Social media — the medium that defies discussion

The semaphore line, a 18th century system for sending messages across France. Its speed of communication allowed the French state to develop a more effective infrastructure and improve national security.

I wanted to write something on the politics of social media and how it played a part in Trump’s victory — probably it adds nothing, but writing is the only way to think things through.

The point I’ll try and make is that social media is so opaque and slippery that it cannot itself become the subject of proper scrutiny — filter bubbles and viral misinformation are second order effects, the result of a phenomena that we each experience so differently we cannot subject it to collective scrutiny.

Yet we have to design our public sphere to promote democracy, just as we have to design any other public institution. Achieving this without undermining liberal principles is the key challenge that modern societies face. The modern idea of a nation is dependent a functioning public sphere. How would people see themselves as part of a nation without a national conversation? Fragmenting nationhood is exactly what we saw in the US, and (perhaps more literally) in the UK too.

The fact that social media is provided by private companies shouldn’t prevent us from regulating it any more than it stops us regulating TV or newspapers. We refuse to let food companies put dangerous additives in our food to make it taste better — why should we accept Facebook doing the same to our conversation?

If we don’t do it now, it may be too late forever. Trump will never change the system that elected him, and leaders in other countries may very well face similar incentives.

It’s commonplace to note that social media assists the spread of lies, or functions as an echo chamber, but neither of these things would matter (at least as much) if it was possible for national media to address these issues directly.

Echo chambers and published falsehoods clearly predate social media. Newspapers and TV stations — to varying degrees — act as filter bubbles. Some media outlets also have a deserved reputation for publishing lies. These phenomena have been around for many election cycles.

What’s really different about social media is that it cannot be fitted into the wider media ecology. If the Daily Mail prints a lie, it can be countered by other papers of a different view. The Daily Show can take apart the Fox News narrative, and vice versa. Over time, each of these outlets get a reputation that lets everyone know about their political leanings, and a reputation for impartiality (or otherwise).

What has changed is that one person cannot know what another person’s media experience is. What do you see on Facebook, what does your Twitter look like? What email scare stories are circulating? We are collectively blind to the media environment.

Let’s look at the Trump horror story as a case study and how this effect plays out.

Theories on how someone as transparently unpleasant as Trump could win include, firstly, a classic tyranny of the majority. Some white, rich or male, voters may have seen Trump as the last bastion defending their status within American society. (Strangely the large poor and white Trump constituency may have voted for him both to preserve their privilege and out of economic desperation.) This factor was perhaps missed in polling, with voters unwilling to admit to this type of motivation.

Secondly, the idea that some people voted for Trump to ‘send a message’, in the expectation that he would lose. This is born out by down ballot voting, which may, via the principle of anticipatory balancing, signal that voters expected a Democratic government. This theory is also supported by the evidence of ethnic minorities voting for Trump when his policies clearly stand to disadvantage them.

It’s impossible to exhaustively list the theories, but the final one I’ll address is perhaps the most widespread: analysis that credits Trump’s success to rejection of globalisation which he embodied. His promises that played on the notion that China or Mexico took advantage of the US, or his idea of ‘extreme vetting’ of immigrants. Again, if this is your motivation for voting, perhaps you don’t want to tell a pollster about it.

Each of these has an analogue in the way Brexit played out.

All of these explanations have a least partial causes in the structure of social media. Steven Pinker has emphasised the importance of print media’s role in allowing us to empathise with others, is it possible that the filter bubble negates this effect, or even makes empathy with other groups more difficult? Certainly, that could exacerbate the ‘tyranny of the majority’ effect that might otherwise be tempered by communal feeling. When we see Trump voters expressing not just delight in their candidate, but outright hatred for Hilary and her supporters we are witnessing the consequences of social media echo chambers. It’s not just political disagreement, it’s the full dehumanisation of your political opponents — people’s whose views you’ve only seen represented as stupid or mendacious.

Attempting to ‘send a message’ by voting Trump while assuming Clinton would win speaks to me of the homogeneity of governments in the preceding decades — and the perceived impossibility of change. The same logic might cause a voter only tepidly supportive of Clinton to stay at home.

Conversely, the opportunity to vote down that same homogeneity might have galvanised Trump’s supporters, and helped them forgive him his abundant failings.

Which leads us to the nature of that homogeneity — the preceding two decades of neoliberal political monoculture, the view that more trade and immigration is good always and everywhere. This view has is accepted by economists across the spectrum, but the argument has never been won in the public sphere — in fact it has mostly never been made. From NAFTA to the UK Government’s acceptance of unfettered EU immigration, those making the policy and negotiating the agreements have operated according to principles unknown to, or unconvincing to, the wider public.

As former Labour ministers belatedly admitted of immigration from EU accession countries, and Hilary seem to have acknowledged too late in abandoning TPP, political consent was never obtained. Gaining that consent is a formidable challenge. The ‘lump of labour’ theory might be a fallacy, but it makes common sense and resonates much better than some abstruse economic theory is ever likely to. Virtually no one knows about the theory of comparative advantage that is the foundation of neoliberalism. Good luck to anyone wishing to persuade a majority of the electorate to vote for you through the theory of comparative advantage — you might as well start spouting particle physics on the stump as far as most voters are concerned.

So how did governments manage to support a neoliberal agenda for 20 years if they never had political consent? Because the media. Journalists from left and right signed up, and as a result that battle never had to be won. They understood the economics so you didn’t have to. The ‘mainstream media’ reached a consensus of their own, and simply presented neoliberalism as a fait accompli. It wasn’t a conspiracy, everyone truly did believe the theory.

Not any more. The media is no longer in the control of an elite, and as a result concepts that cannot be immediately grasped by voters don’t stand a chance. As the neoliberal consensus unwinds the vacuum is being filled with dysfunctional unreason.

That dysfunctional unreason is the disembodied fragments of content splintered across Facebook. The crazed rantings of any tiny band of conspiracists can present themselves as the lone voice of reason. Debunking by mainstream journalists is impossible: they cannot discover all the toxic misinformation flowing around, and even if they did they would spend their time debunking stories most of their readers had never encountered in the first place.

It’s the fragmentation of the content, the shattered dialogue that’s the root of the problem — not it’s fact checking or the existence of filter bubbles in themselves.

A popular reaction to the result — aside from grief — has been to hail it as the beginning of the end for the system, which will be inevitably follow by the reformation of the world according to your own pet theory (Zizek, Paul Mason). These ideas have a compelling mystical quality, but I’m never sure what actions these visions recommend, if any — though of course they are great if you need some cryptic words of wisdom to fill in while your grapple with the intertwining, polymorphic consequences of Trumpism.

What events recommend to me is the need for conscious design of social media systems.

Freedom of speech has been at the heart of many Western theories of morality and politics — Voltaire to Rawls. But we’ve never had, and would not want, totally free speech. The law prescribes libel, copyright infringement, incitement to hatred, restricts pornography, regulates advertising, controls who can fund political adverts and limits ownership of media companies — to name but a few restrictions. All developed states across the world have these rules — though obviously in many different forms. Even things that one is free to say can have terrible consequences. Denying the holocaust, even if legal, could see someone shunned by their friends and unable to get a job — a punishment worse than prison in many ways.

Really, freedom of speech is a shorthand for a permissive public sphere where wide ranging debate of political issues can take place, so that a society can collectively reason to good conclusions that satisfy diverse groups within a society.

This is not a marginal issue. Democracy is not the act voting — North Korea has voting. Democracy means empowering people to meaningfully represent their own views. A functioning pubic sphere absolutely necessary, and very nearly sufficient, for democracy.

It’s commonplace among techies to assume that privacy is the only political issue the Internet faces. It is not — the design parameters for social media that serves the people who use it are considerably more complex. Facebook and Google are apparently looking at fact checking, but it’s not enough.

Some concrete suggestions include that social media companies above a certain size ought to:

1) Reveal the algorithms they use for rankings showing content

2) Make the reasons for showing a particular item legible to the user

3) Make content receiving a high volume of traffic visible transparently to everyone

4) Make statistics regarding political advertising public

5) Flag filter bubble effects to users

6) Allow publications to publish retractions that reach users

This most certainly is not a definitive list, but hopefully it gestures at the kind of design choices we must make — all, of course, while respecting privacy.

It may already be too late — surely Trump won’t inhibit a media system that has just elected him. Could other countries become stuck in the same loop? If nothing is done democracies may start lurching around politically as they elect leaders with ever less effective discussion of the issues at hand — one disappointment following another.

Trump’s election was a kind of deranged public entertainment masterfully conducted by a reality TV star. His fans frequently have the demeanour of wrestling fans — something that may be no coincidence.

Am I just a bad loser — not even a loser since I’m not American and didn’t vote? What if this is democracy actually working? To play devil’s advocate, what if Trump voters were voting — quite reasonably — for someone who isn’t from the same political mould as all the others, who can shake things up a bit, and who isn’t hypnotised by neoliberalism? They ‘held their nose’ and voted for him despite all his faults. Trump voters weren’t hoodwinked by fake news, they rationally acted in their own self interest.

This may be the case — but if Trump is the best alternative a democracy can provide, surely something isn’t working.

PhD on digital systems for collective action and social network analysis.

PhD on digital systems for collective action and social network analysis.