Lexicon Valley

Sorry, That’s Not an Emoticon in a 1648 Poem :(

Robert Herrick, not a pioneer of the emoticon :`(

“I discovered what looks to be the first emoticon!” So read an excited blog post from Levi Stahl, an editor and publicity manager at the University of Chicago Press, who had been reading the 17th-century poet Robert Herrick. Stahl came across this line in a 1648 poem entitled “To Fortune”:

    Tumble me down, and I will sit
    Upon my ruins, (smiling yet :)

Stahl’s discovery quickly made the rounds online, trumpeted in The Atlantic, The Huffington Post, Mashable, Engadget, and elsewhere. Now, it would obviously be big news to find a sideways smiley face lurking in a text published more than three centuries before the official birth of the emoticon in 1982. But a careful look at punctuation practices in the 17th century reveals that this is nothing more than a typographical red herring.

Despite the fact that, as Alan Jacobs has pointed out, Herrick’s poem did not consistently include that colon-parenthesis combination in later editions, it really does appear that way in the original 1648 publication of his collected works. Bonnie Taylor-Blake, an industrious researcher on the origins of words and phrases, shared this scan of the original, pulled from the Early English Books Online database:

But Taylor-Blake also noticed that on the very next page, in a poem entitled “To Anthea,” the colon and closing parenthesis once again appear side by side:

Would a modern-day reader like Stahl even have thought twice about this second example? Probably not, since it’s the happenstance appearance of the word “smiling” in “To Fortune” that primes an interpretation of the colon and parenthesis as a proto-emoticon. Much the same thing happened a few years ago when Jennifer 8. Lee made her own excited announcement about what looked like a winky emoticon in a New York Times transcript of an 1862 speech by Abraham Lincoln:

“… there is no precedent for your being here yourselves, (applause and laughter ;) and I offer, in justification of myself and you, that I have found nothing in the Constitution against.”

The Lincoln example is slightly more plausible, given that the first undeniable glimmers of emoticons appeared a couple of decades later. As I discussed recently in a video for the PBS Offbook series (and have written about elsewhere), typographical play imitating facial expressions goes back to an 1881 item in the humor magazine Puck, followed six years later by Ambrose Bierce’s suggestion of a jocular punctuation mark that he called a  “snigger point” (a parenthesis on its side). So clearly there had been at least a century of groundwork before Carnegie Mellon’s Scott Fahlman proposed smiley and frowny emoticons in 1982.

I still don’t think the Lincoln transcript is winking at us, and I’m even more sure that Herrick did not intend his readers to tilt their heads to see a smiling face. It’s not just in Herrick’s poetry—the colon-parenthesis combo is easily found in other 17th-century English texts.

When I read about Stahl’s discovery, I checked in with Benjamin Schmidt, a digital humanities scholar at Northeastern University who is a master at wrangling historical collections of texts known as “corpora” (that’s the plural of “corpus”). In no time at all, Schmidt conducted a quick text-mining expedition, plumbing pre-1700 texts collected on the Internet Archive’s Open Library. As is often the case with such early materials, scanning and digitizing the texts leads to numerous errors involving optical character recognition (or OCR). But Schmidt found that most instances of :) in the digitized texts were not OCR errors but were legitimate strings of punctuation – even if they seem peculiar according to modern stylistic standards.

Here are a few examples from Francis Bacon’s Sylva Sylvarum, published posthumously in 1669:




And in the same text you can even find the “winky” version with a semicolon instead of a colon:

Some writers, or perhaps more accurately some printers, were more drawn to tucking colons and semicolons before closing parentheses. In Richard Baxter’s 1653 Plain Scripture Proof of Infants Church-membership and Baptism, I count no fewer than 14 real examples of the :) punctuation pattern. (The Open Library digitization erroneously identified 7 more examples due to OCR errors. But the flipside is that there may very well be several more legitimate examples in the text that OCR failed to pick up.)

All of this should come as no surprise to connoisseurs of punctuation from the early modern era. As Keith Houston wrote in his wonderful book on the secret life of punctuation, Shady Characters, “That writers and printers suffered at one time from a tendency—or more precisely, from an irresistible, overwhelming need—to burden the written word with unwonted punctuation first became apparent in the seventeenth century.”

Houston points to the effusive punctuation found in the 1622 edition of Shakespeare’s Othello. The writer Nicholson Baker had previously noted the prevalence in the First Quarto of such combinations as the comma-dash, the semicolon-dash, and the colon-dash—which, in a 1993 New York Review of Books article, he dubbed the commash, the semi-colash, and the colash.

Baker clearly has a soft spot for such typographical anomalies, as revealed in his 1990 novel Room Temperature. Mike, the protagonist, harshly critiques his mother’s “informal punctuation” in an op-ed letter that she wrote, but then discovers some historical precedents:

[T]he fact that I had been instinctively so cruel made me double up with misery when, after I was married, I came across some sentences in Boswell that were punctuated just as hers had been. Boswell (and De Quincey, Edward Young, and others) had treated the sunken garden of a parenthetical phrase just as my mother had — as something to be prepared for and followed by the transitional rounding and softening of a comma. And such hybrids — of comma and parenthesis, or of semicolon and parenthesis, too — might at least in some cases allow for finer calibrations between phrases, subtler subordinations, irregular varieties of exuberance and magisteriality and fragile conjunction.

So let’s not engage in “presentism” and interpret sequences of colons (or semicolons) and parentheses as spectral versions of our own emoticons. Instead, let’s appreciate early-modern texts on their own terms, and celebrate the wild exuberance of their now-forgotten punctuation hybrids. That’s still worth a smiley face.