October 27, 2005

Artifacts of the spellchecker age

The New York Times has yet to issue a correction for the joke-ruining error in its Oct. 25 review of "The Colbert Report." This is somewhat surprising, since the Times prides itself on eventually rectifying even the most minuscule of typos that slip by the copy editors. According to Slate media critic Jack Shafer, there has been a "corrections culture" at the Times since the '70s, which "seems to revel in correcting every misspelling, transposed digit, historical inaccuracy, and boner." (The Times even authorized a book collecting its most amusing corrections, called Kill Duck Before Serving.)

But another recent error that did get corrected sheds some unexpected light on the situation.

In the Oct. 26 edition, we find this doozy of a correction for a sports article by Ray Glier:

Because of an editing error, a sports article in some copies on Sunday about the University of Alabama's 6-3 football victory over the University of Tennessee misstated the given name of a linebacker who is a leader of the Alabama defense. He is DeMeco Ryans, not Demerol.

Ouch! Warren St. John, one of Glier's fellow sports reporters at the Times, saw the correction and had this to say on the blog for his Rammer Jammer Yellow Hammer site:

All rightee then.

While we're at it, we've been meaning to run the following correction for a while: a previous post on the RJYH blog misstated the name of Alabama's head coach. It is Mike Shula, not God-I'm-dying-for-a-bourbon-and-water Shula.

How could Glier possibly have come up with Demerol instead of DeMeco? Could it have been some inside joke about the narcotic qualities of Alabama's defensive line? I doubt it. I think Glier innocuously wrote DeMeco in his story, but then a nefarious force changed the spelling for him: a runaway spellchecker. If you type DeMeco in a Microsoft Word document, you'll get the telltale squiggly red underline that indicates the word is not in the custom dictionary. And if you ask for a suggestion, the very first alternative provided is none other than Demerol.

So what about Alessandra Stanley's goof, replacing Stephen Colbert's sublimely silly truthiness with the pedestrian trustiness? I had assumed the error was a sort of anticipatory assimilation, since the word trust appears later in the same paragraph. But that may just have been a coincidence. If you type truthiness in MS Word, sure enough you're given trustiness as the first suggestion (followed by trashiness, frothiness, and trotlines, the last of which refers to a type of fishing line).

We don't know if Glier and Stanley applied the wayward spellcheckers themselves or if some intervening copy editor is to blame. Either way, this sort of spellchecking artifact is all too common these days, as anyone who has had to slog through college term papers can attest.

Some of the substitutions are rather startling. Back in 1996, this example was noted by a contributor to the Usenet newsgroup alt.usage.english:

This happened a few weeks ago to the menu of a well-to-do restaurant here in San Francisco.  The menu was spell-checked, printed, and a copy displayed in the window of the restaurant (as is the custom here). Nobody noticed that the spell-checker turned "warmed spring salad greens with prosciuto" into "warmed spring salad greens with prostitutes."

I'm guessing that the restaurant menu actually had singular prostitute in place of the intended prosciutto (or prosciuto, if the writer missed the extra t) — amazingly enough, the custom dictionary in my copy of MS Word accompanying Office XP still doesn't recognize the spiced Italian ham and suggests prostitute instead. (There's a Sopranos joke in there somewhere.) Others have fallen prey to the same unfortunate replacement, as in this recipe appearing on a message board for Italian food:

Crumble bread sticks into a mixing bowl. Cover with warm water. Let soak for 2 to 3 minutes or until soft. Drain. Stir in prostitute, provolone, pine nuts, 1/4 cup oil, parsley, salt, and pepper. Set aside.

Some spellchecker artifacts only show up when a particular typo is made. In another case noted on alt.usage.english, the misspelling of acquainted as aquainted has caused some spellcheckers to suggest aquatinted instead. (That word, by the way, refers to etchings made using aquatint, a process that makes a print resemble a water color.) Thankfully, it appears that MS Word has fixed this one, as aquatinted now comes in second place to acquainted in the list of suggestions. But the damage has been done, as evidenced by thousands of Googlehits. (Another example of this kind showed up not too long ago on the Eggcorn Databaseamature, a misspelling of amateur, is often transformed by spellcheckers into armature.)

One very odd type of substitution started popping up a few years ago among users of Yahoo Mail. If an email in HTML format was sent to a Yahoo address and contained the string eval, it would mysteriously get changed to review when the message was received. So medieval became medireview, retrieval became retrireview, primeval became primreview, and so forth. (In French messages, the word for horse, cheval, would become chreview.) It turned out that eval was one of a number of strings that Yahoo's security filter automatically replaced in order to prevent cross-site scripting attacks. (Bizarrely, it also replaced mocha with espresso and expression with statement.) This was far more insidious than spellchecker substitutions, as the replacements were made automatically, without the user's knowledge. Yahoo eventually changed its security filter, but once again the Googlehits live on. Future generations will no doubt chuckle at our technologically medireview times.

[Update, 11/1/05: The Times has issued a correction for the truthiness/trustiness confusion. The spellchecker remained blameless.]

Posted by Benjamin Zimmer at October 27, 2005 01:00 AM