When Everyone Knows That Everyone Knows . . .

In my game theory classes, I show part of a 1961 Peter Ustinov movie, “Romanov and Juliet.” The gist is that a small country (Concordia) must pit the Soviets and Americans against each other for survival, using only diplomacy. The Concordian president (Ustinov) discovers a small piece of important information and immediately tells the Americans. “We know that, of course! Have a bourbon!” says the ambassador. Next stop is the Russian embassy, to tell that ambassador that the Americans know.

The Soviet is scornful. “We know they know; we’re counting on that! Have a vodka!” Ustinov goes back and forth, getting more and more tipsy; eventually, he tells the American, “You know, the Russians know you know that they know that you know that they know you know.”

The American jumps up, thunderstruck. “No! That can’t be!”

Ustinov gapes, drunkenly. “You . . . din’ know?”

“No, we had no idea! I have to advise the Joint Chiefs immediately. Have a bourbon!”

This little vignette illustrates the importance of “common knowledge,” the deceptively simple idea that in many situations our ability to coordinate actions and make plans depends on our beliefs about what the beliefs of others are, and vice versa. In a “game,” I need to know the available strategies, how my opponent values those strategies, and I need to know that my opponent knows that I know those things.

Steven Pinker’s book generalizes and expands on the importance of common knowledge as the basis of trust, or lack of trust, the source of “cancel” culture, and a delightful variety of other things you might not expect to be connected. A major theme of the book is that social media has made common knowledge both more and less available to society. It is easier and cheaper than ever before to share an opinion about an event and have that opinion reacted to and transmitted around the world in seconds. But the content of the “information” may be wrong, so while it is more “common” on Twitter it may no longer be “information” in the usual sense of being true.

The simplest version of the problem may be the emperor who had no clothes. The lazy tailor tells the emperor that he is wearing an extremely light and elegant suit; the emperor believes this, and prances around outside to show off his new finery.

Everyone looks, and they can plainly see the emperor has no clothes. But does
anyone else know that? I don’t know if you know, and since you expect that I don’t know if you know, then you act as if you don’t know, either: “Hey, Emperor! Nice suit! It really brings out the color of your eyes!”

The example is not trivial, because many social revolutions—at least, before social media—depended on just this mechanism. Nicolae Ceaușescu, Romanian dictator in the 1980s, staged elaborate mandatory public displays of support through workplaces and in schools. Since failure to participate enthusiastically was punishable by imprisonment, or worse, everyone showed up and cheered loudly. Everyone thought that everyone else liked the guy.

The façade cracked in the western city of Timișoara in mid-December 1989. The regime tried to deport a dissident pastor, László Tőkés, an ethnic Hungarian who’d been publicly critical of the government. His congregation and then a broader crowd gathered to protect him, and open anti-government protests spread. Security forces opened fire on demonstrators, killing a number of people (initial rumors wildly inflated the death toll, which helped galvanize outrage further).

On December 21, Ceaușescu held a mass rally in Palace Square in Bucharest, meant to denounce the Timișoara “hooligans” and rally support, the same kind of stage-managed event that had worked for decades. This time, partway through his speech, the crowd started jeering and booing—visible live on state television before broadcasters cut the feed. That moment mattered enormously: it broke the spell. People across the country saw, even briefly, that dissent was possible and that the crowd wasn’t just a compliant mass.

Imagine being in Palace Square on that date. You think everyone else likes this guy, but you hate him. Someone somewhere in the crowd snaps, and starts to boo, loudly. You think, “Really? Someone else hates Ceaușescu. Boo! BOOOOO!”

The problem before had not been that no one hated the dictator; most people felt that way. But they didn’t know others felt that way. And collective action could only happen if many people knew that many people knew they hated the regime; you don’t want to be the first (and only) person booing.

It was like a switch had been thrown, because now everyone knew protests exploded overnight into open revolt, nearly nationwide. Even (or perhaps especially!) the army switched sides and stopped supporting the Securitate. Ceaușescu and his wife Elena tried to flee by helicopter on December 22, 1989, but were captured by the air force. They were given a hasty (frankly dubious) military trial on December 25 and were executed by firing squad that same day.

This story of information and deception was told by my Duke colleague, Timur Kuran, in his 1997 Harvard University Press book, Private Truths, Public Lies: The Social Consequences of Preference Falsification. It is surprising that Pinker doesn’t reference the book, because it has made the notion of common knowledge in social movements a centerpiece of the literature. This kind of dramatic, seemingly instantaneous collapse is exactly what you’d expect in a system built on preference falsification (“hey, Emperor, nice suit!”): it looks rock-solid right up until a small crack in public consensus lets everyone’s true feelings become known to everyone else, and then everyone knows that everyone knows, and will join a revolution. The “solid” system collapses almost overnight.

It may be I am expecting too much, or at least too much focus on politics and political economy, for what is in fact a general interest book. But for readers of this journal, the most interesting questions lie precisely where Pinker spends the least explicit analytical effort: the relationship between common knowledge and the enforceability of laws, contracts, and social norms.

A rule against jaywalking, a property right, a constitutional limit on executive power—none of these functions as a mere private belief. A law must be “on the books,” so everyone knows, and everyone knows that everyone knows. If I know the law, but I also know that no one obeys the law and it is not enforced, then it is not “really” a law.

The transaction-cost analyst will recognize this immediately as a coordination problem in Thomas Schelling’s sense: legal rules function as focal points, and focal points require common knowledge to do their coordinating work. Pinker leans on Thomas Schelling’s work on this point more than any other single source, which is a welcome corrective, since Schelling’s insight—that coordination, not just cooperation, is often the binding constraint on social order—has been underexploited relative to its influence. But there are myriad political examples, and the omission of the Kuran analysis is disappointing.

After the rise of social media, the means of gaining access to what everyone knows became universally available. Tahrir Square, in Egypt in 2011, saw protests routinely organized with hashtags on social media; ditto for Hong Kong in 2020. A change in public perception no longer required the single visible act of defiance, because each of us can look at social media. Norms can collapse not because anyone’s mind was changed, but simply because we learn today that everyone’s mind was already changed, months ago. We start booing, by retweeting criticisms and mocking memes. It is very nearly costless, and very nearly instantaneous.

Pinker gives a nice account of the Justine Sacco episode, where an ill-judged joke became, within hours, common knowledge of a very different sort, marking the first instance of “cancel culture.” Sacco had only 170 followers on Twitter, but during her long flight the world all came to “know,” and reach conclusions, that meant she was fired before she landed. That illustrates how the same mechanism that topples dictators can produce disproportionate, uncoordinated-looking, but actually highly-coordinated punishment of ordinary people for minor transgressions. What Pinker calls the recurrent vulnerability of human societies to punitive mobbing is, from a public choice perspective, a species of the tyranny-of-small-decisions problem: each individual act of piling on is nearly costless, and individually “rational,” given what everyone now knows that everyone knows. But the aggregate outcome can be wildly disproportionate to the offense and resistant to correction by any single actor.

Institutions that rely on due process, proportionality, and appeal exist precisely to interpose friction between private outrage and the instantaneous manufacture of common knowledge; Pinker’s framework helps explain why platforms engineered to accelerate that manufacture are so corrosive of such institutions, though he does not develop the constitutional-design implications as far as economists working in this journal’s tradition might wish.

This is, in fact, the book’s central limitation for a scholarly policy audience. Pinker is a psychologist and linguist by training, and the book reads as a work of cognitive science and popular exposition rather than institutional analysis. He is far more interested in showing that common knowledge explains an enormous range of behavior—indirect speech, blushing, veiled bribes, the awkwardness of a first date—than in working out the comparative-institutions question that should follow: which legal and constitutional arrangements economize on the difficulty of generating common knowledge, and which make the manufacture of common knowledge too cheap, with the shaming-mob and bank-run pathologies as the price? A public-choice reader will want more on how constitutional rules themselves function as devices for stabilizing common knowledge across changes in government—precisely the role Friedrich Hayek assigned to general, abstract rules known in advance, and that James Buchanan assigned to constitutional-level agreement as distinct from in-period political choice. Pinker gestures at this territory (his discussion of money and of political coalition-formation both depend on it) without quite entering it.

None of this diminishes the book’s usefulness to economists and political scientists working on constitutional political economy, law and economics, or the economics of norms. Pinker writes with his usual clarity, and the case studies—financial bubbles, the pricing of currency, the etiquette of hypocrisy, the mechanics of revolutionary tipping points—are well chosen and genuinely illuminating even where the theoretical apparatus is thinner than a specialist would like. Readers of this journal will find the book most valuable as a well-written primer on a concept their own literatures have used for decades without always naming: that laws and norms are not self-enforcing facts but public, mutually recognized signals, and that the same mechanisms that let common knowledge coordinate a peaceful transition of power can, absent institutional buffers, coordinate a mob. The book does not solve the design problem that observation poses. It does, however, state it with unusual clarity, and for that alone it deserves the attention of scholars in the public-choice and constitutional-economics traditions.