The internet may be a gaping maw that feeds on snark and chaos, but it does have at least one rule: Poe’s law, which holds that without a clear indicator, a sarcastic remark will inevitably be taken by some people as sincere. It’s a rule the Onion knows a thing or two about. So does, well, pretty much any human who’s ever spent a meaningful amount of time on the internet.
It’s funny, though — in a lot of ways, if you think about it, the internet shouldn’t be a place where sarcasm thrives so fully. Sarcasm done well is subtle, intuitive, pointed; online, most of the signals we use to express it just don’t exist.
When I’m talking to you in person, I can raise my eyebrows, roll my eyes, emphasize certain words, or otherwise alter my face and my tone in any number of ways to make sure the point comes across. “Generally, you can see slower speech rates, longer syllables, more pauses,” says Eleni Kapogianni, a linguist at the University of Kent who studies irony. But when I’m typing into the void, all I have are the words and the hope that you’ll get what I’m trying to say; all the other tools I’d otherwise have at my disposal are useless.
“What’s interesting,” Kapogianni says, “is that in computer-mediated communication, you would expect that because there aren’t many opportunities to signal irony and provide cues — like gestures or facial expressions — you would expect that people avoid it.”
But they don’t. The internet speaks for itself on this point, but research also backs it up: In 2004, Stanford communications professor Jeff Hancock published a study in the Journal of Language in Social Psychology suggesting that people may use sarcasm more frequently online than they do in face-to-face interaction. Sarcasm is alive and well online; that much is obvious. (A good, withering no kidding seems appropriate here.) More subtle, though, are the ways in which it’s changed as it made the leap from spoken conversation to social media.
Before we go any further, let’s define the terms. Kapogianni, who earlier this year published a paper on the linguistic elements of irony, explains that verbal irony is built on contextual contrast, or a separation between what you say and what you mean. Sarcasm is the same concept with a nastier bent: verbal irony deployed particularly for negative uses. “There needs to be a contrast between, for example, the ideas of the person who’s being sarcastic and the person he’s addressing, or a contrast between reality and what the actual experiences were,” she says.
Online, that kind of context can often be hard to come by. If I tweet, for example, that I thought Obama’s performance was so great at the White House Correspondent’s Dinner on Saturday, and you happen to stumble upon it, how — apart from already knowing what I think of Larry Wilmore — are you supposed to know if I actually thought he was great, or if that was my way of trashing his routine?
“There’s no single cue. There’s no Pinocchio’s nose of sarcasm,” Hancock says. And the most foolproof ways to circumvent Poe’s law — things like hashtags or emoji — can also defeat the purpose of sarcasm in the first place. “When we use sarcasm, you would never in real life say ‘sarcasm,’” Kapogianni says. But when people mark their sarcastic tweets with hashtags or emoji, “they make it absolutely explicit, which sort of goes against this general idea that sarcasm is always implicit.”
The hashtag has also stymied attempts to better catalogue and understand online sarcasm. Last year, two researchers published a paper on how to train a computer to detect sarcasm on Twitter, with one major caveat: Because the system taught itself on tweets marked with #sarcasm, it didn’t learn to predict subtler examples. “We were good at detecting whether a tweet contained the hashtag sarcasm, but that doesn’t mean we’ve solved the problem of recognizing sarcasm in general,” says co-author Noah Smith, a computer-science professor at the University of Washington. “The bottom line is, it’s really hard to find positive and negative examples of sarcasm and be confident that that’s what they are.”
But there are a few patterns that have emerged. In a column for the Toast last year, linguist Gretchen McCulloch explained some of the less obvious ways people signify sarcasm online: funky punctuation (putting something “in quotes,” or between ~*~tildes and asterisks~*~), lack of punctuation, or deliberately misspelling words (she invoked the Twitter account Women Against Feminism as an example of this last one). And both Smith and Hancock mention intensifiers — words like super, too, so, really. (Smith also found that being an unverified user, male, and from a U.S. time zone were all strong predictors of sarcasm on Twitter — which, as anyone who’s spent any significant amount of time on Twitter can attest, is shocking.)
How we use sarcasm online is one question; another is why. In some contexts, sarcasm can be used as a bonding tool, or to create what Hancock calls “in-groupness” — there are people that are in on the joke, and then there’s everyone else. “You certainly see cases where people can tease each other and be sarcastic with each other, and that leads to closeness. And the reason for that is, it indicates, “‘I think you and I know each other well enough that I can say this thing and have you know what I mean,’” he says. “A lot of people that are quite charming will use sarcasm with another person very early [in the relationship], and it demonstrates, ‘Hey, we’re on the same page, right?’”
On Twitter, though, that type of in-groupness is harder to come by, or at the very least more blurry; it’s one person tweeting to tens or hundreds or thousands of people they may or may not know, and attempts to foster the sense of an in-group can backfire. If I sarcastically tweet an enthusiastic endorsement for a presidential candidate, for example, I may want everyone who reads it to get the joke — my followers, in this case, are my in-group — but Poe’s law will inevitably do its thing. (Even so, there are ways to create that feeling of closeness: In Smith’s research, when people tweeted at one another, they were more likely to use sarcasm if their profiles showed that they had interests in common.)
Complicating these online interactions is the fact that in real life, we tend to use sarcasm differently with people we’re close to. Adam Galinsky, a social psychologist at Columbia Business School, last year co-authored a paper that portrayed sarcasm as “a double-edged sword,” as he describes it: The study found that using sarcasm tended to boost people’s creativity, but it also tended to diminish the goodwill they felt toward whomever they were interacting with. The second effect had one notable exception: Sarcasm didn’t have the same negative effect on a relationship that was already strong.
“When can sarcasm be successful without creating conflict? It’s when you have a high level of trust with someone,” he says — in those cases, a sarcastic statement, even though it’s negative, can make a point without coming off as hostile. But on social media, statements aren’t necessarily tailored to one specific audience with a specific level of familiarity — when you send a tweet, you’re sending it to friends and strangers alike, meaning the mitigating effect of closeness is often lost. Something intended as gentler sarcasm, in other words, may still come off as harsher than intended.
But Hancock argues that sarcasm on social media may be less about relationships — connecting to, or shutting out, another person — than about presenting oneself in a certain light. It’s less a social act and more a performative one. “A lot of times there, you’re signaling from an identity point of view,” Hancock says. If you tweet a sarcastic comment about a celebrity, for example, “you’re identifying with people who don’t like, say, Kim Kardashian, and you’re making fun of people that do. So even when you’re doing it into the void, to some imagined audience, you can still be signaling whose side you’re on.”
In some ways, the internet is also a lower-stakes environment. For one thing, people can choose to work behind the cover of anonymity; for another, even when they use their real names, there’s still the perception of a wider lag time between your typing the words and someone else’s reading them. “I don’t really see how other people are responding. So if I’m sarcastic with you and I see that you don’t get it, or that it’s upsetting you, I would immediately change track and speak in a different way,” Hancock says, but “in many forms of internet communication you don’t get immediate feedback.”
Except, of course, for when you do. “We know sarcasm can inspire feelings of anger and conflict,” Galinsky says, “and on social media you can imagine that anger and conflict boiling out of control.”
Yep. Imagine that.