Steven Pinker Explains Cancel Culture, Common Knowledge & AI

Steven Pinker just told me something that honestly blew my mind. He said that Malcolm Gladwell's recent cancellation attempt was not only predictable, it was mathematically inevitable.

There is another thing. As a social media shaming mob, why do people feel the urge to pile on and collectively punish someone who says something that seems to violate some moral norm?

And here's the kicker. Pinker started coming years ago using something called common knowledge theory.

Common knowledge means 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, ad infinitum.

Picture this. You're in a meeting where everyone privately thinks the boss's idea is stupid, but nobody speaks up. Then one person finally says what everybody's thinking and suddenly the whole room erupts in agreement. What just happened? According to Steven Pinker, you just participated in one of the most powerful forces shaping human civilization. And today we dive deep into it with one of the foremost public intellectuals of our time. Let's go. Professor Steven Pinker, welcome back to the into the Impossible podcast.

Thank you.

In this book you've written that civilization itself depends on common knowledge. But my first question to you is what happens when that common knowledge is wrong? Or even when technologies like artificial intelligence threaten us by hallucination? What happens to the very foundations of what we thought was secure?

The answer is we don't know. Track record in general of predicting the downstream consequences of, of technology are poor. Even the best superforecasters are at chance for cut and dry questions five years out, let alone open ended questions like what will happen to society, let alone civilization. I don't know if when social media were introduced 15 years ago, if people could have predicted social media shaming mobs, the spread of disinformation, the, the encouragement of conspiracy theories. I bet people thought Facebook was a rather benign and minor phenomenon. Likewise with the effects of the Internet, of cable news programs, let alone electricity, steam power, the automobile, the telephone and telegraph, television and so on.

Along those lines, it came to me to mind when I read these books and your wonderful writing. Throughout the years, I always try to form, you know, kind of the opposite, the strawman version of it. But to me, the opposite of common knowledge is sort of expert knowledge. And there's a term for this which is ironic because it's sort of tautological. It's a shibboleth. A shibboleth is sort of an abstract term or something of art that only the insiders really know about. It's also A shibboleth. To say that what a shibboleth is, of course, it's tautological.

Why is there, you know, such a prevalence nowadays in the age of reason and science and rationality as you promote? Why do conspiracy theories, UFO sightings, why do they spawn such explosive, you know, violent rhetoric? Is that because of this chasm, this, this gap between expert knowledge and common knowledge, that people abhor the vacuum and rush to fill it in?

I'm not so sure. So the common knowledge as the theme of the book, when everyone knows that everyone knows, the starkest contrast is with private knowledge. That is, what I really explore is not a lot of people knowing something versus only a few people knowing something, but rather the difference between everyone know something and everyone knowing that everyone knows it. That's really what the book is about. Not so much about who knows what, but about that vital difference. Because that difference, it turns out, drives a lot of phenomena. It is, to begin with, necessary for coordination for two people being on the same page, meeting at a certain time and place, observing certain standards, holding up certain norms, belonging to following conventions like paper currency, driving on the right the words of a language. I suggest that it also drives our social relationships, that what it means for people to be friends or lovers or superior and subordinate or transaction partners is that each one knows that the other one knows that each one knows that the other one knows that they have that relationship.

And the phenomena that I talk about are not so much a few people knowing something versus a lot of people knowing it, but rather everyone knowing something versus everyone knowing that everyone knows it. And I suggest that many cases we try to prevent common knowledge. Even if all of us share the same knowledge. We tend to ignore the elephant in the room. We resort to innuendo and euphemism so we don't blurt out exactly what we mean. That's the main contrast. Now I can also say a few words about conspiracy theories and quack cures and UFOs and paranormal woo woo. It's a different phenomena, one that I actually spent more time discussing in my previous book, Rationality.

But it does make an appearance in this one because I have a chapter called the Canceling Instinct on why in academia of all places, the the forum where you think people express their ideas to find out which ones are true, which ones are false. But there has been so much censorship and canceling punishment of speech. Why don't scholars and scientists accept that some things may be true or some things may be false, but we gotta hear them to find out what they are there, and therefore to determine whether they're true or false. And I suggest this is an answer to both questions. Why do people believe in so much woo woo and flim flamingo and crack pottery? And why are academics sometimes afraid of ideas being expressed? Why aren't they just in pursuit of disinterested truth? I think the answer to both those questions is that when it comes to belief about big, abstract, important cosmic questions as opposed to day to day life, when it comes to is there food in the fridge? Is there gas in the car? Are the kids clothed and ready to school? I think we're all pretty rational. We kind of have to be. But when it comes to questions like why are there depressions and recessions? Why are there pandemics? Why do bad things happen to good people? Who really drives the currents of society? What really happened at the founding of our republic? How did our species come into being? All these big, fascinating, difficult questions. And of course, the subject matter of a lot of academic research.

People's core gut feeling, their deep intuition is you can't know, no one knows, and you can't find out. And that, of course, was true for most of our histories as species. It was only the last couple of hundred years, with the scientific revolution and the enlightenment and data science and statistics and data sets, that we could even ask questions like why are there accidents? Why is there disease? But our brains aren't prepared for the possibility that these are objective questions of truth and falsity for which we have the means to determine them. It's rather no one knows, you can't find out. So what we should believe is what makes us feel good? What are the uplifting myths? What's good for the morals of children? What makes our side look good and the other side look bad? And true or false? Well, you can't find out anyway, so you may as well believe what does the most good.

Today's episode is sponsored by Shortform, the smartest way to keep up with the world's smartest ideas. Look, I'm a professor, podcaster and parent, and I don't have hours to spare. When I needed to prepare for a guest like Steven Pinker, I start with short form first. I started with their chrome extension, which gave me crisp AI summaries of his articles and past books. Then I dove deeper. As soon as I got the basics out of the way for today's interview, I leaned on their guides to Pinker's early, earlier books, the Better Angels of Our Nature, Enlightenment now, and Rationality. I was on a deadline to release this episode the same week his book got published. So short form was indispensable.

I needed to make the connections fast, and they didn't disappoint. Steven's books reference other leading luminaries, like Daniel Kahneman's Thinking Fast and Slow and Richard Dawkins the Selfish Gene, perfect for seeing how psychology and evolution shaped the bigger picture. But I didn't have time to read all those books and Steven's newest book to stay on deadline. Shortform bailed me out. But it's not just books. Shortform ads, articles, podcasts, and even AI generated summaries. I'll use it for research on AI hype cycles, for catching up on the latest astronomy papers, or, yes, even sometimes for parenting questions. Diving into their podcasts and articles.

It's become my daily driver, my curiosity companion, helping me see save hours and hours a week. And it will for you, too. You can try it for free, and if you go annual, you'll get three extra months on me. So whether your appetite's been wet to revisit Pinker's earlier works or you just want to explore fresh insights on AI psychology and evolution, short form's got you covered. Start your free trial plus three bonus months@shortform.com Impossible or click the notes below. Now back to my conversation with Steven Pinker. That reminds me of the, you know, kind of foremost conflict in my branch of physics, podcasting, which is around consciousness and free will. Nowhere is sort of the tacit assumption that free will doesn't exist meets the kind of hard fact that no one who believes that we don't have free will acts as if they don't have free will, nor knows anyone who's not a psychopath who acts as if they don't have free will.

Why do you think that is? I mean, why do people Sam Harris I'm thinking about Robert Sapolsky. The only one I interviewed that, you know, sort of believe that we do have free will was the late, great Dan Dennett. What do you make of this concept that everybody in the erudite academic circles, they believe in something called super determinism, that it's all basically founded in the underlying laws of nature, that I have no ability to not think about a white polar bear, as you point out in the book. But in in practice, they always act as if they have free will and can't point to a single normal, rational person who acts as if they don't have free will. Why do you think that? Is that cast?

So there are A number of things. You know, on the one hand, it's a deep sacred moral belief of anyone with a scientific mindset, including I assume me and you, that you know, there's no mystical, paranormal, extra scientific causality in the universe. No, no magic, no occult forces. What exists is what our best science says exists, and that a lot of phenomena that seem inexplicable paranormal are just things that we haven't nailed down yet or the result of human folly and misperception. Remember, if you think dreams foretell the future, it probably means you're remembering the few anecdotes in which by coincidence, you dreamed about something that happened. You forget all of the many, many times you dreamed about something and it didn't happen. And so people over interpret coincidences, but anyway, so what that means is there's no room for some miraculous soul separate from the laws of cause and effect that somehow manage to pull the levers and push the buttons of behavior separate from physical processes in the tissues of the brain. So that's a deep conception among naturalists, scientists.

And so the idea that there's something that somehow escapes the laws of physics, not just molecular motions in the brain, doesn't sit well. And they're adamant that that can't be true. Now I think they sometimes overextend it, that mindset which I share to a hostility to a more nuanced and reasonable position of free will, which is that there's a lot of sheer indeterminism, or if it's deterministic, it's not predictable either because of nonlinear dynamics. Maybe for all we know, things at the quantum level of Brownian motion of particles in the brain. But even if it isn't that, it's just so complex that you just can't predict everything that a person is going to do and that kind of looks like it escapes the laws of cause and effect. It may just be they're too complicated for everyone to figure out. There may be, as Dan Dennett suggested, even a designed random number generator in the brain to make us unpredictable so that we can't be gamed. So just as when you have an outguessing standoff, and I talk about this in the book, does the hockey player shoot left or right? And does the goalie defend left or right? And when you play scissors, paper, rock, you try to display exactly what the other guy isn't guessing.

And we know from game theory that the optimal strategy in an outguessing standoff is to be random. If you're not random, if there's Any kind of tell or statistical predictability, your opponent can use it to your disadvantage. So there may even be in the brain some randomizer that makes our behavior unpredictable above and beyond the fact that something this complex as a brain, it's unlikely that every last twitch will be predictable by some antecedent. So there's that. There's also a. So anyway, there's, I think, an illusion that behavior is free of the laws of cause and effect that some naturalists, I think, try to explain away as if, or wish wasn't there, as if everything was ultimately predictable. And I think it's perfectly acceptable to use free will for these not entirely predictable processes. Although free will.

Nor does free will actually mean random. If someone really behaved at random, we wouldn't call that free will either. We call it crazy, but it's not completely predictable. I think the other moral conviction that leads people to insist that there's free will, even if they haven't thought through exactly what that means is the idea that we can hold people responsible, including ourselves. That you can't just say, well, it's my genes that did it. It's my evolutionary history, it's my neurotransmitters. You know, it wasn't me. I can't be punished.

And I don't have to esteem or reward someone for their accomplishments because it wasn't them, it was their brains, whatever that would mean. I think Dan Dennett in his book Elbow Room, the Varieties of Free Will Worth Wanting, was pretty clear on that confusion. If anything, holding people responsible. There's a sense in which the last thing you want is complete unpredictability. That is complete randomness. Because then you really couldn't hold anyone responsible with the hope that holding them responsible will affect their behavior because they just blow off your credit. Your blame, your approval, your disapproval, your rewards, your punishment. You wouldn't be able to get someone to stop from robbing a liquor store or sexually assaulting someone by the threat of opprobrium or actual criminal punishment because they could do whatever they want.

They have free will. We're kind of hoping they don't have so much free will that they could blow off what everyone thinks of them and therefore do whatever they want. So I think because of the desire to uphold a moral order via holding people responsible, people think that that needs some concept of free will. I think mistakenly, because it's quite the opposite, it's predictability that makes holding someone responsible itself a cause of them behaving in ways that we want.

Part of what you're mentioning really comes to the fore when I asked people like Robert Sapolsky. I said, well, you know, what if, God forbid, your dog was run over, you know, intentionally, gruesomely, you know. And he said, of course I'd want the murder, you know, I'd want the murderer, you know, punished or whatever. My dog. And all the more so, as we say, you know, for a human being, of course. But before I get into, you know, some of the original kind of work on this topic of free will and consciousness done by, of all people, my hero, Galileo Galilei. Come back to him later. The so called, what's called Galileo's Error.

I do want to get into the thing that you're never supposed to do until you generate enough common knowledge about something, which is to judge a book by its cover. Hey, book lovers, we're judging books by the covers. We know we're not supposed to do it, but into the impossible. There's nothing to it. Let's take a look and judge some books. Stephen Taylor. Take us through this book. The title, the subtitle and the COVID illustration, please.

When everyone knows that everyone knows common knowledge and the mysteries of money, power and everyday life. So common knowledge in this sense, it's actually a technical term. So it's not the same. Exactly the same as what people usually refer to when they talk about common knowledge. It's almost the opposite. We usually ordinarily use common knowledge to refer to, like, an open secret. Like it's common knowledge that the police can be bribed around here. It's almost, you know, sad to say.

It's just, it's a technical term that stuck. I don't like it. There's nothing I can do about it. That's what it's called in the literature. So in the technical sense, common knowledge means 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 ad infinitum. That's why the COVID illustration has a thought balloon inside a thought balloon. Now, two immediate comments.

One is, that sounds impossible. My head starts to spin with two layers of I know that she knows, let alone an infinite number which can't fit in a finite brain. And there are a lot of allusions in popular culture to how hard it is to keep track of embedded levels of thoughts about other people's thoughts, about your thoughts, about their thoughts. There's an episode of the situation comedy Friends in which Rachel says to Joey, they don't know we Know they know we know. And Joey, you can't say anything. And he shrugs and he says, I couldn't even if I wanted to.

Don't know that we know they know we know. Joey, you can't say anything. Couldn't if I wanted to.

The humor coming from the fact that there are some times in life when there's deception, when there's bluffing, when there's humoring, there are mysteries where it's hard to keep track of who knows what, although we often try. But the reason that common knowledge can still be a coherent concept even though you can't literally think, in fact, affinity of thoughts is that we intuitively sense it. When there's an event that is self evident or public, which we experience as it's kind of out there, you can't take it back. So if you see something and you see other people seeing it and they see you seeing it, that's kind of all you need to generate common knowledge. At a stroke. We can go back and forth between peeling out the layers of I think that she thinks and appreciating it in one stroke when something is just public, unignorable, undeniable. Now why does this matter? Well, it's because common knowledge is necessary for coordination for two people being on the same page. So if a couple is separated, for example, and the cell phone of one of them is dead, so you can't generate common knowledge through language.

It really isn't enough for each one to know where the other one is going to go. If he thinks that she's going to go to the bookstore because that's where she who likes to hang out, that's not going to work because she might think that he's going to go to the camera store because that's what he likes to come out, hang out, and he can't even think, oh well, she's going to go to the camera store because she knows that that's where I like to hang out, because she also knows that I know that she likes to go to the bookstore. So she's going to go to the bookstore after all, until it dawns on her that I'm going to know that she's going to think, et cetera. Nothing short of actually generating the knowledge that you know when you know that the other person knows it can ensure coordination. So that's a very simple example, but there are more complex examples. What makes money valuable? It's just a green piece of paper. Well, it's valuable because I know that other people treat it as valuable. Why do they treat it as valuable? Because they know that other people will treat it as valuable.

Conversely, you can have phenomena like hyperinflation, bank runs, bubbles, crashes when people lose confidence in the value of some commodity or entity, Political power. What makes a leader a leader? Well, they probably can't train a gun on every last citizen. They would want to and often they don't need to. My department chairman doesn't monitor my comments conversations 247 and fire me for insubordination if I, if I blow one off. You just know and I don't have to think that way, I just accept that, you know, she's the department head and she's got some authority over me and she knows that I treat her that way, et cetera. And then as I mentioned earlier in the conversation, a lot of our informal relationships, what makes two people lovers, what makes them friends, is a matter of common knowledge. Each one knows that the other one knows that they know. So that's what the topic of the book, that concept of common knowledge, well known in certain sectors of academia, in game theory, in economics, among some political scientists, surprisingly not as well known in my own field of psychology, considering that it is a psychological phenomenon.

It's about knowledge known among linguists because language generates common knowledge. I know that you're intending for me to believe something when you explicitly. Anyway, I wanted to put it all together and what's inside. Oh, is it right side up? Yes, all of the outer thought balloon is a sample of the phenomena of money power in everyday life that I talk about, including blushing, tears, cancel culture, sating face, battle of the sexes, games of chicken, sexual come ons, hyperinflation, norms falling through the cracks, bystander apathy, plausible deniability, shaming mobs, white lies. All of these phenomena I think you can't understand without grasping the concept of common knowledge.

Yeah, and the more you see it, the more you notice it. It's sort of this self reinforcing feedback cycle where I was walking across the street yesterday and there was not a stop sign and I looked at the car and the driver, did they want me to go first? Do I go first? And you actually mention that as an example in the book you also mentioned a political example, a famous one of the recursion for the detriment, but also illustrating the dark humor of the denizens of the former Soviet Union where they would say, say the saying, we know that they're lying, they know we know they are lying, but they keep lying anyway. And actually I think the original Quote.

Maybe it's a Solzhenitsyn. I think what's crucial there, at least the comment I make on that is that even though he brings it up to several layers of embedded thought, like, we know they know we know they're lying, and still they continue to lie. How come? Well, it's because even those two levels, three levels are short of common knowledge. And so if you were to say in the press, holding a sign, their line, you'd be off to the gulag. So there really is a difference between common knowledge, that is saying something in public, so everyone knows that everyone knows that everyone knows something. And even several layers, finite layers of knowing that they know, short of common knowledge. And speaking of the. The former Soviet Union, it's also, I think, beautifully illustrated by an old joke where a man is handing out leaflets in Red Square, and of course, they're kgb, arrest them.

They take him back to headquarters only to discover that they're. They're blank sheets of paper. They confront him, what is the meaning of this? He says, what's there to say? It's so obvious. The point of the joke is he was being subversive because he was generating common knowledge. In a dictatorship, everyone might be aggrieved by the regime. They may know that it's cruel, repressive, inefficient, but they don't know that everyone else knows it because dissent is punished. Everyone falsifies their preferences, even if they suspect being human and kind of knowing what other people are like, that they're also discontented. They don't know that other people know that they know or that enough other people know.

And so what dictatorships try to do is prevent common knowledge. They engage in censorship. They repress public protests where everyone could see everyone else seeing everyone else, simply because if millions of people coordinate their resistance, they can bring the government down. No government has enough firepower to intimidate every last one of its citizens. I quote the character Gandhi in the. In the movie by that name, where he tells a British officer, in the end, you will leave. Because 100,000 Englishmen simply cannot control 350 million Indians if the Indians refuse to cooperate. So that's the subversive power of common knowledge.

Generating demonstrations, protests, even blank leaflets had a blank leaflet saying, we all know, and we all know that we all know there's something to be discontented about.

Yeah, the lack of words leapt from the page in that example. The Catholic church of the 17th century was no stranger to censorship, and they had sort of ultimate power, but One event that was, you know, pointed out to me by your late great colleague Owen Gingrich was the notion that the Catholic Church didn't actually forbid my hero, Galileo. As I said, I'm proud that I did. The very first ever audiobook of Galileo with Frank Wilczek and Carlo Rovelli and many other luminaries. But the. The ultimate sin that they forbid him to do is to teach it. So the Sidereus Nuncius was written in Latin. That was not a common language of.

Of ancient Rome or not ancient Rome, of Renaissance Rome. Right. So that meant that you could do research which was conversed in Latin. And they didn't have a problem so much with the Sidereus Nuncius or work after that he did in Latin, but they had a big problem with the Dialogo, which was written Italian, obviously. And so what do you make of this? That they. Did they make a mistake? Were they. Were they actually trying to prevent sort of a massive, you know, at that time, we didn't have evidence. Galileo kind of fudges a little bit the evidence for heliocentrism.

It actually wasn't proven by his favorite method of the earth's tides. That was shown to be wrong.

Again, bogus theory.

Completely totally bogus. So he was right for the wrong reason, and yet he's portrayed as martyr for free speech and suppressing. The Church is guilty of suppressing, you know, what should be common knowledge. So where does he fit in in this grand scheme?

It fits in exactly because what they were repressing was not expression, but common knowledge. That is teaching it in the same way that the dictatorships like the Chinese Communist Party, which intensively surveils people's electronic communication, not for criticism of the government. In fact, they use criticism of the government as a way to run the government more efficiently, to find out if they're failing to kind of mollify the masses as it's kind of useful intel. But what they do repress and come down like a ton of bricks is anything that would coordinate people's behavior. Meetups, demonstrations, coordinated protests. Why they're sharing knowledge. That's what they can't tolerate. Likewise, protests in public, where people can generate common knowledge by all seeing each other at the same time and place dictatorships clamp down on that.

And, of course, censorship in the press. Again, the reason is, in the case of the Catholic Church, their power depended on their authority enforcing certain dogmas, that being the only pipeline to knowledge. So what Galileo was doing was. Was deeply subversive if people heard about it and knew that enough other people had heard about it and in turn knew that because that would get allow people to in cahoots, challenge the authority of the church if they thought they were the only one to know how the universe worked or the solar system, at least there's not much they could do with that knowledge.

About 10 years ago, I saw a tweet when it was still Twitter, which you're very active on. We'll link to you on Twitter. The, the tweet was from none other than Elon Musk, who is the now proud owner, maybe we could say, of that platform now rebranded X. And he said, enjoy Twitter. Something to the effect of enjoy Twitter, you know, post a tweet and lose your job in five years. Just last week, you know, our mutual, you know, sort of public intellectual Malcolm Gladwell was, you know, kind of subjected to a modern day cancellation, or it's become known as cancellation, thanks in large part to Twitter causing this phenomenon. Because if he has admitted that he had been misleading people about his knowledge, which he should have been better about, about women, men competing in women's races, et cetera, I don't want to get into the details of what he was accused of doing, but the mere fact that he had admitted that he actually revealed preference or revealed the common knowledge that he suppressed, why do we do that? Is that, what is that sort of a general phenomenon that could have been predicted due to psychology and the human brain, brain not changing in hundreds of thousands of years? I mean, could you have predicted.

There is some of that and I think, I think Gladwell deserves enormous credit for publicly changing his mind and anticipating and enduring the consequences. But none of this is infallible. We need more people saying, I take back what I said. I've been persuaded. I was intimidated at the time, for which I also apologize. Ironically, Gladwell was punished for exactly the wrong reason. But there is such a thing as a social media shaming mob. And I have a chapter and a discussion in another chapter on that phenomenon.

The chapter is called the Canceling Instinct, as why do people feel the urge to pile on and collectively punish someone who says something that seems to violate some moral norm? Now, part of it we discussed earlier, the very idea that venturing an opinion should be evil, condemned, immoral, I think comes from this feature of our psychology that we deep down don't really believe that there's truth or falsity when it comes to big questions. But then when there is what is perceived as a breach of immoral norm, there is, since moral norms, social Norms are only exist to the extent that people know they exist. They aren't enforced by Big brother surveilling you 24, 7. Other people will condemn you, will ostracize you. And you know that. And they do it because they know that everyone else expects it. So when a norm is flouted in a public forum where everyone can see it being floated while they see other people see it, that triggers the urge to prop up the norm by punishing the norm breaker again in public. And the norm is policed.

It's reinforced only when it is publicly punished. That's why traditionally criminal punishments were done in a public square. Public hangings, crucifixions, disembowelments, even if they weren't that gruesome, like pillories or stocks, it had to be in a public place. Social media give us the equivalent of the stocks or the pillory or the public hanging in an electronic format. Namely, you pile on in a forum where there's some good chance that it'll go viral and everyone will see the punishment. Therefore they will not be tempted to float the door.

You know, reading this book, you cite studies by Nobel laureates in economics, there is a logically non stochastic, completely predictable or calculable, non complex method of reaching agreement as long as both actors are rational. So why do you say that common knowledge makes it impossible for rational people to disagree? Why is that?

This isn't me. The credit goes to the Israeli mathematician Robert Alman, who won a Nobel Prize, although not for this, even though he is, as with many Nobel Prize winners, the work that doesn't win a Nobel Prize is also brilliant. That's right. So this is a simple theorem. You can say. I'm going to say it in a couple of sentences. The proof is actually pretty short, but as Alvin himself put it, not completely obvious. So here's the theorem.

It's called the agree to disagree theorem, whose result is you can't. Rational agents cannot agree to disagree, which sounds very intolerant and counterintuitive. But here's the theorem.

Irrational.

Yeah. Two rational agents who have the same priors in the Bayesian sense of a degree of credence in a hypothesis prior to examining evidence. So if they come to a hypothesis with the same priors and their posteriors are common knowledge. Again, posterior in the Bayesian sense of degree of credence after having looked at the evidence, then those posteriors must be the same. It's not that. If I show you all the reasons I have modified my belief, you must Be persuaded by the reasons that you can kind of swallow it. Just all I have to do is give you on a scale from 0 to 1 or 0% to 100% how much I now believe it and you do the same. And we each know each other's posterior well, then they have to be the same.

That is, if they differ, we've got to bring them into alignment if we are rational. Therefore rational agents cannot agree to disagree while remaining rational. It's both intuitive and unintuitive. You know what's unintuitive is that you don't need to look at the other guy's evidence. All you need to do is look at the other guy's degree of confidence. What's also not intuitive is that if you do have two agents who start off disagreeing, they come into convergence not by getting closer and closer and closer, but rather it's a random walk. They might completely flip their belief. They might leapfrog each other, they might lurch randomly until they end up at the same place and they stay there.

Which is kind of not the way human arguments tend to go, typically. Although Scott Aronson, the computer scientist, says that actually mathematicians sometimes do behave that way when they're arguing about digging a truth of the theorem with the soundness of a problem. It's not the way we usually argue. Usually we. It's more like bargaining. You know, we agree, if we're reasonable people, we kind of compromise and meet in the middle. But Almond's theorem says that that's actually not the rational way to do it. Because for one thing, there's no reason to think that the truth lies exactly halfway in between the opinions of two guys who happen to be arguing.

Also, if you're going to be just moving closer and closer, there's no reason you should do it in dribs and drabs, as if you're trying to get the most out of the other person. That's not a rational way to to land on the truth. You should move your opinion as far as the evidence and arguments warrant and no more or no less. Okay, so that's it's attitude. Except that there is actually an everyday, pretty well known example, and that is investing, speculative investing. If you buy a sound book on what you should do with your retirement funds or your savings. Aside from the con artists and the various people who are out to sell some. But the reasonable investment advisors pretty much say the same thing.

Don't try to outwit other investors. Don't try to time the market. Don't act on a hot tip. Unless you really have insider information, just buy an index fund, a basket of 500 securities, because you can't out predict the market. The reason that you can't is it's basically Almond's theorem, which is that a price is a posterior. It's an estimate of the present value of some security. If there is some actionable information that commends rationally adjusting the price. Given the way that information is transmitted, it's almost certainly common knowledge.

If you know that Apple is coming out with an insanely great gadget, you're probably a lot of other investors know as well.

That's the efficient market hypothesis.

It's the efficient market hypothesis. They've already bidden. Bid it up. The price. The price now being common knowledge, has already incorporated every scrap of available information. Since everyone knows it, if you depart from the posteriors of other rational agents, then you're the greater fool that everyone else is hoping to sell to.

Next, want to turn to the concept of charity? You talk about tzedakah. We'll get into our Talmud, you know, havrusa study. And later on, on after the retention curve flattens out. So what I want to know based on this book kind of brought this to my attention, although I've known it before. You know, there's a reason that all these buildings on campuses have people's names. Hospitals have people's names on it. So my question to you is, what makes generosity so psychologically motivating and powerful? Is charity sort of a signaling mechanism that only works if everyone else knows that they value it as well? In other words, this Johnstone person, you know, is he only sponsoring you as a. As a chairman?

And the Johnstone family. The Johnstone family and Reed and Brett.

I think he's just signaling, Stephen. Are they just. They want to hitch their wagon to your star and they want everyone to know it. And they want they know because everybody values such things, especially at a place like Harvard. Is that the motivation?

There's a bit of a paradox here, because people want to be known for their donations, but they don't want that to be known, that they gave it in order to be known for their donations. The reason that people want to be known is that, you know, each of us likes, Wants to be around generous people, people who give more than they have to and who kind of offer the best deal in life. I mean, you know, we like generosity, we like altruism. And of course, if you're identified as someone who gives something away without expecting anything in return, then it Shows you're a generous person and other people like you. On the other hand, the paradox is that other people liking you is really a good thing to cultivate. And so there's a motive to flaunt the generosity with the named professorship or the named building, or even the acknowledgment at the last line of a program or the little plaque on a synagogue pew that is kind of self defeating. So the ideal state is when you want to be anonymous, but everyone knows about it. A paradox that was worked out by a great Jewish thinker, Larry David, in an episode of Curb youb Enthusiasm in which Larry David, playing himself, donates money for a wing of a new nonprofit building.

He's then the other wing has a plaque saying donated by Anonymous. And then to his chagrin, it starts circulating that the anonymous donor was his social rival, Ted Danson. And so Danson gets the glory both for the donation and for the anonymity, which is, you know, paradoxically, has to be breached for everyone to know that he wanted to be anonymous. So all of this. So it's a subject of humor in that episode of Curb youb Enthusiasm. But it's also a paradox that the medieval Sephardic rabbis debated in the 12th century, most famously in Maimonides Ladder of Church charity, or more accurately, ladder of Righteousness, Tzedakah I learned about when I was 11, and it was my first exposure to the concept of mutual knowledge, knowledge about knowledge and the fact that it seems to matter so much in human affairs. So Maimonides ladder, what he claimed was holding constant the size of the charitable gift. There are eight rungs of righteousness.

The top one is not too controversial. It's if you eliminate the need for future charity by giving someone a starter loan where they can set up a business teaching them a trade, going into partnership with them. You know, teach a man to fish and you feed him for a lifetime. You know, that kind of thing, you know, fine. But it's the other levels that I found puzzling as an 11 year old. So the worst form of charity is the rich men say, puts money into the hands of a poor person. So the poor person, the beneficiary knows the donor, the donor knows the beneficiary, and each one knows that the other one is. So it's common knowledge, common knowledge being the lowest, least righteous form of charity.

The highest form, after teaching someone a skill is double blind. The donor puts money into a secret chest in the temple, goes away, the poor person comes and retrieves it. Donor doesn't Know who the recipient is and vice versa. Kind of double blind. Like a double blind clinical trial that's up there, second highest rung. In between are rungs of different levels of mutual knowledge. So if the donor knows the beneficiary, but not vice versa, that's one rung down. So that's the donor puts money on the doorstep of a poor person in the dead of night, scurries away.

Poor person opens the door, the next morning they get the donation. The donor knows the beneficiary, but not vice versa. One rung below that is the converse. A wealthy person walks around with an open backpack full of money, and the poor person can scamper behind them and pick money out of the backpack. The recipient knows the donor. Donor doesn't know the recipient. That's one rung down. I was kind of puzzled.

Why should it make a difference if you're a poor person, what difference does it make whether the donor is known or not? You buy the same amount of groceries with the gold coins of the cash. The answer that now I'm going to switch to we, we being my students at the time, Julian DeFreitas, Kyle Thomas, Peter De Sholy figured that going back to the beginning of this discussion, why are people. What's the ulterior motive for being altruistic? Is that other people think you're generous. And a person who gives away more than they have to is more generous than someone who gives it grudgingly. And someone who gives anonymously must be really, really generous because there's nothing in it for them. They can't call in the favor. They can't develop a reputation for generosity. So that's the real in the bone, dyed in the wool, altruist.

And that's the guy that I want to be hanging down, not the guy who dribbles out favors in expectation of being repaid in the future. And so we verified in several studies that people more or less agree with Maimonides. They're not so sure about the picking money out of the open backpack. But aside from that, they agree with them on most of the rungs. And we think the reason is that they judge that someone who is anonymous really is more dispositionally generous than someone who is public and therefore could call in the favor.

Getting back to this notion of what is the antithesis of common knowledge, again, I was thinking it's uncommon knowledge or expert knowledge. And I want to ask you, with the risk of expert knowledge being so much more, as I say, risky, perhaps to the possessor or the explicator of the uncommon knowledge. In other words, if I say there's a purple unicorn on Mars's North Pole, that's a very specific claim, and it could in principle be falsified. If it does, that risks tremendous amount of reputational harm, of social shaming, of. Of professional shaming, of course. But there are instances, and since you mentioned the words Thomad Ham, Rashba Rambam, Larry Ben David, you call him in the book. So I feel confident.

Great Jewish thinker and other great.

Yeah, we could do just. And Seinfeld makes many appearances. Matzah balls are flying around. But I feel confident I can ask you just two questions about the Talmud and the Torah. So one is this claim that the. The Torah was given at Sinai and it was given by God. That is a very, very easily, you know, potentially easily falsifiable claim in some sense. There are many contingent claims upon that.

For example, the most famous one is the animals that are kosher. There's only one animal which you and I know is not kosher, that has hooves but does not chew its cud. Those are the two signs that a land animal needs, and that's a pig. Right. So if the Torah, the logic goes, and this is related to the Curzari, which is also, you know, tangentially referenced in this book, so the argument goes that in order for the Torah to basically for a human to make that claim, that there will be one animal with only one of the two signs, unlike, you know, there's many animals that have scales but no fins, and fins but no scales, both of which render them invalid as kosher. So how would a human being 30, you know, 500 years ago, if he wrote it or she wrote it, how are collective of, you know, jw, whatever the acronym is that Richard Friedman likes to use, how would they know that on Antarctica, where I've been twice, you know, that they wouldn't find another animal that, you know, is, you know, kind of a. Isn't that a huge matzo ball? You know, to use the Seinfeld? Like they were really risking this by stating this in a document that could easily have been falsified but hasn't been?

Well, did they make the empirical claim that there are no animals other than swine that have cloven hooves but don't chew their cud? I mean, how embarrassing would it have been? Let's say we discovered a heffalump in Antarctica, there are two mammals you can't eat. Would that have been embarrassing to them? I mean, they didn't claim that the pig is the only one. They're just cleaning.

The pig is the only one that has. In other words, if you eat shrimp or you eat pork, they're both non kosher and so you're violating the laws against kashrut. But the pig is especially earmarked for detestability and it is considered sort of an affront to God Hashem himself. So there is a unique status of the khazir of the pig.

It's hard for me to put myself in the. I don't know if we're talking about say the original, the JW, whoever wrote that 2500 years ago or just rabbis today making sense of it. My sense is that the fact that they're only pigs is kind of a small part of it. However many there are, you can't eat any of them. And it just happens that pigs are the only ones. But actually aren't there other. Well, I don't know, who knows. But I have not enough of a I ungulate specialist.

Your Talmud Hakam days are.

I'm not a Talmud hachim, but maybe.

I'd love to, I'd love to learn.

But I think they weren't me. It's my sense is that this was really a moral or a normative prescription, not so much an empirical or a factual claim.

Well there are, yes, you're right, there are sort of prescriptions against it because in a symbolic sense the, the pig presents itself. Oh, look at my hooves, you know, you should want to eat me. But internally it's neshama, it's soul is polluted and not good for you. So there are sort of emotional appeals and non rational reasons for it, but those aren't in the Torah itself. The last thing I want to point out is, is the Curzari principle which is basically again the claim is that for you have to have been lied to about the divine nature of the Bible, your, your Bubby, you know, would have lied to you and her bubbies Bubby and her bubbie, you know, in other words the infinite chain of, of regress that suggests that there was common knowledge at the very beginning of time. The Torah actually does say, you know, you are the only group, you know, search around and you are the only group that heard the voice of God and lived. And I think in Brahman religions they, they heard the voice of God but they died. So in other words, people would have had to know or thought to believe that they know that everyone in the generation before them believe the same thing ad infinitum.

Back to Sinai itself. How do you explain that in the context of. Of common knowledge, when again, it could be falsified? So what role does this common knowledge play in the mass belief? Or maybe some would say delusions. Etc. How much of an incorrect fact would it take to promulgate or to undo what should be truly common knowledge?

A few things. So I do talk about cultural and religious rituals, which are common knowledge generators within the group that alert people to the fact that they're in a circle of cooperators, of coordinators that do things that benefit all of them, because all of them know that all of them know that all of them know that they do it. So, for example, do you stay home from work on Saturday or Sunday? It doesn't really matter. But it does matter that you all stay home the same day and we've got different pools for whom it's Saturday or Sunday or. When do you become an adult? Is it when you're 18? Is it when you're 13? There's no answer to that question because you don't become an adult overnight. But it really helps to, to forestall endless haggling. Does this guy get to drink? Does he get to drive? Does he get to vote? You gotta draw a line somewhere and everyone's gotta recognize the same line. And you know, and cultures and religions do that and they do it publicly.

There are investiture ceremonies, there are bar mitzvahs. Different people have different coordination equilibria ratified by different common knowledge generators. Sometimes you choose a signifier that just makes it public who's in the group and follows these norms. Who's in that group, follows their norms. Now there's a special role to play, though, in unfalsifiable beliefs, because if there's some advantage to being a hanger on an exploiter, a fair weather friend. So you want to get all the goodies and perks from a group, but you don't really want to do any of the sacrifices. Sometimes there are costly initiation ceremonies that prove you're really committed, like you cut off your newborn's foreskin. But there can also be untestable, indeed, preposterous beliefs.

God is three persons and one person at the same time. Jesus rose from the dead. God wrote the ten Commandments on a pair of sapphire tablets and gave them to the entire Jewish people, including you and me. We were actually there. Those are really hard to believe. If it's just. Do you believe that rocks fall down or rocks fall up? Well, everyone believes that rocks fall down. So what does that prove if you believe that all Jews were actually on Mount Sinai 4,000 years ago, that kind of shows that you're willing to put yourself out there, that you're willing to make a sacrifice, if only in your own rationality and credibility, to be considered a member of the group with all of the perks and coordination equilibria that come from it, the trust, the ability to do business with them, the ability to marry with some of them, et cetera.

Mm. Okay, the last question I have is not about the Torah, don't worry. But you've of course known for your language instinct, your writing on language itself. I want to ask you about this, you know, 64 trillion dollar question, which is that large language models that we all use and love are trained on centuries, you know, perhaps millennia of common knowledge, the billions of words that humans have published online and will publish in the oncoming days, terabytes every day are being added. But if all that training data is full of the same peccadilloes, flaws, foibles, lacunae, half truths, nonsense, the model confidently puts out hallucinations that look like it's a consensus. So I ask you as a final question, does this turn the notion of common knowledge itself into kind of a polluted training set that should be an FDA cleanup site or whatever?

Exactly. Because what it's trained on is it doesn't have to be common knowledge. It just has to be a kind of text that's out there. You know, it could be a Reddit post that, you know, 10 people have seen, and it's perfectly happy to suck that in and absorb the statistical regularities in that too. And the problem with hallucination is not that the input that it's trained on has errors because it generates its own. They generate their own hallucinations. The problem is that they don't have any propositions, any knowledge base, any model of the world. It's just mashed up statistical patterns now, incredibly significant, subtle and complex and abstract ones.

But nonetheless, the models don't actually consult some knowledge base on what is in the world, who did what to whom. It's just what kinds of words occur with what other kinds of words, plausibly, statistically, frequently. And so sequences of ideas, that, or abstractions, I should say, that occur together often enough, get mashed together in new combinations. The odds are that they tend to occur together. The fact that they didn't actually occur in reality is not something that the model actually registers. And that's where the blends and hallucinations and confabulations come from.

Steven Pinker, author of When Everyone Knows that Everyone Knows. Thank you so much for joining us for your second appearance on the into the Impossible podcast. May there be many more. Halavi thank you. Stephen.

Thank you so much. Brian. Pleasure. Bye bye.

Common knowledge shapes more than just our social interactions. It fundamentally alters how we perceive reality itself. If you enjoyed exploring these hidden mechanisms of human coordination, check out my first episode with Stephen right here. Don't forget to like, comment and subscribe. Stay tuned for next week's episode of into the Impossible.