The Twittering Machine

by Richard Seymour

  If this were true, however, it would make trolls bizarre and incomprehensible. If they truly attack their victims on the basis of no discernible norm or value, then they have distilled punishment to its purest, pointless essence: you are punished because you are punishable. This would be the height of superego irrationality. Yet, to the extent that the yield of trolling is the outcry of the aggrieved, trolls need there to be some sort of moral order. There have to be just enough people ‘taking things too seriously’ to keep going. An indifferent shrug means the troll has failed. The death of moral value, the supposed aim of trolls, would be the death of lulz. Further, campaigns like that against Dawson make the troll look a lot like a closeted moralist, or a vigilante. To put it another way, the troll tries to have it both ways, claiming to be both magnificently indifferent to social norms, which he transgresses for the lulz, and often at the same time a vengeful punisher: in his fantasies, both the Joker and Batman.

  The ways in which trolling and vigilantism resemble one another are not incidental. When Stranger Things actor Millie Bobby Brown quit Twitter, it followed months of harassment under the hashtag, #TakeDownMillieBobbyBrown.25 The attack began with a tweet by a user making a spurious and unsubstantiated claim of having been victimized by Brown. The user claimed that she had encountered Brown in an airport and asked for a photograph. She claimed that Brown said, ‘Only if you remove the hijab,’ and then aggressively pulled it off her head and stamped on it. There is no evidence this ever happened, and the user’s profile picture showed a white woman without a hijab. Yet the hashtag was used to spread outrageous stories about Brown, often linking the LGBT advocate to homophobic or racist ideas. The story is complicated further still by the fact that some of the misattribution of homophobic sentiment started out as a satire by gay users, whose joke was that it was wildly implausible that Brown would be homophobic. Somehow, on the Twittering Machine, the transition from pure irony to zero irony is fast and frictionless.

  It would be literally impossible to disaggregate this toxic combination of banter, punitive spite, misinformed outrage and sheer glee at someone being ‘taken down’. On the Twittering Machine, they quickly become indistinguishable. And this ambiguity, this family resemblance between trolling and witch-hunting, is part of what makes it so viral, and so deadly as weapon.

  V.

  Trolls have become the main folk devil of the internet, the monstrous metaphor for everything that is wrong with it. Perhaps it’s telling that this happened around 2010–11, just when the social industry platforms went stratospheric. Thanks in part to the spread of smartphone ownership, Twitter gained a hundred million active monthly users for the first time, whereas Facebook was close to gaining a billion monthly users.

  Since then, trolls have been blamed for everything from hate crime to sharing leaked nude images on the internet, the term metastasizing so that there can now be everything from ‘gendertrolls’ to ‘patent trolls’. Politicians often use the term to deride their social media critics, which at its most cynical works to deprive the criticism of its political substance. Previously, the role of internet folk devil was occupied by spammers, stereotypically represented as a Nigerian man trying to con a little old lady out of her savings, despite the fact that most spamming came out of the United States. Anti-spammer vigilantes often targeted Nigerian men for sexual humiliation in ways that were classically racist.26 Fittingly, since it is also a tactic of war, trolling is represented instead through reheated Cold War stereotypes about meddling Russians, in a way that serves Washington’s traditional self-image as a defender of a liberal and open internet.

  If trolling has generalized on the Twittering Machine, it is probably due to an elective affinity: that is, the practice coheres with the social patterns encouraged by the protocols of the machine. Trolling, like all manipulative communication, from marketing to military propaganda, reduces language to its effects. That is to say, it uses language in the way it does, not to persuade you of an idea but to change your behaviour. The social industry platforms have invented a form of teaching machine that uses reinforcements to induce users to respond accurately to marketing signals. In so doing, they’ve created an apparatus that can easily be gamed by trolls, who simply use it as it was designed to be used.

  This is one reason the social industry bosses, despite noisy protestations of good intentions, seem unable to do much about trolling. The machine is perfectly congruent with the tendency that Raymond Williams described, wherein the New Right sought to rebuild societies to resemble the brutal struggle for survival among states.27 These societies were nihilistic, he said, their goal ‘a willed and deliberate unknown, in which the only defining factor is advantage’. The platforms have distilled the idea of advantage into perfectly abstract metrics of attention and acclaim. Attention that is wrested most efficiently through manipulation, from the ‘thirst trap’ to ‘fake news’.

  In a way, the social industry platforms have turned John Forbes Nash’s ‘fuck you buddy’ game into a principle of interaction. In Nash’s game, played with chips and cards, each player tries to win all the chips. To win, however, they have to make temporary agreements with other players, which they ultimately go back on. They have to screw each other over. On social media, there are incentives towards short-term cooperation, ‘signal boosting’, the better to win more ‘chips’ (likes and followers). But these same incentives lead to users treating one another as raw material for their own success, and turn on one another with startling ease. They also lend themselves to a kind of hypervigilance – to what Eve Kosofsky Sedgwick calls ‘paranoid reading’.28 One of the by-products of trolling anonymity is that it is often hard to tell whether or not one is being trolled. Friendly criticism appears as flaming, queries as concern trolling, a mild joke as an attack. Trolls have always relished their ability to trigger chain reactions, and as trolling goes viral, it flows seamlessly into online vigilantism, from which it is often indistinguishable. Trolling has become generalized because a directionless war-of-all-against-all is exactly what the machine is designed for.

  But while a war against vulnerability is by definition a war against anybody, not everyone is equally vulnerable. The logic of online social Darwinism favours the dominance of the least vulnerable. When the irony-Nazi and celebrity troll Andrew Auernheimer, or ‘weev’, bombastically declared that ‘trolling is basically internet eugenics’, a way of driving the ‘filth’ and ‘retards’ off the internet and into the ‘oven’, he was giving a programmatic expression to tendencies already present in trolling.29 There is no clear correlation between trolling and support for right-wing politics. If anything, many right-wingers seem to have adapted the cultural style of trolls for their own purposes. But even among the majority of trolls who claim to be ‘equal opportunity offenders’, their victim choice betrays a tacit morality.

  Trolls, in the subcultural sense, are overwhelmingly white men from anglophone and Nordic countries, who disproportionately attack women, queer and transgender people, black people and the poor.30 Their vaunted detachment performs a familiar white-male fantasy of ironclad superiority.31 When Anonymous trolls first donned V for Vendetta masks for their campaign against the Scientologists, they could hardly have better demonstrated this fantasy. On the trolling message boards, it doesn’t pay to admit to being a woman, unless one is prepared to post naked photographs or ‘camwhore’. As Jamie Bartlett describes, one woman who did agree to ‘camwhore’ for the ‘/b/tards’ inadvertently supplied enough information to enable the trolls to track her down.32 While she watched helplessly, they found her location, contact details, Facebook and Twitter accounts and university. They doxed her and shared her nude images with her relatives. One of the trolls called her and reported that she was crying ‘like a sad sad sobbing whale’, but the ‘/b/tards’ didn’t give a shit: it was her fault for being stupid, and she deserved the consequences.

  Trolling for the lulz thus segues into ‘gendertrolling’, wherein the aim is to silence vocal women
through swarm-like harassment, rape threats, epithets like ‘cunt’ and ‘whore’, and the threat of ‘doxing’. Faced with an actual, committed ‘gendertroll’, most trolls would see it as an opportunity for some countertrolling. Anyone who took the issue that seriously would be asking for it. Nonetheless, the spontaneous ideology of trolling is masculinist, and it’s often impossible to tell the difference between a ‘real’ sexist attack and something said to provoke, for the lulz. Pew Research found that a quarter of young women have been sexually harassed, and another quarter stalked, on the internet. Danielle Citron’s study of online hate crime finds that 53 per cent of non-white women, and 45 per cent of white women, have suffered harassment. Everyday sexism is everyday psychological warfare.33

  VI.

  A Talmudic saying has it that to ‘shame another in public’ is a sin ‘akin to murder’. As if shame was something like a death sentence. Jon Ronson mentions the startling finding that 91 per cent of men and 84 per cent of women can recall at least one vivid fantasy of murdering someone.34 Almost all of these fantasies were driven by the experience of humiliation, as if the worst thing you could do to someone is to destroy their idea of themselves. Most find a way to sublimate the desire. Some people murder themselves.

  In 2006, a thirty-one-year-old Neapolitan woman, Tiziana Cantone, committed suicide by hanging.35 This, the last of several suicide attempts, followed years of public shaming over a leaked sex tape. The tape went viral and became the basis for mocking memes, sometimes printed on t-shirts or mobile phone covers. This was ‘revenge porn’, a malign form of internet celebrity. Cantone had sent the footage of herself having sex to an ex-boyfriend and some other friends on WhatsApp. And it was her ex who decided to troll her by posting the footage online. The devastating shaming visited on Cantone forced her to leave her job, change her name, move to Tuscany and fight in the courts to get the footage removed from the internet. Recognized everywhere, she was subjected to mocking skits by Italian footballers, turned into a joke by radio hosts and even denounced in smarmy terms by a politician from the Democratic Party. She was doing everything to erase all trace of herself, short of suicide. Until she committed suicide.

  Whether or not the uploading of the footage began as an attempt at trolling, the subsequent campaign of moralistic spite, the fetishistic, detached nature of the mockery and the rapidly commodified memes, quickly came to resemble trolling. As Ronson points out, this kind of lethal shaming is hardly new to the platforms. News media have often hounded people to misery or death with exuberant public shaming campaigns. In recent years, this included the smearing of the so-called ‘poverty hoaxster’, blogger Linda Tirado, the public humiliation of Australian Duncan Storrar for asking a question inconvenient to the government, and the journalist Richard Littlejohn’s hounding of trans woman Lucy Meadows until she committed suicide. Many others have been taken apart with cheerful amorality, usually without the elan or wit of the better trolls, or with a scintilla more justification. But the social industry has now greatly expanded the potential ranks of previously anonymous individuals who are susceptible to this kind of predation, as well as the ranks of potential predators. Not only that, but the way in which social media mobs form to shame an individual can provide legacy media corporations with a prefabricated story which they can quickly monetize.36

  The monstering of Justine Sacco over a tasteless joke is one of Ronson’s most telling case studies.37 Sacco had tweeted ahead of a flight to South Africa, with what she says was deliberate dark irony, ‘Going to Africa. Hope I don’t get AIDS. Just kidding. I’m white!’ With only 170 followers, she had no reason to expect widespread attention. But during her flight, Twitter exploded with rage over what was seen as an intentional and literal racist provocation rather than, as she intended, a commentary on white ignorance. The paranoid reading prevailed. As soon as she landed, Sacco was submerged by angry tweets and concerned messages from friends. The furore was then taken up by newspapers and broadcasters. The Rupert Murdoch-owned New York Post sent journalists to follow her around. Her old tweets, often deliberately tasteless, were mined by BuzzFeed. For a badly worded joke – or worded all too well, hitting its mark too surely – she lost her job and spent years in misery – tormented, perhaps above all, by how happy everyone seemed to be by her ruin.

  The schadenfreude of those looking forward to Sacco’s devastation upon landing is recognizable. Ronson sees in it his own initial ‘happy little “Oh, wow, someone is fucked.” ’ But he also draws attention to the necessary detachment of this punitive glee: ‘Whatever that pleasurable rush that overwhelms us is – group madness or whatever – nobody wants to ruin it by facing the fact that it comes with a cost.’ Whatever it is that enables social industry users and journalists to overlook the cost of their buzz-driven show trials, to refuse context with a certain invested glee and wilful philistinism, to refuse the slightest scrap of interpretative generosity, it is just as much a fetish as the ‘mask of trolling’. From one perspective, it looks like hypocrisy: you can have outrage or gleeful schadenfreude, but not both. From another perspective, just as the laughter of anons is secondary to that of the collective, the outrage of individual tweeters is secondary and vicarious. The main job of participants is to fuel the outrage of the anonymous collective. In that case, the main difference between trolls and shamers is one of emphasis. The former often mistakenly think they don’t have a moral commitment; the latter often mistakenly think they do.

  Sacco was, in a way, a small-scale troll who inadvertently provoked, in response, a mass trolling campaign. Her victimization, the fusion of trolling and vigilantism, was extraordinarily lucrative for media firms. Ronson estimates that Google alone may have made $120,000. Perhaps this collusion between troll and witch-hunter proves so extraordinarily volatile because it plays out something we already do to ourselves, intra-psychically. As if the Freudian slip or gaffe is just a way of trolling ourselves, inciting and enjoying the rage of our own internal Witchfinder General. Or, as if trolls operate on our existing unconscious dissent towards the identities and ideas we take too seriously, while online witch-hunters magnify to gigantic proportions the ways in which we are already punishing ourselves for our dissent.

  Popular internet wisdom warns, ‘Don’t feed the trolls’. A logical corollary might be, ‘Don’t feed the moralists’. They are both part of the same spiral.

  VII.

  Trolling, and the backlash against trolling, is for the most part good money. But even when it starts to cost them business, the social industry giants are consistently bad at dealing with trolls. ‘We suck at dealing with abuse and trolls on the platform,’ then Twitter CEO Dick Costolo lamented in 2015, ‘and we’ve sucked at it for years.’38

  This is an understatement: it is difficult to suck at something you’re barely trying to accomplish. Twitter’s Trust & Safety Council, responsible for protecting users, says the company should not get involved in distinguishing between good and bad speech. Twitter is ‘the free speech wing of the free speech party’, exclaimed then vice president Tony Wang. The ‘marketplace of ideas’ should sort it out, argues its user-safety chief, Del Harvey. Bad speech, she insists, is best countered with more good speech.39 Of course, since Twitter can’t distinguish good speech from bad, it can’t know that the ‘marketplace of ideas’ will promote good speech. But whatever the problem, more monetizable content is the answer.

  Still, by 2015 Twitter’s share values began to suffer as user growth stagnated in response to the nastiness of the medium. Under Costolo’s successor, Jack Dorsey, the company responded to the problem by taking an idea from Facebook. Rather than lose profitable content, they used algorithms to change user experience.40 Rather than seeing tweets in the order that they were posted, users would see tweets aligned to their tastes. Subsequently, harassment has been addressed by means of tweaks to the algorithm. It was an optimal solution. Even if it didn’t reduce bullying, it shifted the conversation and it mitigated Twitter’s long-term problem w
ith user engagement.

  Intriguingly, this solution paid no attention to user demand. The social industry bosses don’t trust that users know what they want. As Facebook’s former chief technical officer, Ben Taylor, explained: ‘Algorithmic feed was always the thing people said they didn’t want, but demonstrated they did via every conceivable metric.’41 The metrics in question are those of user engagement, which fuel the mobile advertising business. Even if users complain bitterly about being trolled and abused, as long as we stay hooked, then the metrics will say we love the system. And if trolling provokes us to engage more intensely with the machine, typing out angry replies until the early hours of the morning, that still looks like pleasurable engagement in the metrics.

  The doctrine of ‘free speech’ on the platforms is both a business doctrine and the assertion of a kind of sovereign power. When Reddit was used to circulate leaked celebrity nude images, then CEO Yishan Wong rallied strongly to ‘the ideal of free speech’.42 Reddit was not just another corporation, he insisted, but more like ‘the government of a new type of community’. And a government should exercise ‘restraint’ in its powers. But Reddit is not a constitutional republic, and Wong was misappropriating the language of ‘free speech’. On Reddit, as on almost every platform, speech is controlled. It is subject to user-engagement protocols determined by the commercial aims of the owners. In defending the ‘free speech’ of users, Wong was asserting a state-like monopoly over speech on his platform. He was defending the company’s sovereignty against challenges from governments, rival companies and citizens. The only successful challenge to this monopoly comes in the form of property law. When Reddit finally deleted the threads containing leaked celebrity nude images, it was under threat of copyright suits. But this strategy only works for those with resources. For most of us, it is completely ineffectual.43