It's not Facebook's fault, it's your fault

Ethnic hatred on social media says more about us than about social media.

It is near-impossible these days to scroll through a social media feed for 30 seconds without someone losing their head over how fake news on social media leads to polarization and radicalization. Frankly, it’s wearying to see this many people misidentify the real problem. After all, there are so many other ways to lose one’s head.

Like this one: in 1866, Bulgarian nationalist writer and journalist Georgi Stoikov Rakovski wrote, “Let Turkish heads roll, and their corpses become playthings.” There was no Facebook in 1866. There were, however, far more Turks as a percentage of the population of what is today Bulgaria than there are now.

In 1866, the Bulgarians were seeking independence from four centuries of Ottoman rule, which they eventually got in 1878. At the time, most Bulgarians couldn’t even read. (As late as 1900, almost three quarters of the population remained illiterate.) Nevertheless, somehow they were able to take Rakovski’s poetry to heart. The Turkish purges (and infamous Ottoman reprisals) would continue, on and off, for the next 133 years, the last (as of this writing) taking place in 1989. By then, Europe was industrialized, the Cold War was ending, nearly everyone in Bulgaria was literate, and radio and television were well-established. But social media and the internet did not exist yet. The ethnic cleansing happened anyway.

The intercommunal violence and strife taking place all over the world today—Myanmar, Sri Lanka, Libya, India, the United States, etc.—is not new. Nor is it particularly worse than what has gone before. For as long as there have been human groups, these groups have deployed the hot tool of communication du jour—social media, radio, newspapers, children’s books, the printing press, rumor and hearsay, perhaps even complex language itself—in the execution of their designs upon one another. Blaming social media for inciting ethnonationalism is like blaming the opposable thumb for inciting people to carry torches and pitchforks.

Sri Lanka and Myanmar: same as it ever was

Let us consider, as an example, Sri Lanka, which was rocked by anti-Muslim riots in 2018. Facebook took sufficient heat for its role as a platform for incitement that it apologized, and was temporarily blocked by the Sri Lankan government. John Harris, writing in The Guardian, proclaimed, “In Sri Lanka, Facebook’s dominance has cost lives.”

This might be a compelling argument if Sri Lanka had been an uninhabited island before Facebook discovered it this past decade. However, as it happens, people have lived there for a long time, and charting even their recent history makes it fairly obvious that today’s developments are more of the same. Starting in 1983, the country fought one of the 20th century’s more demographically predictable civil wars. An 11% Tamil minority overwhelmingly concentrated in the country’s north felt marginalized by the 75% Sinhalese majority and fought for independence. After 26 calamitous years, the war culminated in a hideous 2009 denouement ignored by everyone outside of Sri Lanka (despite the rapper MIA’s best efforts). Government forces killed as many as 40,000 Tamils in the final phase of their campaign and then machine-gunned the entire Tamil Tiger leadership to death after international diplomats had negotiated their surrender before processing most of the surviving Tamils under Tiger control through internment camps.

Facebook penetration in Sri Lanka in 2009 was nil. At the time, less than 9% of Sri Lankans even had the internet. Yet, somehow, majoritarian Sinhalese Buddhists, rather than rejecting the architects of this bloodbath like good pluralistic global citizens uncorrupted by greedy tech titans and their clickbait, instead enthusiastically re-elected hardline incumbent president Mahinda Rajapaksa the following year. They did not need social media, or even the internet, to do this. They did not do it because algorithms told them to. They did it because they wanted to, because they are people and they have agency.

Nor is the more recent targeted violence against the island’s Muslim population new either. Anti-Muslim attacks in Sri Lanka have taken place before. One particularly nasty episode in 1915 was sparked by the diversion of a Buddhist procession away from a mosque, ostensibly because the mosque’s worshippers did not like the noise. The Buddhists considered this affront to be yet another example of British colonial favoritism towards the Muslim minority. Riots quickly spread across the island. The procession-goers did not need mass WhatsApp chat groups to broadcast their displeasure or spread bogus rumors; local newspapers and the spoken word did the trick.

Or consider Myanmar, where hateful Facebook comments were ubiquitous during the ethnic cleansing of most of the country’s mostly Muslim Rohingya minority during a vicious campaign in 2017. Advocates protested that Facebook was responsible, having allowed itself to be weaponized to stir up hatred that would otherwise not exist. Again, from The Guardian, this time John Naughton: “Facebook’s global monopoly poses a deadly threat in developing nations.” A UN investigator called Facebook “a beast.”

Once again, it is worth comparing pre-Facebook Myanmar to post-Facebook Myanmar to see if the humans who live there changed their behavior after the social network arrived. Spoiler alert: they didn’t. Myanmar:

Yes, the Tatmadaw, Myanmar’s paranoid and insular military, apparently used the platform to spread hate, but its actual 2017 military operation against the Rohingya has been well-documented as having been pre-planned, systematic and orchestrated from the top down. The propaganda campaign seems mostly to have reinforced views already widely held among the citizenry in order to legitimize a campaign that was going to happen anyway.

But why would the Tatmadaw do such a thing? That question leads us away from the blue herring that is Facebook towards the actual explanation.

The real culprit: us

I suspect there are two reasons people blame our social media companies for hate speech.

The first is unmet expectations. We were promised that an interconnected international marketplace of ideas would usher in an era of global harmony, and we are deeply embittered to learn that people on the internet behave the same as people always have.

The second is that the obvious alternative explanation makes us uncomfortable. But here it is: demonizing minorities is popular. People like it. It is easy to make groups fear each other, especially in conflictual and unstable demographic configurations. Groups do not like to be controlled by other groups, and so fearful groups will support the targeting of any other group of people deemed a threat, whether or not the threat is real. This is, I cannot stress enough, standard human operating procedure.

And thus, may I suggest a far more salient event in Myanmar than Facebook’s arrival that doomed the Rohingya to this latest round of horror: elections.

In 2015, Myanmar held its first free and fair elections after decades of military rule. Voters overwhelmingly supported the National League for Democracy (NLD), a Burmese-dominated nationalist democratic movement led by global icon and Nobel Peace laureate Aung Saan Suu Kyi. The elections also opened up a previously closed political environment, and anti-Rohingya chauvinism flourished.

The NLD was far more popular than the Tatmadaw, and so the Tatmadaw needed to increase its appeal among the citizenry. They chose the one thing where their credentials with the public were even more impeccable than the NLD’s: their anti-Rohingya bona fides. Researchers told The New York Times that the campaign’s purpose “was to generate widespread feelings of vulnerability and fear that could be salved only by the military’s protection.” Put another way, actually accountable to public opinion for the first time, the Tatmadaw scapegoated a hated minority group to boost its own popularity.

This sort of thing is, unfortunately, so normal it would be banal if it wasn’t so violent. The arrival of elections and the opening of previously closed political systems in multi-ethnic societies has a well-established track record of tipping off fratricidal violence and disintegration.

Consider:

  • Yugoslavs got to vote exactly one time before immediately ripping their country into seven;

  • Post-independence Nigeria managed exactly one election/census cycle before the Biafra war;

  • At independence in 1960, Cyprus managed one election cycle before UN peacekeepers had to be called in in 1964, and things deteriorated from there until Greek and Turkish intervention in 1974 left the island totally cleansed and divided, as it remains to this day;

  • The Republic of the Congo (take one, 1960) managed one election cycle before Cold Warriors and former colonizers exploited its internal divisions to set off a ruinous war that took the lives of its first prime minister Patrice Lumumba and the UN’s most famous Secretary-General Dag Hammarskjold and eventually led to Mobutu’s dictatorial Zaire kleptocracy;

  • The Democratic Republic of the Congo (take two, 1997-present) played host to “Africa’s World War” for its first six post-Mobutu years (estimated death toll: as many as five million people, maybe more), then fell apart almost immediately after its first election in 2006, with the victor and the runner-up’s respective praetorian guards engaging in pitched street battles in the capital and the runner-up eventually seeking refuge in the South African embassy, then Portugal, and, later, The Hague (where he was, to the consternation of international justice advocates, acquitted);

  • As soon as the iron hand of the Communist Party was lifted from the Soviet Union, the state fractured along straight ethnic lines, and a number of ongoing conflicts today stem directly from the misalignment of people and borders in that partition (see Nagorno-Karabakh, Crimea, Abkhazia, and the Donbass, for starters);

  • When Sudan achieved independence in 1956, South Sudan immediately launched a 16-year civil war to secede from Sudan, then launched another 22-year civil war to secede from Sudan, finally seceded from Sudan in 2011, and held exactly one election before its two largest ethnic groups launched a civil war against each other for control of the new country;

  • Iraq’s simmering insurgency against occupying American forces predictably turned into a full-blown civil war after the country’s 2005 elections, and the 2010 and 2014 polls didn’t exactly make things better;

  • Nine years after the Syrian people demanded the right to vote in the face of a minority dictatorship, their country lies in ruins;

  • and in the first two years after Ethiopia’s political opening under Nobel Peace Laureate Abiy Ahmed, three million people have been ethnically cleansed from their homes… and they haven’t even voted yet! (This year’s elections sure looked like they were headed for the rocks before being mercifully postponed due to COVID-19.)

There are many more of these, the pattern repeating across the modern age, mostly without the accelerant of social media. Indeed, ethnic conflict as a result of demographic change or political upheaval occurs even in an alternate contemporary internet ecosystem. Consider the ongoing suppression of the Uighurs in China, a classic demographic strangulation. In 1945, Uighurs constituted somewhere between 75% and 83% of the population (sources vary) of the province of Xinjiang, their home turf. By this decade, that number was 45%, thanks to systematic ethnic swamping by Han Chinese. Unsurprisingly, resenting their second class status and feeling their control of their surroundings slipping away from them, some Uighurs revolted violently against their Han overlords. Equally unsurprisingly, in response, the Han-dominated Beijing central government gave the Uighurs the hammer; perhaps a million of them are now in camps, to the gratitude of a great many Han. No Facebook was required for any of this; it’s banned in China. Instead, the country has its own social networks with strict controls, and the results speak for themselves.

Perhaps, then, the algorithms aren’t corrupting us, rather we are corrupting them. Tasked with keeping our attention, the algorithms are showing us what we wish to see… and what we wish to see turns out to be inflammatory defamations of our identitarian adversaries.

In the United States, polarization has reached fever pitch in predictable conjunction with an identity-based demographic transition that has set off some sort of conflict in virtually every known prior instance anywhere in the world where it has occurred. Our stories, our news, and our narratives reflect our divisions, rather than causing them. Fake news and partisan rumor play out not just on the new-fangled internet but on good ol’ fashioned television. We do not have the same realities because we think of ourselves as distinct. We fear the designs of the other, whoever the other may be, and thus we are spring-loaded to react immediately to any claim of wrongdoing by other groups against our own. “The human mind,” as Ezra Klein writes in his book Why We’re Polarized, “is exquisitely tuned to group affiliation and group difference.”

In such polarization, truth is irrelevant. Ask the Armenian and Azeri partisans to the long-running Nagorno-Karabakh dispute precisely who ethnically cleansed who. Compare notes, and try to square their stories. Then throw up your hands. There is no truth in Nagorno-Karabakh, there is only fear of them. “Impartial” media cannot long endure two-sided sniper fire in tribal no-man’s land. And so they don’t. And if they try, they just make people mad. As in the Caucasus, so too now in America. It is our demography, not our algorithms, that makes us do this. We are driving the machines, not the other way around. If the machines were different, we would drive them differently, but in the same direction, for we would still be ourselves.

And so if, just as a hypothetical, the 2020 US Presidential Election begets a similar outcome to the aforementioned examples, we can surmise that it will be because of America’s internal dynamics, not whether or how those dynamics were strewn across the internet.

Emperor Zuckerberg?

If you accept the above argument, then it follows that to ask social media to regulate content is a Sisyphean battle against human nature. But it is worse than that; it is swallowing a spider to catch a fly, because it grants tech giants dangerous powers that they should not have, all to solve a problem that they cannot solve.

Why should they not have such powers? Because they are unelected and unaccountable. Facebook chief Mark Zuckerberg and Twitter CEO Jack Dorsey fear becoming “the arbiters of truth,” and in this they are far-sighted. Their successors might not have their scruples, or might face disturbing public pressures unimagined today. This is an extraordinary amount of power to put into the hands of private entities. Yes, some sort of oversight board of our social media behemoths seems like a good idea, but this body is unelected also, and were it actually subject to popular vote by Facebook users, it would immediately run into the same paralyzing and destabilizing polarization experienced by other elected institutions, such as the US Congress. It is provocative to instruct one person or company to “do the right thing” when the body politic deeply disagrees on what the right thing is, and to whom or on whose behalf it should be done.

Perhaps you think that this, given the current chaotic internet environment, is nonetheless what we need. After all, from Franz Joseph to Tito to Haile Selassie, diverse polities have often required an unelected, all-powerful emperor who can manage the interests of different groups and crush any identity-based division and dissent to keep the peace, and these polities have often immolated once this strong leader left the scene. So it can work, at least for a while. Is this what we want? An Emperor of the Internet? Choose carefully: emperors, once enthroned, are difficult to unseat. And, après eux, le déluge.

So yes, we all know that Facebook, like any form of communication, can be used to spread messages of hate in ethnic conflict. And yes, the site has some responsibility to try to rein in the worst of it. And yes, we would all probably be better off if internet economics depended a little less on monetizing our attention. But at the end of the day, when it comes to ethnonationalism, the medium is just the medium, and the message is the same across space and time. It is you who clicks on clickbait. It is you who disappears down internet rabbit holes. It is you who gets a jolt out of reading identity-based content, be it incendiary op-eds or Buzzfeed listicles. It is you who fears them, whoever your them may be. It’s not Facebook’s fault, it’s your fault.