Politics and the prisoner’s dilemma

Jamie Thunder
Jul 30, 2017 · 8 min read

If you’re reading this, I assume you’re aware of at least one of politics, the media, or Twitter. And if you’re aware of at least one of these, you probably despair at the state of the world. More than ever, it seems that lies, half-truths, and malicious misrepresentations dominate:

  • Campaigns print lies on the side of a bus;
  • Political parties shoddily ‘cost’ their manifestos — or don’t bother at all;
  • Politicians make promises about things that are out of their control, or that they’ve no intention of doing;
  • A lie in a tweet gets thousands of RTs, while a correction gets six; and ultimately
  • People believe all sorts of things that aren’t true.

The fundamental driver of all of these problems is a version of the prisoner’s dilemma operating at various levels of society. If you’re not familiar with the prisoner’s dilemma, it works like this: there are two prisoners, who are each given the option of informing on the other (‘defecting’), or co-operating with one another and staying silent. The combination of their choices will determine their sentence (or outcomes):

So if both co-operate, they get a year each (a total time in prison of two years). If both defect, they get two years each (a total time in prison of four years). If one co-operates and the other doesn’t, the co-operator gets punished (a total time in prison of three years).

This dilemma is at the heart of the problems with our media, our politics, and our social (media) interaction. It’s better, altogether, for everyone to co-operate and work in good faith — and not to lie, or misinterpret, or throw out policy aims with no idea if they can be achieved. But there’s always the possibility that your good faith will be exploited: turning the other cheek risks getting punched some more.

And so, it becomes perfectly sensible for sites like The Canary and Skwawkbox to pop up to counter the Daily Mail with lies and exaggerations of their own. It becomes perfectly sensible for Jeremy Corbyn to say he’ll ‘deal with’ existing graduate debt without knowing its size (or looking too hard for it). It becomes perfectly sensible for the Leave campaign to claim trade with the Commonwealth will replace the advantages we’ll lose from Brexit.

It becomes a world where instead of turning the other cheek, everyone’s just punching each other pre-emptively. There’s no trust: everyone is assuming bad faith on the part of everyone else, and reacting according.

You can see this immediate leap to bad faith clearly on Twitter, as people make real-time snap judgements about other people’s motives. Over the last few days I’ve made a (long) list of some of the impacts this assumption of bad faith has on people’s interpretations of each other:

  • Getting something wrong becomes lying
  • Having a different view of what is right becomes being evil (being right-wing becomes being fascist; being left-wing becomes being Communist)
  • Not knowing something becomes ignorance (and depending on the issue might also be taken as evidence of racism, sexism, ablism, homophobia, etc)
  • Misreading something becomes manipulating
  • Questioning someone becomes kicking
  • Supporting someone becomes cheerleading
  • Disliking someone becomes hating
  • Liking someone becomes adoring
  • Thinking people don’t co-operate becomes being cynical
  • Thinking people aren’t self-interested becomes being naive
  • Lacking data becomes attempting to pull the wool over someone’s eyes
  • Saying something that’s been said before becomes passing someone’s ideas off as your own — or needing a crutch to prop up your beliefs
  • Being clumsy becomes being (micro)aggressive
  • Correcting an error becomes making an embarrassing climbdown
  • Taking a decision that causes short-term harm (for long term benefit) becomes being uncaring
  • Taking a decision that causes long-term harm (for short-term benefit) becomes being childish/unserious
  • Not noticing, or even not mentioning, a particular downside to a policy becomes blindness to suffering
  • Not noticing, or even not mentioning, a particular upside to a policy becomes blindness to benefits

It’s the extremisation of political discourse: can you feel the sneer and veiled hatred in the words in bold?

(This does not mean that the words in bold are never justified. A good recent example of is professional dickhead Paul Joseph Watson, who tweeted some rubbish from his mum’s house about a cartoon that featured a black man in ancient Rome. This was then gleefully and entertainingly picked up by a historian who pulled apart the idea that Rome wasn’t ethnically diverse. But it did make me think. I had no idea Roman Britain wasn’t extremely, Trump-administration levels of white. And my first reaction when I saw his tweet was ‘I don’t really care, but yeah that seems wrong’. If I’d then tweeted that, particularly with a false sense of certainty that’s easy to do on Twitter, would I have been pilloried?)

If that’s how you interpret other people’s motives, it’s no surprise that you then react accordingly, either by accusing them or telling them to ‘come back when you’ve got a heart’, or ‘get back to the drawing board’, or ‘read some [academic] then we’ll talk’. Or just writing them off.

And these reactions mean that the first person, no matter what their original intent, is likely to take umbrage, which in turns solidifies the second person’s assumption of bad faith, and so on. It leads to a poisonous spiral: ultimately, disagreeing becomes disliking becomes hating becomes sending death threats or even physical attacks.

Not that this necessarily leads to death threats. Another reaction is that people retreat: they stop engaging, getting quicker with the mute/block button (Twitter allowing you to block people from your entire timeline, not just your mentions, for being irritating rather than abusive doesn’t really help here), or even limit their @s to people they already follow, because it’s exhausting to debate someone who’s in bad faith, and they’ve had their fingers burned. And so they only see people with whom they disagree when they are being mocked in their timeline. Which exacerbates the bubble.

In a world where assumptions of bad faith are common, arguing on Twitter has becomes a numbers game: what are the odds that the stranger who’s just popped up in your mentions to inform you that ‘sorry, I don’t follow’, or ‘just to clarify’, or ‘so you’re saying’ is one of any of the below? Assuming equal distribution (which on Twitter seems unlikely, but hey), you’ve got

  • a 20% chance of a troll;
  • a 20% chance of a productive if lengthy discussion as you both try to make nuanced points 140 characters at a time (at the end of which you may/may not agree);
  • a 20% chance of a productive discussion once you’ve ironed out some factual issues (at the end of which you may/may not agree);
  • a 20% chance of someone actually agreeing with you; and
  • a 20% chance of banging your head against a wall because someone doesn’t believe a fact (or rejects an extremely likely prediction/plausible interpretation).

And that’s before you get to the people who sarcastically quote-tweet you to the applause and occasional piling-on of their followers.

It’s nigh-on impossible in practice, and especially at a first interaction, to distinguish between (1) ‘trolls trying to win an argument’, (2) ‘true believers who might be up for a proper discussion even if you ultimately bit different bullets’, and (3) ‘people whose analysis is just very poor’ (who may/may not be willing to consider this possibility). Much like Van Halen and the brown M&Ms, therefore, a stranger popping up in your mentions becomes a good indicator of someone you shouldn’t bother with, however unfair that is on the person themselves.

And in time, the only people who are left arguing are the ideological warriors, the teflon-skinned, the people with extremely fixed opinions (which will now not be challenged as much), and the people who just like a fight.

(for what it’s worth, I try pretty hard to be (2) above, or if I fail, then at least a (3) who’s willing to adapt his analysis in light of new information — but it does grate to be told your analysis sucks in any case, never mind by someone who is, or whose followers are, then triumphant when you admit you’re wrong)

What on earth do we do about this?

What’s the way out of the prisoner’s dilemma? There’s a range of academic literature on strategies to ‘solve’ the prisoner’s dilemma over single and multiple iterations (according to Wikipedia, to solve it in a single iteration the best strategy is to ‘cheat’, and over multiple it’s some form of ‘be nice first, then do whatever your opponent does, with occasional random acts of forgiveness’).

But this strategy has problems: the random acts of forgiveness (which are intended to stop the whole thing descending into a cesspit) do still allow some level of scot-free defection. It requires a heroic adherence to the strategy to keep forgiving people — which is easy enough for an algorithm, but harder for a human who may be hungover, sleepy, in a rush, or simply have had enough.

Then there’s also the fact that simple prisoner dilemma setups have two people involved. But as anyone who’s ever tweeted knows, no online conversation is really between two people (even private messages or emails can occasionally be used as ammunition). You and I might be willing to take each other in good faith — but maybe we have a mutual follower, or just someone who stumbled across our conversation, who’s not, and so in spite of our good faith interaction, I’ve still got whacked. And so I either have to suck it up and assume good faith, probably repeating much of what I’ve said to you already, or I lash out and create a new poisonous cycle (and will then be criticised for inconsistency).

Another solution that has been tried is fact-checking and mockery of individual claims. These are valuable and worthwhile. But they will always have two problems:

  • There’s a preaching to the choir effect, which makes them not especially effective: does The Daily Mail really care if Gary Lineker criticises them?
  • Even where they can work, they can only address individual instances of defection. As I’ve tried to argue here, the problem is the entire structure of the prisoner’s dilemma, not that some people don’t co-operate

What you need to solve the prisoner’s dilemma is some sort of common set of rules. In theory that could be ‘trust’, but I can’t see that catching on when it comes to politics. The opportunity for abuse of trust will always be there and will always be taken.

And so I’ve ended up in the position of deciding that we have to bite one of two pretty bitter bullets:

  • We accept the continued existence of racism, sexism, lies, homophobia, misrepresentation, etc in the name of free speech; or
  • We get serious about the regulation of free speech in media, political adverts, manifestos, and social media (with high reach) to cut out outright lies, adjudicate on contested claims and accusations, give prominence to corrections and rights to reply, and enforce against breaches, including levying fines, blocking websites, and so on.

It’s one or the other, and doing nothing will have consequences: the consequences in the first bullet. I believe in owning my decisions (insert joke here about my decisions getting owned), and I would prefer some version of the second outcome.

__

If you want a condensed version of my argument here, it’s in the tweets below:

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade