Private but Viral

A few years ago, #KillAllMen made its divisive rounds on social media. This phrase was controversial for all the reasons you might imagine.  One point of conflict was identity-based: those outraged by it complained that substituting almost any other identity group for “men” would be deemed abhorrent by the very people using it in their posts. The phrase’s defenders countered that almost no other identity group enjoys the same clear-cut dominance in modern and historical power structures, and so “men-bashing” was acceptable in this context. 

The other conflict arose from the actual wording of the hashtag: its users said that the phrase was shorthand for something along the lines of “we’re sick of dealing with a subset of unpleasant men that perpetuate sexism and misogyny.” Their critics’ counter was that, in a public sphere like Twitter or Facebook, such shorthand is unacceptable. Where you stand on the first issue depends on a host of priors that one article cannot resolve. But the second issue warrants more discussion because technology’s ability to move human expression out of its intended context has upended definition of what is and is not a public forum. 

Large groups of humans have always broken themselves into smaller, closer groups, and the Internet has not changed this trend. Facebook, Twitter, Instagram, and any other platform are home to countless sub-communities, from dog moms to graphic designers to deer hunters. Communities have their own cultures, worldviews, and rhetoric that can be confusing or offensive to outsiders. 

For the first time in history, people are regularly talking with people they are not talking to.

What distinguishes these social media communities from their historical counterparts is that outsiders have unprecedented access to these groups’ inner dialogue. The mechanisms for sharing and liking/disliking posts can elevate a phrase or idea meant for a small, specific audience to extreme publicity among a wide range of different groups. In fact, on platforms like Facebook, Twitter, or Instagram, unless you intentionally change your privacy settings, every post or tweet may as well be published media content. 

Back to #KillAllMen. The communities that used this phrase were largely female and feminist. Their members had lots of context and experience that contributed to both the motivation and meaning of that phrase, so it was neither offensive to members of that group nor a call for the slaughter of 3.5 billion people. But to many people outside those communities, it was both those things. A 65-year old man who just got on Twitter and doesn’t spend much time with millennial feminists will likely take the hashtag at face value. A feminist posting on Twitter may rightly protest she is not addressing the 65-year old. She is, however, talking within virtual earshot of him. 

For the first time in history, people are regularly speaking with people they are not speaking to. Politicians and journalists are no longer the only people who must write and speak with the knowledge that every word may be interpreted in the worst way possible, and that there are strong incentives to spread those interpretations. The statements by an outgroup that will get the most attention from an ingroup are those that make the outgroup look the worst. An avid chicken griller will be far more likely to share a news story about a vegetarian who wants to put all chicken grillers in jail than they would a story about a vegetarian who doesn’t like the taste of meat.  

This phenomenon of content traveling far beyond its intended audience has a name: context collapse. Coined by researchers at MIT’s Media Lab, context collapse refers to many disparate audiences simultaneously encountering content that may not have been targeted towards any of them. Context collapse is the realization that a conversation during your 8th-grade sleepover is being transcribed onto a billboard overlooking the nearby interstate. 

For the first time in history, people are regularly talking with people they are not talking to.

Context collapse goes beyond individual reactions to the statements of those outside their tribes. 20 years ago, the idea that you could be fired for expressing what you think about a high school buddy’s fishing picture would be insane. Now, while the intended context of your comment may be a joke targeted at an old friend and no one else, its potential audience could interpret it as a PR statement that represents your employer. The threat of virality makes everyone associated with a person potentially liable for every reaction anyone on the Internet might have to everything that a person might say. That’s terrifying.

Social media, and especially platforms with character or video-length limits like Twitter, are also optimized to remove even face value context. A post may be one part of a long series, but none of those other parts matter if that one contains the right buzzwords. For example, a post may be clearly satirical in the context of the series it is part of, but then outsiders interpret it literally when it goes viral as a standalone piece of content. The speed at which something can go viral also means that by the time someone realizes a post of theirs has been misinterpreted, it is far too late to clarify. A final inhibitor is that a person’s body language and facial expressions provide lots of context for their statements; 280 characters eliminates these signals. 

Over the past 15 years, people have become more aware of how context collapse can take their statements to places they never expected. Many people have begun to self-censor what they post online, whether by removing drunk photos of themselves or refraining from tweeting jokes for fear that they will be misinterpreted. But while people have begun to understand context collapse, they are still quite unsympathetic towards its victims. And often its most severe victims are people who did not understand the risk associated with their online statements and could not have known better when they said the things they did. These people are our former selves. 

Time periods are also their own contexts, and they can collapse just like any other. For example, as a 9th grader, I did not consider what my college friends would think of the Instagram photos and captions I posted back in 2014. Although I could have deleted those photos at any time if I had posted something that my high school soccer team would have understood but my freshman floor would not, all it would have taken was one person having the wrong reaction for the whole thing to blow up in my face. The idea that your birthday wishes to your cousin in 2009 can be seen to represent your employer in 2019 is something that most people now accept as fact. But most people do not extend any grace to the individual in 2009; they evaluate people as if their past selves are specifically addressing everyone that hears them in the present. 

The threat of virality makes everyone associated with you potentially liable for everything that you might say. That’s terrifying.

People’s online actions can wreck their careers and relationships, sometimes because they simply don’t realize the scope of their audience. However, as more people come to terms with the extent to which their content can move beyond its targets, the way we use social media will change. Similar to the way corporate speak converges on a series of bland, universal buzzwords, and mainstream media outlets rarely say anything their audiences won’t like, people are becoming more aware that what they say has to be acceptable to people whose reactions they can’t anticipate. 

This adjustment extends past social media users to the platforms themselves. They are beginning to realize that people don’t want to talk to everyone all the time. In the past several years, Snapchat and Instagram have introduced features that allow users to limit who can see their public stories. These private stories help users reclaim their ability to address a specific audience and no one else. Facebook has also contributed in this sphere in the past five years by making more users’ content invisible to everyone except their friends. Twitter has, well, not followed suit, and only gotten worse if anything, likely because ordinary users going viral has always been one of the platform’s core attractions.  

In 2017, Facebook changed its core mission statement to “bring the world closer together.” But people often separate themselves because there are quite a lot of people that they would rather not be close to. Facebook and its fellow platforms have brought many worlds together that were never meant to be that way, and it has only driven us further apart. Mitigating the negative effects of this context collapse requires more than privacy options on their part or self-censorship on ours. It also requires a normative shift towards suspending instinctual reactions toward a person’s online statements until we know their full context. We should not count decade-old posts against a person, except in the most extreme circumstances. We should not be eager to spread content to people that will angrily misinterpret it. And we should think about the group a person is addressing, not the crowd listening in. After all, we would all love to receive the same grace if we were the ones under scrutiny.