“Wretch,” it tweeted Friday morning, a momentary blip in Twitter’s vast, primordial stream.
“Wretched,” it followed. Like clockwork. Wretcheder. Wretchedest. Wretchedly.
Surely no one would read through a dictionary this way, index finger underlining each consecutive word — so many of them either mundane (“a”) or trivial (“aalii,” a Hawaiian plant). And yet, since @everyword revved up in the fall of 2007, the account has attracted more than 92,000 followers and inspired a wave of copycats and spin-offs. The man behind @everyword, a programmer and New York University professor named Adam Parrish, suspects his creation is the most widely read piece of conceptual literature in existence.
“Just the fact that it has so many followers is a constant surprise,” Parrish said from his home in Brooklyn. “No one’s ever accused me of being accessible.”
Parrish, who is 32, has always thought of computers in a way most do not. As a child in Centerville, Utah, he pored over the programming manuals that came with his family’s Tandy TRS-80 Color Computer 2, intrigued, he says, by the idea that a machine could do his bidding. During high school, he completed a two-year computer science program at a local community college before going on to get a four-year degree in linguistics. Now Parrish plans to put two titles on his business card. First, programmer. Second, poet.
Of course, to people unfamiliar with Twitter — or the longer, more robust history of creative computing — it’s hard to see the poetry. But as @everyword’s nearly 100,000 followers suggest, there’s something oddly, compellingly hypnotic about the perpetual drip of the account.
The words, divorced from any and all context, seem to take on lives of their own: people project additional meanings and connotations onto them, retweeting and tweeting back to the account, even though they know it can’t understand. Stefan Beckett, a social media editor at New York magazine, has followed @everyword since 2012 and has tweeted at it at least a dozen times; in that time, he says, the account has become something of a metronome — and an inside joke — to the people who follow it.
“I guess there’s comfort in having a perpetual, predictable presence in your Twitter stream that pops up every half hour, regardless of all the other noise that’s going on,” he said by e-mail.
Sometimes, @everyword’s random discharges coincide perfectly with events in the news. It reached “woman” just as Jill Abramson’s ouster from the New York Times leaked onto Twitter. Other times, the humdrum words — blended into a user’s usual stream of tweets from friends and celebrities and news organizations — create juxtapositions, little threads of association and whispered hints of narrative.
That’s part of why Parrish dismisses critics who say the bot he created simply tweets “contextless words” — just as he dismisses the thousands of people who have, over the years, accused @everyword of tweeting words that aren’t actually words. The bot, contrary its aspirations, is just using a script to pull text from a list of words Parrish found online in 2007. (He’s no longer even sure where the list came from.) As such, @everyword misses words. It also appears to make some up.
“But that’s the point,” Parrish said, adding that it’s a commentary on language, specifically language online, in its endless, imaginative mutation. “There’s an implied commentary there — it’s satirical in that it’s trying to do something impossible.”
That type of abstraction may not resonate with many of @everyword’s followers, who are just hanging around to retweet “sex,” “weed” and “vagina” — at this point, the account’s most-retweeted words. But the idea of using Twitter as a medium for serious art and social commentary has increasingly caught on with a ragtag group of conceptual writers, generative poets, and performance artists — to such a degree, in fact, that the “botmakers” gathered, in person and by video-chat, at a summit in Massachusetts last year.
“In the past year, the number of botmakers has really increased — and the types of people making bots have really diversified,” said Darius Kazemi, the host of the summit and the creator of more than two dozen Twitter bots himself. “A year ago I think you’d be hard-pressed to find a botmaker who was not a white dude programmer in his 20s or early-30s. But that’s completely changing, and it’s a really exciting, encouraging trend.”
The diversity of the botmakers means their projects span a range, as well. Today’s bots can do everything from imitate the affectations of a teenage girl (@oliviataters) to mash-up news headlines (@twoheadlines) to tweet accidental axioms on the human condition (@horse_ebooks, before its human acquisition). Many of the botmakers have been inspired by @everyword, Kazemi says, as well as the older, broader heritage of “creative computing” — an experimental field that dates to the 1960s, when a program called ELIZA, developed at MIT, began simulating human conversation.
Since then, artful programming has let people in chatrooms communicate with bots as if they were other people — and even communicate with other people through the intermediary of bots. The Lonely Project, by the Philadelphia botmaker Lauren Hallden, anonymizes tweets from people who say that they’re lonely, and invites readers to send them a comforting tweet, also anonymized through the bot.
“Someone over here wanted you to know that you’re not alone,” the tweet reads. It’s one of those rare instances when computer code and social media racket coalesce into something like magic.
Still, botmakers have had to contend with the skepticism of those who don’t — for lack of a better term — “get” what they’re doing. Popular culture frequently associates bots with spammers or comedians, not artists and poets. Parrish, for one, is aggravated by the frequent suggestion that @everyword is somehow autonomous and self-sustaining, as if he didn’t write the code behind it.
“People are really eager to attribute agency to computers,” he said. (He suspects it has something to do with the whole dystopian narrative of computers replacing humans.) “But I worked really hard to make that program. Would you say a typewriter wrote William Carlos Williams’s poems? Or a paintbrush made Jackson Pollock’s paintings? [Programming is] just a procedure to make art.”
Ultimately, of course, Parrish — not his program — will decide how @everyword ends. When the script reaches the last word in Parrish’s vocabulary list (he wouldn’t say what that was; that would spoil the surprise), it will look for the next word, find nothing, and go silent. A particularly dedicated contingent of @everyword followers, what Parrish calls the “doomsday cult,” have already begun dreading that day, joking about possible outcomes. Will Y2K strike? Will the apocalypse come? Will Parrish just start the whole thing over at “a,” or move on to another language, or let it all fade away for good?
Every half-hour @everyword is getting closer to its end. And Parrish honestly doesn’t know what will happen then.
“I haven’t decided yet,” he sighed. “I know that sounds irresponsible, since a lot of people want to know. It’s on track to end the weekend of June 6 … so I guess I have two weeks.”
(Due to a transcription error, an earlier version of this story said Edgar Allan Poe when it should have said William Carlos Williams. And due to an editing error, that same version misspelled Edgar Allan Poe.)