Steven Pinker: When Everyone Knows That Everyone Knows...

The Book Club

In this episode of Spectator's Book Club, Sam Leith interviews cognitive scientist Steven Pinker about his new book on common knowledge, exploring how recursive awareness shapes human

Key takeaways

• Steven Pinker defines common knowledge as infinite recursive awareness ('I know that you know') rather than simple widespread belief
• Thomas Schelling's rendezvous problem demonstrates how public salience solves coordination dilemmas without requiring mental recursion

Main topics

• Common knowledge & coordination theory
• Language evolution & public salience

Notable quotes

Common knowledge is the state where I know something, you know it, I know that you know it, you know that I know it...

Conclusion

Pinker's framework reframes human social behavior as coordinated efforts to

Transcript preview

Speaker 1 (0:00) Subscribe to The Spectator and get 12 weeks of Britain's most incisive politics coverage, unrivaled books and arts reviews, and so much more, all for just £12. Not only that, but we'll also send you a £20 Amazon gift card, absolutely free. Go to www.spectator.com forward slash voucher to claim this offer now. Speaker 2 (0:30) Hello and welcome to Spectator's Book Club podcast. I'm Sam Leith, the literary editor of The Spectator, and this week I'm very pleased to be joined by the linguist, cognitive scientist and historian Stephen Pinker, whose new book is When Everyone Knows That Everyone Knows, Common Knowledge and the Science of Harmony, Hypocrisy and Outrage. Steve, welcome. Pivotal concept of this book is common knowledge. What exactly do you mean by that? How would you explain it? It Speaker 3 (0:55) has a sense that's slightly different from the ordinary vernacular sense. It doesn't refer just to something that is widely known, but something that is known to be known. So common knowledge is the state where I know something, you know it, I know that you know it, you know that I know it, I know that you know that I know that you know it, and so on ad infinitum. Speaker 2 (1:12) And that's, I mean, just to start with, that that endless recursion is a problem in cognitive science for you, isn't it? How do we manage that? It's a problem because the... Speaker 3 (1:22) Interesting common knowledge comes from the fact that it is necessary for coordination, which humans do a lot of. Coordination would be two or more people making complementary choices, being on the same page, doing something that has no value other than if other people do it. Driving on the left versus driving on the right would be an obvious example. Or a simple rendezvous. Two friends may not care where they end up other than that they want to end up in the same place. And the reason that common knowledge is necessary is that merely knowing the other person's intention is not enough. If you imagine a rendezvous, let's say, to take an example from Thomas Schelling that initiated this field of research more than 60 years ago. Imagine a couple is separated in New York. They don't have working cell phones. Somehow they have to find each other. It won't do for him to think, well, she likes to go to a bookstore, particularly a bookstore, so I'll meet her there. Because she might be thinking, well, he likes to go to his favorite camera store, so I'll meet him there. And then he would think, well, she knows that I know that she knows that I like to go to a camera store, so she'll go there after all. And each one could be whipsawed by this kind of futile empathy, each one thinking what the other one wants is not enough. They have to each know that the other knows. Schelling suggested they could solve their dilemma by common salience, relative of common knowledge. They might meet at the clock at Grand Central Station, even if it wasn't particularly close to where they'd been separated, even if it wasn't a favorite haunt of either of them, simply because it was salient enough that each might anticipate they would pop into the mind of the other and vice versa and would solve their problem. Anyway, to get to your question, and I apologize for the background, given how important common knowledge would seem to be in a lot of human affairs, paper currency, institutions like governments and nonprofits and schools, any case in which humans coordinate, how do we do it if the logicians tell us that it requires a hall of mirrors of thoughts about thoughts about thoughts, particularly since that kind of thinking is notoriously taxing, that your head starts to spin as you start to think of like two or three levels of I know that she knows, let alone an infinite number. The likely solution is, as with the meeting at Grand Central at noon, something that is publicly salient, that is conspicuous, that you see at the same time that you see other people seeing it. Yeah. I Speaker 2 (3:50) mean, Speaker 3 (3:50) how Speaker 2 (3:50) would you explain the word salient to this? Salient Speaker 3 (3:54) meaning it pops out. But it has to be even more than that. It pops out in a way that you know that it pops out to others at the same time. And that can give... common knowledge at a stroke. You could reel out the implied I knows that she knows. But if you have a simple mental state that says it's public, it's out there, it's common knowledge, then that gives you the benefits of coordination. You Speaker 2 (4:19) make a distinction in the book between public knowledge, common knowledge and private knowledge. And this has a really strong impact, doesn't it, on the way we react. And in a classic instance, I think you use the, where the emperor has no clothes. That's how I Speaker 3 (4:32) begin the book, because it's a story about common knowledge, exactly. The catch in the story is that when the little boy said that the emperor has nothing on, he wasn't telling anyone anything they didn't already know. They knew he had nothing on, but he was still changing the state of their knowledge, which is why that could be the climax of the story. Namely, by blurting it out with an earshot of everyone else, what everyone... privately knew, at that moment they knew that other people knew it. And so that not only was a different state of knowledge, but crucially to the story, it changed their social relationship to the emperor from obsequious deference to ridicule and scorn. And an idea that... I introduce in the book, I like to think adding to the literature on common knowledge, is that our social relationships are also coordination games, like a rendezvous, like respecting paper currency, like driving on the left or driving on the right. Two people are friends, not because they've signed a contract. What does it actually mean to be friends? Well, it kind of means I know that he knows we're friends, and he knows that I know that we're friends, and I know that he knows. Or lovers, or a... dominant and subordinate or transaction partners. The stable social relationships that we have are coordination games that are ratified by common knowledge. And going back to the story of the emperor's new clothes, even though everyone knew that the emperor was ridiculous, each one could think they were the only one who knew. Once it was common knowledge, they could no longer maintain that hierarchy Speaker 2 (6:02) of status. And this is a kind of extremely generative idea, isn't it? I mean, Because the Empress of No Clothing, as you say early on, this is relevant to revolution. I mean, the Arab Spring, there was this whole thing about essentially a game theoretical. Everyone knows it sucks to be living under the Iranian theocracy, but nobody wants to be the first two or three people to go out on the street and say so. Yes, I quote the Gandhi character Speaker 3 (6:27) from the eponymous movie where he, at one point, he tells a British colonial officer, in the end you will leave because 100,000 Englishmen simply cannot control 350 million Indians if the Indians refuse to cooperate. He could have said coordinate. So that's the dilemma of people under an autocracy. Everyone might despise the regime, but think that everyone else adores the dear leader. And that's why regimes can be vulnerable to a public demonstration. If people turn out en masse in a public square, then everyone can see everyone else seeing everyone else. And that gives them the safety in numbers that could be sufficient for overthrowing the regime, whether by literally storming the palace and outnumbering the regime or by bringing the apparatus of the state to halt through work stoppages and strikes, as Gandhi advocated. You mentioned that the concept of common knowledge has broad sweep, and I collect many instances where it drives our humor, our politics, our economics. Just to give an example of a joke that depends on common knowledge from the Soviet era. A man is handing out leaflets in Red Square, and of course the KGB arrest him, bring him back to headquarters, only to discover that the leaflets are blank sheets of paper. And they confront him. What is the meaning of this? He says, what's there to say? It's so obvious. Now, that is a joke about common knowledge, because what he was doing was subversive. Namely, he didn't have to tell people what was wrong with the regime. The mere fact that in public he was saying that there was something, a common grievance. And when people saw other people taking the leaflet, it ratified their own private knowledge. And indeed, in the joke, the KGB would not allow that to pass. And in real life, Putin's security forces have arrested people for carrying blank signs. Speaker 2 (8:13) Yes, it kind of went from Dvarslova to simply a blank sign. Speaker 3 (8:20) Blank signs, which has also happened in China. It's happened in a number of Speaker 2 (8:24) autocracies. Yeah. There is a kind of, you know, deep game theoretical sort of mathematical structure underpinning all this. Towards the beginning of the book, you talk about things like, you know, Hawk Dove games and this Nash equilibrium. How much or how perfectly does that map