Social media is an adversarial system, the MSM is inquisitorial

God, what a horrible fortnight. So much has been written. I’ll just record two brief thoughts.

First, one idea circulating has been that the Palestinians have kept alive grievances of their homes lost in 1948, and that it would be better if they moved on, the way Europe’s many post-war refugees moved on and rebuilt their lives, avoiding irredentism, treating the post-war boundaries as settled. I get this, up to a point. Yeah, Arab states have treated Palestinians shamefully and have prioritized Palestinian statehood over the welfare of people from Palestine, as Matt Yglesias points out, by effectively forcing them to stay as refugees.

But I do think this argument can spring from a kind of neoliberal arrogance. In David Goodhart’s terminology, it’s an argument made by Nowheres, telling Somewheres “don’t be so backward”. (An argument made in Thomas Friedman’s The Lexus and the Olive Tree.) Well, if you have moved thousands of miles and broken ties with home to pursue your dreams, good for you, but respect others who care for different things. And most people’s roots are shallow, compared to those in the Middle East. There’s a quote from Heirs To Forgotten Kingdoms, a great book about vanishing religions. Someone asks a Samaritan guy about how long his family has been in the area. “Seventy-five,” he says. “Seventy-five years, that’s a long time!” “No. Seventy-five generations.”¹

This isn’t a policy argument. I think it is obvious, now if it wasn’t before, that you cannot expect Israel to literally let in millions of people whose grandparents once lived there, and of whom some unknown proportion hates Jews and wants to kill them all. It is just an argument against being too confident about what other people ought to prefer.

My other thought is just this. All the talk, all the anguish, the circulating pictures and videos, the unfocused rage, the accusation and counter-accusation, the horror, the thoughtful analysis, the real and fake experts, the tweets and thinkpieces, the ramifications into every other issue, the culture war discourse sucking this up as it sucks up everything else: is it actually any different, does it contribute any more, than a chimpanzee band screaming and pant-hooting around a fight to the death?

I suppose I hope that it does.

I want to think about one particular ramification: the events in Israel have restarted the Forever War between social media and the MSM. The case against social media, in particular X/Twitter, is obvious. It’s been an open sore of extremism, craziness and disinformation. Pretty soon after the original attacks, people started to look for the exit, and there was a flurry of public moves to Bluesky. (I did it myself and got an invite; I’m hughjonesd.bsky.social, and maybe I’ll be more active there in future.)

But the mainstream media has been criticized too. The New York Times first blamed Israel — well, reported Hamas blaming Israel — for a “strike” on Al-Ahli Arab hospital, then downgraded it to a “blast”.

Then in a final, ouroboros ignominy it became its own story:

The changes are now being posted on Twitter as evidence for the Zionist-controlled/Hamas-controlled media (take your pick).

Ouroboros

Social media also got credit for how fast open source intelligence analysts worked on images of the hospital blast, and (from some) because circulating videos gave users a visceral sense of the horror of the October 7 attack:

Adversarial and inquisitorial media systems

Here is a useful way to think about it. There are two kinds of judicial systems, adversarial and inquisitorial. In inquisitorial systems, an appointed judge is responsible for getting at the truth of a case. He can call witnesses and may participate in the investigation before a charge is brought. In adversarial systems, the judge is a referee. A criminal case works like a battle between the prosecution lawyers and the defence. The truth is supposed to emerge from this clash between interested parties.

Most Anglo-Saxon countries have adversarial legal systems. But until the internet came along, our media system was inquisitorial. The journalist played the role of the judge. He gathered facts, interviewed witnesses and summed up the case. He was trusted as an expert. Most people only took one newspaper, and (at first) only two or three TV channels existed. So they were only exposed to a limited range of views. More importantly, the market for points of view was not very contestable. It took a lot of money to run a newspaper or a television station. If the media class as a whole agreed on something, it was very hard for that view to get a hearing. Today’s crazies, like anti-vaxxers and conspiracy theorists, think they are still living in that world: they believe their ideas are unpopular not because they’re daft, but because they’ve been suppressed.

The difference between media and social media is not one of personnel. Every professional journalist and every media organization is on Twitter. But social media puts the ideas together differently. The journalist is not guaranteed any position. He can present his angle just like anyone else. In other words, social media is an adversarial system.

This helps to resolve the paradox that Twitter can be a “hellsite”, a “dumpster fire”, but also an important place to learn things that the MSM don’t get right. There are no barriers to entry. Anyone can tweet. And while the average person is neither particularly smart nor dumb, the author of the average tweet is somewhere between stupid, evil and deranged. (Just read absolutely any popular account and look at the replies.) Luckily, you don’t have to read the average tweet, and if you follow the right people, you will listen in on an intelligent, ever-flowing conversation.

The professionals are still actually the best at this! About half of the people I follow are journalists, because they are good at finding out and writing about things. But anyone can chime in, and a good comment by anyone — a topic specialist, or just a person who accidentally hits the nail on the head — can briefly rise to the top. As Aristotle pointed out, average people put together can be better than good people individually.

Engineering the public square

The problem of social media is how to surface what’s good in a scalable and attack-resistant way.

It has to be attack-resistant, because bad people will try and manipulate your system, for greed, propaganda or lulz. If your algorithm prioritizes people with many followers, a market for followers will arise. If you prioritize upvotes, people will form voting rings. Twitter’s new community notes are already under this kind of attack.

It has to be scalable because the “fragmented web”, which Noah Smith celebrates, is not a good enough solution. Small web forums, personally curated, can be cool, but there are irresistible gains from layering other services on top, like identity verification, upvoting and sharing. Also, these work best if they can cross the narrow boundaries of one small group. Small forums could develop standards for these things, and people have tried, but none have really caught on.² That in turn is because software is a public good. One organization needs to create, maintain and protect these systems.

One of the most successful reputation mechanisms was developed quite early by the internet forum Slashdot. With the tagline “News for Nerds, Stuff that Matters”, Slashdot got huge during the first internet boom. Before Kim Kardashian broke the internet, there was the Slashdot effect: a linked headline on the homepage could take down a server just by the volume of traffic. As the number of users expanded, discussion quality became… variable. There were smart and knowledgeable comments, but also the famous Slashdot trolls. They rushed to type “first post” on a new story; posted links to a photograph so infamous that, like Voldemort, it must not be named; drew ASCII art of a small bird inexplicably perched on a penis. (Today it all seems so quaint.) To fight this the forum developed moderation. Anyone could vote comments up for being interesting, insightful or funny, or down for being a troll, flamebait, spam or irrelevant. And to stop that system being abused, there was metamoderation: every day a few randomly elected metamoderators could vote on whether other people’s votes were fair. The system kind of worked. Interestingly, Twitter’s “community notes” now have metamoderation built in: a note moderates a Tweet, and users can vote on whether notes are useful.

Once the basic pieces are in place, an adversarial system can arise naturally. Some people earn reputation for what they say. Others are listened to because of their real-world expertise, achievements or position. Others become trusted intermediaries at relaying the “good stuff”.

Evaluating the two systems

The adversarial versus inquisitorial metaphor helps to get past the idea that social media is a “free-for-all”. Sure, anyone can say anything! But actually, well-designed systems will let an influence aristocracy develop. That doesn’t mean that the influencers will be the people you personally prefer. No design will prevent people listening to pop starlets, phony life coaches, or populist politicians. We cannot make people smarter than they are.

In law, the advantage of the inquisitorial system is that one person, the judge, has responsibility. Ideally, they will be fair to all sides. The risk is that this one person might be biased. The advantage of the adversarial system is that it gives both sides an incentive to make the best possible arguments. The disadvantage is that they also have an incentive to make dishonest or manipulative ones.

The same holds true in media systems. The ideal of journalistic responsibility is meant to discipline New York Times or BBC reporting. When they get it wrong, perhaps like last week, they are criticized for it, which just shows the high expectations we have of them. Whereas “someone on the internet is wrong” is so unsurprising it’s a meme.

But media organizations have biases, which are especially insidious when they are shared. You cannot watch the American media around election time without knowing what side it’s on, less by what it says, which is usually scrupulous, than by its very obvious silences on one side of the story — the corruption of Hunter Biden versus Donald Trump, say. In the old inquisitorial system, pointing out the biases of the mainstream media was a job for lonely extremists in academic departments. Now, shouting at the MSM has gone mainstream itself. An adversarial media system stops any one group getting a monopoly of the truth.

The adversarial system doesn’t just exist in the courts. The idea of getting at the truth by the clash of competing interests is a part of Enlightenment culture, especially of its Anglo-Saxon variant, as informed by the experience of the coffee house and the broadsheet. Perhaps free speech, which has worked so well over the past three centuries, will fail in the twenty-first, and be turned by our enemies into a weapon against us. But we cannot really discard it. A commitment to it is baked into our societies.

The risk of social media is that it gives the algorithm designer a hidden power. You think you are seeing and responding to what you chose, but really the algorithm may be subtly adjusting your world view. That is dangerous, whether it is optimized to rile you up or tranquilize you. Meta has chosen to de-emphasize politics and disagreement on Facebook and in Threads, after screwing up very badly in the 2010s. But our public square has to live somewhere. The adversarial metaphor partially describes how social media works today, but it can also act as an ideal to help us evaluate and improve it.

---

If you’re interested in how communications tech can change society, you might like my book Wyclif’s Dust: Western Cultures from the Printing Press to the Present. It’s available from Amazon, and you can read more about it here.

You can also subscribe to this newsletter (it’s free):

1. I am quoting from memory, since I couldn’t find the book text online.

2. Notably, mastodon hasn’t. Bluesky has a commitment to allowing alternative service providers, but nobody bothers to use them.