Common knowledge

Why a fact everyone already knows can change everything the moment somebody says it out loud.

July 16, 2026 6 min read

A crowd watches a naked emperor walk past. Every adult can see. Not one of them says a word. A child says it out loud, and the spell breaks.

Notice what the child did not do. The child did not inform anyone. Every person there had already seen the emperor perfectly well; the announcement carried, in the strict sense, zero information. And it changed everything.

The gap between a fact everyone knows and a fact everyone knows that everyone knows is one of the most useful distinctions I have ever been handed, and most people are never handed it. It has a precise name and a real formal history.

Two kinds of knowing

Mutual knowledge: everyone knows p.

Common knowledge: everyone knows p, and everyone knows that everyone knows p, and everyone knows that everyone knows that everyone knows p, and so on, without end.

That infinite tower looks like philosophical showing off. It is not. The first full analysis came from David Lewis in his 1969 monograph Convention, and Robert Aumann gave it a mathematical formulation in his 1976 paper Agreeing to Disagree. The tower is load bearing, and the reason is coordination.

You will not act on what you know unless you are confident that others will act too. Consider a crowd that privately despises a regime. Everyone hates it. Everyone knows they hate it. Nobody moves, because moving alone is suicide, and each person needs to know not just that the others agree but that the others know that everyone agrees. Then someone gathers them in a public square, and a fact that was already in every single head becomes something else: it becomes common. Now the crowd moves. Nothing was learned. Everything changed.

That is the thesis of Michael Chwe's Rational Ritual (2001), which reads public ceremonies, from rallies to inaugurations to weddings, as machines for manufacturing common knowledge. They work not because they transmit meaning from a stage out to each person, but because they let each person see what everyone else is seeing.

So publicity is not primarily about informing people. It is about making knowledge common, which is what lets people count on each other.

Your decisions are mutual, not common

Here is where it earns its keep.

Most of what your team has decided is mutual knowledge. The five people in the room know you chose the boring database over the interesting one, and they know why. Ask any one of them and they will tell you correctly. And yet, reliably:

• Someone quietly does the other thing next month, and they are not being difficult. Nothing ever told them the choice was settled rather than one person's preference.
• The same question comes back next quarter and gets argued from scratch, with the same people, reaching the same conclusion, at full price.
• Someone joins and inherits none of it, because there was nothing to inherit.
• Nobody can point at anything, so the tie gets broken by whoever is most senior in the room that day.

None of that is a memory failure. Everyone remembers fine. It is a coordination failure. The decision never became common knowledge, so nobody could count on anybody else honoring it, so it never bound anyone to anything. It was an impression that several people happened to share.

The worst reason not to write it down

Which brings us to the trap, and it is a good one, because the objection is true.

"We all already know that."

Yes. You do. That is exactly why the record is necessary, and exactly why it feels like busywork. Writing the decision down gives no new information to anybody who was in the room. Its value was never informational. It is epistemic: it is the thing that converts mutual knowledge into common knowledge, so that everyone knows that everyone knows, and the choice can finally be relied on, defected from, inherited, or pointed at. The written decision is the child saying it out loud.

Be precise about what does the work, though, because writing alone does not do it. A decision in somebody's private notes, or in the fourth of nine documents nobody agrees is the real one, is not common knowledge. It is mutual knowledge with extra steps. What creates common knowledge is publicity in Chwe's sense: one place everyone knows about and knows that everyone else knows about. The record has to sit somewhere you can be sure the others are looking too. Otherwise you have written a note to yourself.

A decision that lives in five heads is not a decision. It is five people with the same impression, and an impression binds nobody, because no one of them can be sure the other four are still holding it.

We built Brief so that a team's decisions become common knowledge instead of a shared impression, for the people who have to honor them and, increasingly, for the agents that read them. But the distinction is older than any of that, and it is yours to use anywhere: in a contract, a launch, a negotiation, a marriage.

So the question worth asking about your last important call is not whether people know about it. Everyone probably does. Ask the sharper one instead. Does everyone know that everyone knows?

Frequently asked questions

What is the difference between mutual knowledge and common knowledge? Mutual knowledge means everyone knows a fact. Common knowledge means everyone knows it, everyone knows that everyone knows it, and so on without end. The distinction was analyzed by David Lewis in Convention (1969) and formalized mathematically by Robert Aumann in Agreeing to Disagree (1976). It matters because coordination depends on the higher levels: people act on a shared fact only when they can count on others acting on it too.

Why does writing down a decision everyone already knows change anything? Because the record adds no information and that was never its job. It converts mutual knowledge into common knowledge. Once the choice is stated where everyone can see that everyone can see it, it can be relied on, inherited by people who were not there, and pointed at when someone departs from it. Before that it is only a shared impression, and an impression cannot bind anyone.

Why do teams keep re-litigating decisions they already made? Usually not because anyone forgot. Everyone typically remembers the conclusion. What is missing is the common knowledge that it was settled, so each person is individually unsure whether the others still consider it binding, and the cheapest way to resolve that uncertainty is to have the argument again.