"Peace talks" didn't exist. Civil War soldiers weren't named Kevin. There's a lot to learn from the subtle anachronisms of Tony Kushner's Oscar-nominated script.
Audiences and critics have lauded the Lincoln, which this morning received an impressive 12 Oscar nominations, for its dedication to accuracy. From Abraham Lincoln's high voice to Thaddeus Stevens's black mistress, Steven Spielberg's movie shows a commitment to reproducing some of the less-known aspects of the past.
But when playwright Tony Kushner said in an interview with NPR that he checked every word in his script that sounded out of place, it piqued my curiosity. As a historian working with huge textual databases, I've noticed that even the historians, linguists, and writers most invested in the past can be blind to the many ways language gradually changes over time. Last spring, I wrote an article for The Atlantic about the many minor anachronisms in Mad Men. Given that Lincoln is set a full century earlier, I was pretty confident there was no way Kushner could have succeeded. Lincoln must be full of mistakes. Not that that would necessarily be a bad thing. Kushner is a master of the English language, and Spielberg has been forthright about both striving for accuracy and his studious departures from slavish reproduction. Unlike some other historical dramas, like Django Unchained or Hyde Park on Hudson, Lincoln seems to really care about giving its audience an authentic picture of the past. If we want to know to know what the limits of historical accuracy are, there's no better place to look than Lincoln.
I'm not the first to look at the historical accuracy of Lincoln's language. But by using massive databases of digital texts—in particular, the Ngrams corpus that Google created in collaboration with the Cultural Observatory at Harvard, where I have a fellowship—I can do it very comprehensively. Kushner relied on his ears to know when to look up a particular word. I lack that sensitivity, so instead I have a computer program that can tackle the problem more crudely: It simply checks every single word and phrase of up to three words (in Lincoln, there are 15,000 of them) to flag places where the script seems to be departing from language published in books. This "anachronism machine" produces dozens of potential leads I can track down in dictionaries, old newspapers, and other sources.Using computers for tasks like this is useful because it gives a completely different perspective. The statistics can help uncover shifts in American language and culture over the last century and a half that no one has noticed—although we still have to decide what they mean.
It turns out Kushner didn't reach for that OED as much as he could have. Mary Todd complains about Thaddeus Stevens's "prosecutorial" interest in her accounts, but the word wasn't used at all until 1934, and not widely until Watergate. As others have noted, everyone said "sneaked," not "snuck," until the 20th century, and the "barrage" of artillery Edwin Stanton plans for Wilmington only entered English around 1900. And the "bipartisanship" at the heart of the movie's narrative? The dictionary says the term only entered the language in 1909, although I've found a few occurrences from the 1890s.
What's really interesting about having scores of errors is that they let us see what kinds of mistakes a historical drama makes. Many are of the sort that only linguists would find interesting: extremely rare noun forms for verbs ("abstention" and "frustrating") or verbal forms of nouns (railroads "switch" in the 1860s, but not much else does), and several misused prepositions (changes to instead of changes in, support for instead of support to, and so forth).
But other anachronisms can remind us how diverse the sources of modern English are. When Congressman Ashley frets that the amendment is "absolutely guaranteed" to lose, he's speaking in the language of late 19th-century newspaper advertisements, not mid-century politics. Every war seems to create hundreds of new words, and Kushner can't avoid them: Two different characters talk about "signing up" for the army, which comes into the vocabulary during World War I. Before the Boer war, Mary Todd would almost certainly have worried about sharpshooters, not about "snipers." Even events "overseas" (another Mary Todd anachronism) shift the vocabulary of the everyday. I was surprised when the program tagged a soldier named "Kevin" as extremely unlikely, but sure enough, the Draft Registration records from 1863-65 show only a single "Kevin" pulled into the entire Union army, compared to many thousands of "Georges" and "Michaels." According to the Social Security Administration, it wasn't until 1912—seven years after Sinn Fein was founded, when trends in Irish nationalism could spread to the United States—that more than five American families a year named a child after Saint Cóemgen.
No one will be surprised that swears and slang have shifted. Even aside from four-letter words, which are extremely hard to track through print, milder curses like "get the hell out" and "Jeez" are almost certainly out of place. There shouldn't be three characters who say "Yeah" (which most dictionaries don't have until 1905), for example. A more surprising counterpart, though, is that the language of politeness has changed in ways that we tend not to notice; a frequent category of errors in the script, apparently, are pat phrases like "Nice to meet you," "Could you please," and "Good to see you." Same goes for priggish language: Almost no printed books use George Pendleton's phrase "highly unusual" until 1900 or so.