In an age where the idea of “voting with your dollar” is accepted far beyond libertarian political circles, it makes sense that we would readily accept voting as a sufficient and putatively objective tool for registering and measuring attention. Favorites, likes, and upvotes can come across as straightforward attempts to document the way we’ve spent our highly coveted attention span, displaying what we’ve “purchased” on timelines and front pages. But these efforts to numerically depict attention override the subtleties of attentional focus. Despite attention’s many possible inflections––empathy, anger, laughter, solidarity, hate, sympathy––on popular sites like Reddit, YouTube, and Hacker News it is recorded with a simple up or down vote, and any attenuating comment is rendered into an addendum.
Yik Yak, a location-based messaging service that brings together several of the major recent trends in social media apps—ephemerality (Snapchat), geographic proximity (Tinder and Grindr), and anonymity (Whisper, Secret)—also foregrounds voting in its main feature set. The app allows anonymous users within a 10-mile radius to post 200-character “yaks.” These posts are then up-and down-voted to determine their prominence within the community. Only posts from within the last day are displayed, and those with a score of –5 or below are deleted by the service.
“When we made this app, we really made it for the disenfranchised,” Brooks Buffington, told the New York Times. Buffington, who founded Yik Yak with his fraternity brother Tyler Droll, sees their app as a corrective to the (micro)celebrity culture of Twitter. On Yik Yak, they claim, localized content wins on merit, not because of its byline; cultural capital is subtracted from the calculus. By delinking information from the reputation of its supplier, people are free to speak candidly while also remaining humble. On the app, your accumulated “yak karma” — the numerical compilation of all your activity on Yik Yak — is only visible to you — unlike Twitter’s public follower metrics — and has no bearing on the success or failure of any individual post. It is the kind of meritocracy that everyone from anti-authoritarian hackers to venture capitalists might endorse, but as anyone that has been fired for lack of “culture fit” can tell you, meritocracy is anything but an even playing field.
Unsurprisingly, Yik Yak is not immune to the sorts of interpersonal aggression or structural oppression that pervades workplaces or any other social setting. In the months after its meteoric rise on school campuses, there were too-familiar reports of young adults and teenagers being harassed through the app. There were calls to ban the app at Clemson University after racist slurs were hurled in response to Ferguson protests. Dozens of other campuses weathered their own bomb threats, hateful jokes, and harassing speech much to the grief of students and administrators alike.
One particularly bad case involved a young woman named Elizabeth who suffered constant demands by her high school classmates to kill herself after they learned of her suicide attempt. In response she started a change.org petition that demanded Droll and Buffington build in better reporting features and community standards. The petition, which got over 78,000 signatures, reads in part, “With the shield of anonymity, users have zero accountability for their posts, and can openly spread rumors, call classmates hurtful names, send threats, or even tell someone to kill themselves — and all of these things are happening.”
Yik Yak responded to these events with code-level fixes directed squarely at personal harassment and young users. In January, lead community manager Ben Popkin announced on Yik Yak’s company blog the introduction of several features meant to catch what he called “bullying.” Behind the scenes, Yik Yak partnered “with a company that specializes in identifying threatening or offensive content through natural-language processing.” They also set up geofences to block the app in middle and high schools, enhanced their report feature, and gave parents tools to block the app on their kids’ phone.
But harassment does not begin and end at puberty, nor is it only the action of individuals. Structural oppression doesn’t crumble in the face of code because it is regarded not as a bug but a feature in our present historical moment. As the old saying goes, the system isn’t broken, it was made this way.
Yik Yak users may be all looked over by algorithms of standards and grace, but these are mere thumbs in a cracking dam. Standard social media anti-harassment features make the classic mistake of confusing structural violence for individual bad behavior. This is precisely why community policing on these sorts of apps is a Sisyphean task: Unable to address the underlying structures of oppression, they settle for whack-a-mole reporting features and ultimately rely on toxic communities to regulate their own toxic behavior.
Droll and Buffington’s desire to create an app that liberates the downtrodden has fallen short, but perhaps this is due to a lack of social theory and media literacy on their part, not technical impossibility. In all fairness, there is not a single social media company that has gotten this right. Far too focused on monetizing our digitally mediated communication rather than improving it, social media companies benefit from us being positive (Like!) but not necessarily good (Like Song of the South!).
Anonymity can be liberating if it carves out a void to shout into — a place where you can ask for candid advice or share a deeply personal story in the abstract. “In a society that often feels alienating and impersonal while stigmatizing mental health concerns,” writes media studies scholar Britney Summit-Gil, “Yik Yak allows people to reveal the most intimate parts of themselves without many of the consequences of doing so to people you know in day-to-day life.” Regardless of whether you think posts about broken hearts and existential dread are more performance than earnestness, their presence opens a comfortable space for vulnerability. Anxiety and depression are isolating conditions, and seeing a high-scoring post on Yik Yak can provide comfort and perhaps encourage someone to seek out in-person support.
Just as they believed anonymity was the source of the app’s innovative ability to extract honesty, Yik Yak’s founders—seemingly following Elizabeth and other vocal critics of their work—also believed it was the source from which all the problematic behavior flowed. But its voting system, too, plays a critical role. It’s hard to think of voting as a harassment vector because it seems more like a tool for making decisions, not an opportunity to be hateful. In a system where votes determine a post’s ability to command future attention, they serve to manifest and police a community’s discursive norms. But entire communities can be harassers; indeed entire nation-states have been unified by categorical discrimination, promulgating norms grounded in defining others as unclean, dangerous, or otherwise marginal.
Categorization and quantification are prerequisite for structural persecution. It is impossible to vilify a group of people if you lack a schema to define them. Michel Foucault, in The Order of Things, observed that there are only “two types of comparison: the one analyzes into units in order to establish relations of equality and inequality; the other establishes elements, the simplest that can be found, and arranges differences according to the smallest possible degrees.” The first type is measuring: finding discrete, abstract quantities that can be added up: I need two cups of flour; you are in a low tax bracket; he has the most karma points. Type two is more akin to tagging and categorizing: One might create a Pinterest board of terrariums, tag a Tumblr post as #NSFW, or wait to post your yak when you get back to campus so that the right people see it.
Neither type is wholly benevolent or strictly menacing. Foucault explains how both contribute to hegemonic discourse: The powerful both quantify and categorize everything and everyone in such a way that sustains the systems that made them powerful in the first place.
Some categorical hierarchies are easy for anyone to spot, like the ones that help define the popular jocks that the creators of Yik Yak sought to dethrone. Other structures require more work to reveal not because they are rare but because they are ubiquitous: those of race, class and gender. With the exception of location, Yik Yak doesn’t impose categories, thus letting all the unspoken categories of our social world make themselves apparent instead through the voting mechanism. Ranking systems like Yik Yak’s are especially prone to reifying norms because the incentives are in the common denominators. To become visible, let alone popular, one must try to appeal to the populace’s pervasive stereotypes, which reinforces the impression that they reflect what everyone actually thinks.
Since voting affords social groups the ability to demonstrate their collective “common sense,” votes will adhere closely to the race, class, and gender norms of that community. Implementing a voting system as an information filter signals that group cohesion is prioritized over most other outcomes, including justice or equality for all members.
Anonymity, of course, has long been a friend of voting. Though anonymity protects individuals from political retribution, it nonetheless amplifies voting’s ability to enforce norms. That may seem counterintuitive: Rey Junco, writing in Wired, argues that when retribution seems less likely, individuals are more likely to take a risky stand for an embattled person or an unpopular opinion. Anonymity lowers the barrier to bystander intervention, making it more likely that people will speak up to try to right perceived wrongs. But while anonymity might encourage bystander intervention in individual cases, it is unlikely to spark the kind of structural intervention that could prevent future instances of abuse. It does not agglomerate into an effective challenge to an abusive status quo. Just as anonymity makes retribution difficult, it also precludes positive reinforcement or collective action. And while we may readily accept anonymous statements as fact when they reaffirm our own experiences, we are more likely to ignore them if they don’t, and there is no credible identifiable person, with verifiable life experience, to back such statements up.
Foucault, in his genealogy of the West’s concepts of free speech, or parrhesia, concluded that truth is deeply connected to the speaker’s nonanonymity: A speaker’s belief in a statement’s truth is guaranteed by assuming a distinct personal risk in uttering it. Following contemporary criticisms of Athenian democracy, Foucault argues that “real parrhesia, parrhesia in its positive, critical sense, does not exist where democracy exists.” Only comedians and other reckless public figures are capable of critical free speech in the face of democracy, which demands not what is best but what is most popular — which must be something that appeals to the worst and best members of society. In other words, if you live in a racist or sexist society (and you do), it is a mathematical certainty that racism and sexism will not preclude a piece of content from winning an election, but it may help it succeed.
This is exactly what is happening on Yik Yak. It has become a site where racism, patriarchy, and other forms of dominance are rearticulated by the masses. Norms, whether they are desirable or undesirable, are rewarded by the voting mechanism. Individuals, for better or worse, are insulated from reactions to speech and are incapable of appealing to their standpoints to back up their claims. Jocks don’t get to rest on their laurels, and any Yak that starts with “As a…” must be viewed with suspicion as a potential fraud. Even worse, the same social structures make for high-ranking jocks and yaks. Everyone may still all be paying attention to the popular kids, they just don’t know its happening.
A social network has to be actively antiracist if it is to avoid the same fate as Yik Yak and the apps that have preceded it. Harassment is not an anomaly that should be flagged for review; it is a constant signal that must be anticipated, accounted for, and modulated — not amplified. What exact feature sets will accomplish this can only be known through frustratingly slow experimentation or a serendipitous eureka moment. The first organization that figures it out will have changed the world for the better.
