plural of the common noun ducky -- duckys (Faith) or duckies (WF)?  And what is the plural of the proper noun Germany -- Germanys (Faith) or Germanies (WF)?  Both versions occur (in both cases), and frequently.

I'll have more to say about Faith vs. WF, with several new types of examples, in a while.  For the moment, I'll just point out that eggcornesque misquotations seem to illustrate another type of conflict between the principles.

zwicky at-sign csli period stanford period  edu

Posted by Arnold Zwicky at August 1, 2007 02:42 PM