January 14, 2006

Forensic linguistics, the Unabomber, and the etymological fallacy

It's been noted here at Language Log that mass-media reporting on linguistic topics very often turns out to be frustratingly simplistic or misleading. But the truth is, it's difficult to get journalists interested in writing about linguistics at all. Despite the success of Steven Pinker in popularizing cognitive linguistics and Deborah Tannen in doing the same for gender-based sociolinguistics, most research by linguists remains resolutely unsexy. (American dialectologists and lexicographers find that the only sure-fire way to get mentioned in the press is to anoint a Word of the Year — and if that selection sparks a phony feud, all the better!)

But one subdiscipline that seems tailor-made for media attention is forensic linguistics, the application of linguistic analysis in legal settings, such as criminal casework. The Washington Times, reporting on the field in its Jan. 12 edition, went with the obvious headline, "CSI: Language analysis unit." Hey, if forensic anthropology can get its own network TV show, why not forensic linguistics?

The article touches on the forensic analysis of academic scholars such as Roger Shuy, as well as work done within the FBI. James R. Fitzgerald, the acting chief of the FBI's Behavioral Analysis Unit-1 and a longtime Bureau analyst, spoke of perhaps the most famous application of forensic linguistics in a U.S. criminal case:

[Fitzgerald] recalls how a transposition of verbs in the manifesto written by the Unabomber helped lead to a closer identification of Ted Kaczynski in April 1996.
The latter used the phrase "You can't eat your cake and have it, too," instead of the usual form, which is "You can't have your cake and eat it, too." Like most people, Mr. Fitzgerald thought Kaczynski had made a mistake. But examination of other letters by him contained a similar feature, which, Mr. Fitzgerald says, "is actually a traditionally middle English way of using the term. He technically had it right and the rest of us had it wrong. It was one of the big clues that allowed us to make the rest of the comparison and submit a report to the judge who signed off on a search warrant."

There are a few problems with this account. First, by focusing strictly on forensic linguistics, the article glosses over the role of David Kaczynski, the brother of the Unabomber. It was David who first made the realization that the appearance of "you can't eat your cake and have it too" in the Unabomber manifesto might be an indication of the writer's true identity. [See Update #3 below.] Fitzgerald has elsewhere discussed how David Kaczynski's call to the FBI set the identification of the Unabomber in motion. Following David's hunch, Fitzgerald's team of agents and analysts made a more systematic comparison of the Manifesto with letters written by Ted Kaczynski to his brother and mother. The idiosyncratic use of the "cake" expression, among other stylistic evidence presented in the FBI's affidavit, was enough to convince a judge to issue a search warrant for Kaczynski's cabin in Montana. (See the abstract from a paper presented by Fitzgerald at the 2001 conference of the International Association of Forensic Linguistics.)

But what of Fitzgerald's assertion that Kaczynski's particular usage of the "cake" phrase is "actually a traditionally middle English way of using the term"? Well, the "eat your cake and have it" ordering is indeed older than "have your cake and eat it," though its first dating of 1562 (in John Heywood's A Dialogue Conteynyng Prouerbes and Epigrammes) only makes it Early Modern English, not Middle English. But beyond that nitpick, Fitzgerald's claim that Kaczynski "technically had it right and the rest of us had it wrong" is a clear variant of the etymological fallacy frequently observed by Arnold Zwicky and others (see here, here, and here). As with "could care less" developing from "couldn't care less," it's often claimed that the historically later idiom is less "logical" and therefore incorrect.

But does "you can't have your cake and eat it" really lack the inherent logicality of "you can't eat your cake and have it"? Only if you consider the ordering of the two conjoined verb phrases to imply sequentiality: you can't eat your cake and then (still) have it, but you can have your cake and then eat it. On the other hand, if the and conjoining the VPs implies simultaneity of action rather than sequentiality, then neither version is more "logical" than the other: cake-eating and cake-having are mutually exclusive activities, regardless of the syntactic ordering.

Fitzgerald seems to suggest that Kaczynski's "correct" use of the idiomatic phrase helped guide FBI profilers into looking for an exacting academic type (rather than a mere raving crank), one who knows how to use the "right" language that ordinary folks get "wrong." But of course it's only "wrong" in the sequential-and (rather than simultaneous-and) way of thinking. According to this article, "eat your cake and have it" was also the ordering that Kaczynski's mother used (probably another reason why his brother spotted it in the Manifesto). If that's the case, then it's understandable why he would have grown up scorning the "have your cake and eat it" ordering, especially if his education at top universities (Harvard and Michigan) reinforced an elitist view of language use. The FBI thought they were looking for a paragon of linguistic propriety, when they were actually just looking for a pedant.

Finally, I should note that the "wrong" version of the expression has been around for 180 years or so, at least in American usage. A search on the American Periodical Series and the Making of America databases finds the have-eat ordering in use from 1827, and firmly established by the mid-19th century:

North American Review, July 1827, p. 116
This may have its advantages, but how will he contrive to live below the common standard and above it at the same time? He cannot both have his cake and eat it.

Tennessee Farmer, Feb. 1837, p. 2
We beg of them to look about the River Towns for Farmers who will join them in getting up a sugar refinery, and in that way falsify the old proverb, which says "you cannot have your cake and eat your cake."

Cincinnati Weekly Herald, Nov. 13, 1844, p. 9
If your Jewish creed be right, you are wrong to deny its manifest deduction. If your Jewish creed be wrong, you are right in wishing to explain it away. But you cannot have your cake and eat it, too.

North American Review, Apr. 1848, p. 371
The reading public cannot have its cake and eat it too, still less can it have the cake which it ate two thousand years ago.

Daguerreotype, May 20, 1848, p. 289
The experiment will end in the discovery that "you cannot have your cake and eat your cake."

Debow's Review, Oct.-Nov. 1848, p. 271
Unfortunately, we cannot have our cake and eat it too.

United States Democratic Review, July 1849, p. 82
It would, indeed, allow the stockholders to "have their cake, and eat it, too."

The earliest example, "He cannot both have his cake and eat it," is helped along by the use of both — which, as Michael Quinion notes, assists the reader with the simultaneous-and reading. Later examples from the 1840s onwards simply append too at the end of the expression to imply simultaneity, and this remains an overwhelmingly common phrasing. But this still doesn't seem to satisfy those who consider the sequential-and reading to be somehow more "logical." In fact, Kaczynski said "you can't eat your cake and have it too" in his manifesto, so the presence of too is apparently not sufficient to establish the simultaneous sense of and for those who are committed to the sequential version.

(I certainly don't mean to tar all sequential-and types with the same brush as Ted Kaczynski. But perhaps he's a cautionary tale for what can happen when narrow-minded pedantry goes unchecked!)

[Update #1: Early American Newspapers supplies an earlier variant with keep-eat rather than have-eat in a verse entitled "Guillotina for 1797," first published on Jan. 1, 1797 in the Connecticut Courant and subsequently appearing in other papers (Chelsea Courier, Jan. 18, 1797; Providence Gazette, Feb. 4, 1797):

Thus greedy boys would gladly treat it,
Could they but keep their cake and eat it.

Here the exigencies of verse dictate the ordering, but this example still establishes that the simultaneous-and reading was already available by the late 18th century.]

[Update #2: Richard Mason takes issue with my assertion that "cake-eating and cake-having are mutually exclusive activities, regardless of the syntactic ordering," noting that one "has" cake during the process of eating it. Though this is technically correct, the "having" part of the idiom seems to me to imply possession over a long period of time, rather than the transient cake-having that occurs during cake-eating. (The 1797 example, interestingly enough, makes this sense more explicit by using keep instead of have.) Ultimately, however, such ruminations over logicality are irrelevant when it comes to the popular usage of crystallized idioms. Few people protest the expression head over heels to mean 'topsy-turvy,' despite the fact that its "literal" reading describes a normal, non-topsy-turvy bodily alignment.]

[Update #3: James R. Fitzgerald sent the following email:

I recently read your posting on "Language Log" regarding my interview with the Washington Times. I want to make a few clarifications.
Firstly, if David Kaczynski did know of his brother Ted's non-standardized usage of the proverb/idiom "you can't eat your cake and have it too," he never provided it to me or my colleagues on the Unabom Task Force in 1995 or 1996, or any other time. He was apparently aware of the term "cool-headed logicians," which was found in the Manifesto, and also known to have been used by Ted, as he told various investigators of its use. But, as valuable as he was to the FBI in providing his brother Ted's information to the Task Force, he never mentioned anything about the "cake" proverb/idiom. As I explained in chapter 14 of the book Profilers, I was the first one to recognize this unusual usage.
Secondly, years ago, upon doing some basic research re. this phrase, I dated the idiom to the Middle English period as, according to the Morris Dictionary of Words and Phrase Origins, it was first found in Heywood's "Proverbs" in 1546, but, "...it had been in circulation for centuries before that...." (1988: p 277). While the Modern English period is generally seen as beginning c. 1500, I felt it safe to say that its etymological roots are firmly planted in the Middle English period.
]

Posted by Benjamin Zimmer at January 14, 2006 04:14 PM