Social Common Sense Is Maintained by Everyone’s Conjecture Though Nobody Endorses It

Convergence has a large payoff. Forms of greeting, payment dates, document formats. Negotiate each of them from scratch every time and the transaction never happens at all. Because things converge, the load of judgement drops. Nothing here treats convergence as waste. In most situations, going along with it is the cheaper option.

One thing is missing from it, and only one. The mechanism that produces convergence contains, in itself, no step that checks whether the content is right. Where not fitting produces a visible disadvantage, the passage of time stands in for a check. That condition does not hold in domains where being converged is itself what gets evaluated. Things do not converge because they are correct; they keep converging because they have converged. The converged state therefore survives after its grounds are gone.

Common sense is not something someone decided and then distributed. It is the state in which choices have lined up without any order being given, because each person consulted what others were doing. Those consulting assume the others are doing it with conviction. But what can be observed is only the outcome of a choice, never whether the person is convinced by it. A state can hold in which everyone is making the same conjecture and nobody is convinced.

The object gets narrowed first. Three things get settled here. By what mechanism convergence arises, and how it is promoted into grounds. How the burden of doubting it is distributed, and to whom. And what to look at in order to tell that a position of your own has formed outside the convergence.

Two things are withheld. Which convention to start with, and whether a given convention you currently follow is right. Hand over the first and the person bearing the cost and the person deciding come apart. The second cannot be decided by anyone who does not know the conditions under which that convention arose. Either one, handed from outside, would be followed without the receiver being able to check the grounds — which is exactly the shape this article is about.

The damage comes from reading convergence as evidence of correctness. A surviving convention either survived because it was right or because doubting it was expensive. From outside, those two look identical. Follow without separating them and you keep fitting your time and your pricing to a form whose conditions have already disappeared.

And that misreading concentrates where it costs most on the business side. Going rates, formats, standard procedures. All of them arrive without leaving any memory of having been chosen. What leaves no memory of being chosen does not present itself as something that could be chosen again.

What Can Be Set as a “General Knowledge Quiz” and What Cannot

One word points at two different things.

The first is the sense of a list of manners to be learned. How to hand over a business card, the order of seating, the steps of a report. Common sense in this sense can be set as a quiz question. There is a correct answer and the marking can be done. Collections of general knowledge questions exist on exactly this basis, and the books on shop shelves about social common sense are mainly about the first sense. It is a matter of knowing or not knowing, so not knowing is repaired by learning.

The second is the sense of a judgement used as a premise without being questioned. “In this trade that is how it is done.” “That is roughly what the price should be.” This sense cannot be set as a quiz. Nobody anywhere holds the answer. It names a state in which the thing is known and nobody has been asked why.

Because both senses share a word, questioning the second is received as denying the first. Ask why a rate or a format is what it is and you can be treated as someone who has not learned the manners. The overlap of the word makes the question hard to raise. And the overlap does not dissolve in a dictionary. The line between the two senses lies not in the meanings but in which sense was in use at that moment.

The overlap has a further effect. The first sense carries a moral colour. A person with common sense, a person without it. Those are assessments of a person’s character. Once the colour is there, it transfers to the second sense as well. The object of the question is a judgement, and the reply comes back at the level of the person.

An objection: surely the second kind also survives because its usefulness has been confirmed over long stretches of time.

For some of it, that holds. Manners that have survived a long selection often have reasons. But selection needs a condition: that not fitting produces a visible disadvantage. Where no disadvantage returns, selection does not operate. And in domains where being converged is itself what gets evaluated, the disadvantage of not fitting arises purely from not fitting. Since it arises independently of whether the content is sound, no route exists by which the content could be tested.

So the second sense will not be called “common sense” here. It will be called a converged state. The rephrasing makes the question easy to raise. “Why has that price converged?” carries no moral implication. Changing the word opens up an object that can be handled.

A Norm Nobody Endorses Is Held Up by Everyone’s Conjecture

Convergence does not come from an order. It comes from each person consulting the others.

When making a choice, people use what others are doing as one input. That is rational behaviour. In a domain where you hold no information, treating other people’s choices as containing information is sound.

Yet while other people’s choices can be observed, why they chose that way cannot be. Only the outcome is visible. So each person conjectures that the others are doing it with conviction, and chooses to match that conjecture.

The phenomenon has a name: pluralistic ignorance. The psychologists Daniel Katz and Floyd Allport put the concept forward in their 1931 book Students’ Attitudes, from a survey that asked people both for their own opinion and for what they believed those around them thought. On many contested matters, students’ personal opinions sat on the more moderate side of what they attributed to their peers.

In 1993 the psychologists Deborah Prentice and Dale Miller ran the same design on drinking on a university campus (Prentice & Miller, 1993, Journal of Personality and Social Psychology, 64(2), 243–256). Students estimated their own discomfort with the level of campus drinking as greater than the discomfort they attributed to the average student. Everyone was privately thinking that only they had a problem with it.

That study concerns one campus and one practice; it did not establish that every convergence is made of pluralistic ignorance. Convergences that everyone genuinely endorses exist as well. The test is whether each person can state the reason for the choice. If they can, the convergence is an accumulation of conviction. If they cannot and it is converged anyway, an accumulation of conjecture is likely.

In the pluralistic-ignorance state, each person feels like the exception. Someone who thinks “social common sense is wrong here” does not say so. Because it is not said, nobody learns that others feel the same. Silence reinforces the conjecture.

That state is experienced as isolation. The sense of being the only one who does not fit. Even where the feeling is in fact the majority feeling, being the majority cannot be observed. While it cannot be observed, it remains, as a feeling, the minority position.

Convergence Breaks Not When the Content Changes but When a Minority Becomes Visible

The way convergence breaks follows from the same mechanism, reversed. It does not break because the content has been corrected. It breaks once the number of people saying so out loud passes a level.

The moment one person says it, everyone else gains a piece of information: not only me. Once that information exists, the next person finds it easier to speak. A convergence that held for a long time and then broke quickly did so by that accumulation.

The process has been measured. Damon Centola and colleagues ran an experiment in 2018 in which participants were paired repeatedly and had to converge on a name for an object, then introduced a committed minority insisting on a different name (Centola, Becker, Brackbill & Baronchelli, 2018, Science, 360(6393), 1116–1119). While the minority stayed below a certain share, the convention did not shift at all. Once the share crossed the threshold, the group’s convention flipped to the new one within a short span. In that experiment the threshold sat at roughly a quarter.

There is a reason not to carry the number across. The experiment used an artificial task with small groups, and the threshold is reported to move with conditions such as how much of the past the group retains. No claim can be made that the same proportion applies to a trade’s going rates or dealing conventions.

The shape, though, is usable. Change arrives as a step, not as a slope. So the fact that nothing has happened yet is not evidence that nothing is working. A situation where only one person is visibly doing it differently and a situation just short of the threshold look identical from outside.

The same shape runs in reverse. Let the minority become invisible and the convergence returns. Someone who keeps making the outside choice while saying nothing about it sends no information to anyone. A choice that does not travel has no force on the convergence.

There is a second point about the step. Not moving until the threshold also means that the speed after the threshold is not yours to control. When a going rate moves, it moves because the outside choices became visible, not because you decided to raise your price. So the mechanism cannot be built into your own plan. What can be built in is only whether you keep your own outside choice in a visible form.

In that sense, quietly continuing an outside choice and stating the reason for it do different work, even though the choice is the same. The first adds only to your own burden; the second enters other people’s judgement as information. Since the burden is the same either way, the visible version returns more for the same cost.

A Converged State Gets Reused as Grounds

Once convergence holds, a second stage begins. The fact of being converged starts being used as grounds.

“Everyone does it that way” is not an explanation of why it is done. In practice it functions perfectly well as one. Asked for a reason, the person answers “because that is normal”, and the asker accepts it. With that exchange the question closes. A closed question is not reopened, and each closure strengthens the convergence.

What bears on this is the “self-fulfilling prophecy” described by the sociologist Robert K. Merton in 1948 (Merton, 1948, The Antioch Review, 8(2), 193–210). A belief that was false to begin with becomes true through the conduct of those who acted on it. The prophecy does not predict reality; it produces it.

In the case of common sense, what gets produced is the fact that everyone really does do it that way. A convergence that began as conjecture becomes, over time, an observable state of affairs. Once it is observable, it can be treated as a judgement based on observation rather than conjecture. The grounds are supplied retrospectively.

An objection: if many people really do act that way, is that not sufficient grounds already?

As a description of a distribution, it is sufficient. “The prices in this trade sit in this band” is accurate as a statement of fact. It becomes insufficient at the move to “therefore it should be done that way.” A distribution reports what is happening; it does not report whether that is good. Between the two lies a gap in the argument, and the gap is filled by the weight of the convergence itself.

The structure comes with the erasure of origins. The original reason for the form is lost over time. Since the convergence outlasts the reason, anyone who sees only what remains takes it for the natural shape of things.

That erasure has a practical consequence. The line between what can be changed and what cannot stops being visible. A form that was decided can be decided again; a form that simply came about does not present itself as available for redeciding. When the origin is gone, the first looks like the second.

Digging up the origin does not guarantee that a form can be changed. Often the excavation shows that the form is maintained by reasons still valid today — and that outcome is not unusual. The value of digging is not that the form turns out to be changeable, but that it becomes possible to judge whether it is. While no judgement is possible, neither changing nor deliberately keeping is available as a choice.

Breaking with Common Sense Brings No Punishment; It Brings a Demand for Explanation

Common sense is not a rule. Breaking it carries no penalty. And yet breaking it is heavy. What is actually happening?

What happens is a demand for explanation. A converged choice needs no explanation; a non-converged one does. Only the side that departed is asked why it does what it does.

The demand is not a punishment. It is nonetheless a real cost. It has to be met each time, meeting it does not guarantee agreement, and where agreement does not follow, the relationship takes on friction. And the number of times it is demanded is not up to you. It repeats with every new counterpart, once per departing item.

The second thing that happens is a change of classification. Someone making a non-converged choice is filed as “that sort of person”. The filing carries no good or bad in itself, but once it happens, subsequent exchanges proceed on that premise.

Being filed is not necessarily a disadvantage. In some domains, being outside is the reason for being chosen. Whether you are outside is a property of the choice; whether that hurts is a property of the setting. The same choice is treated as a disqualification in one setting and as a reason for selection in another. Which setting you are in can sometimes be chosen.

The demand for explanation consumes the time of the person who departed, continuously. And that consumed time is never counted. No line item exists anywhere for “time spent explaining”. Costs that are not counted also drop out of the material for judgement. The benefits of going along are visible; the costs of departing are not. That asymmetry is part of the force holding convergence in place.

The demand has one more property. Whether an explanation lands is decided by the other party’s spare capacity, not by its content. The same explanation passes with someone who has time and fails with someone in a hurry. When it fails, what comes back is not a counter-argument but simply non-adoption. A counter-argument is material; non-adoption is not. Nothing indicates what failed to reach them.

Because of this, the outside choice carries a further disadvantage: it is hard to get any handle on improvement. A converged choice, when it fails, comes back with a reason, because the shared premise makes it possible to name where the divergence was. The outside choice has no such shared premise. So the burden of departing falls less on the labour of explaining than on the slowness of learning. That part can only be filled with time.

The Cost of Doubting Is Paid Only by the One Who Doubts

Doubting a convention has a specific property. The cost concentrates on the individual who doubts; the benefit disperses.

Suppose you notice that a format is inefficient and propose a different one. The proposal takes preparation, the persuasion takes time, and failure spends credibility. All of that is borne by the proposer. If the proposal carries and the format changes, the benefit reaches everyone who uses the format. The person who bore the cost does not receive most of the benefit.

Under that structure, the rational move for each individual is not to doubt — not because doubting has low expected value, but because the burden of doubting lands on you alone. And once everyone runs the same calculation, nobody doubts. The state persists even where the content has plainly stopped fitting.

In fact a great many people have doubted a convention and succeeded. If the burden concentrates, should there not be far fewer of them?

They do not disappear. Anyone whose expected return exceeds the burden still moves. But the rate at which they appear is not proportional to how badly the content fits. A convention that has stopped fitting survives a long time if doubting it is expensive. A convention that fits reasonably well gets updated quickly if doubting it is cheap. What weighs most on the lifespan of a convention is therefore less the soundness of its content than the cost structure of doubting it.

That structure can work in favour of someone arriving from outside. A person who does not share the setting’s conventions does not bear part of the cost — the relational friction, the change of classification — because they were already outside that classification. New formats often arrive from outside not because of a difference in ability but because of a difference in burden.

Arriving from outside carries a symmetrical disadvantage: no material for telling apart a convention with a reason from one whose reason is gone. People on the inside know by feel which parts are load-bearing and which are inertia. The outsider does not, so the chance of standing on something that should not be stood on goes up. New formats arriving from outside, and outside attempts dying quickly, come from the same source.

So the advantage does not belong to “the person from outside” but to “the person who spent time inside and then stepped out of the classification”. The material is held and the burden has dropped. That combination can be built deliberately. Change the range of relationships, change how you present yourself, change who you deal with. The burden side is a variable that moves with the design of your position.

Where You Feel You “Do Not Understand Social Common Sense”, the Business Conventions Are Gathered

On the business side, common sense shows up in three places: the going rate, the format, and the procedure. None of them registers as a choice while you are receiving it passively.

The first is the going rate. The price of a certain kind of service sits in this band. Nobody set it; participants maintain it by consulting each other. Each person looks at the rate and sets a price, and that decision then makes the rate. Since the rate can simply be looked up, no memory of an active choice remains.

The second is the format. Proposals are built this way, contracts cover this range, delivery happens in these units. The more converged a format is, the less friction it produces, and less friction produces more convergence. But a format specifies what is handed over and what is not. Following it means letting the format decide the range of what you deliver. And formats arrive in the counterpart’s hands.

The third is the procedure taken as given. In this trade the order runs this way, this step always goes in. Reasons are seldom asked for. Procedures continue as they were taught.

The feeling of “not understanding social common sense” tends to arise in front of these three. And the feeling divides into two quite different positions. One is not knowing the first sense — the manners that can simply be learned. The other is being asked to comply with the second sense while no reason is given. Learning resolves the first. It does not resolve the second, because nobody is available who can explain.

An objection: formats and procedures must have practical reasons, and ignoring them will cause accidents.

Where a reason exists, the concern is correct. And whether a reason exists can be found out by asking. A procedure someone can explain is a procedure with a reason. A procedure where everyone answers only “because it has always been done that way” has either lost its reason or never had one.

Asking produces friction in itself, because the person asked is being shown that they cannot answer. So checking reasons is frequently processed as rudeness. That too is part of the cost of doubting.

Being passive is not always a disadvantage. At the start of a business, adopting the existing forms is a fast route with few failures. What becomes a problem is that state continuing for years. During those years, the conditions those forms assumed have moved. And your side has no occasion to notice. Occasions do not arrive from outside; if one arrives, it arrives as a fall in business.

Doubting Is Not Denying but Checking the Conditions of Establishment

“Doubting common sense” suggests refusal or revolt. What actually gets done is not that. It is identifying the conditions under which the convention held, and checking whether those conditions still hold.

There are three steps.

The first: infer the conditions in place when the form arose. Why did it take this shape? What was the constraint at the time? In most cases the reason is a technical constraint, a constraint on how information travelled, or a property of the counterparties.

The second: check whether those conditions still hold. If the constraint is gone, what remains is the form alone.

The third: where the condition is gone, look for another reason to keep the form. The original reason can disappear while a different one grows in. Then the form should stay. Only the reason has been swapped.

None of the three steps contains a denial. They are checks. And the result of checking often comes out as “there is still a reason”.

The check requires material about the conditions of the time. Without material, inference becomes speculation. Change a form on speculation and you stand on the reason you could not see. The safe handling in the absence of material is to hold, without changing. Holding is not the same as complying. Complying skips the judgement; holding is the judgement that no material is available yet. A held item, once recorded, can be resumed when material appears.

The work is slow and mostly produces nothing. Check many and perhaps a few turn out to need changing. As a return on effort it looks poor. But if one of those concerns pricing or the range of what you deliver, the structure changes on that one. The expected value sits on the size of the hit, not the frequency.

The three steps have a second payoff. A convention that checks out as “there is a reason” is thereafter followed for a reason that sits on your side. Continuing the same format because you were taught it and continuing it because you checked are different states. The first cannot notice when the premise moves. The second remembers the condition it checked, so movement in the condition registers.

The main product of the work, then, is not the number of items changed but the list of conditions. A record of what holds under which conditions stays in your hands. With the record, the next time a condition moves, the check does not have to be redone. The first pass is many checks for few changes; the second pass only covers the items tied to the condition that moved. Return on effort cannot be judged from the first pass.

The selection itself takes judgement. Which parts are load-bearing is not visible from the convention’s side. The criterion for deciding that has to sit on your own side. Start doubting without one and whatever you adopt outside will also be borrowed from outside. How that criterion gets built is handled in Being unable to decide is not character but the consequence of a structure in which judgement has been handed over.

Whether Stepping Outside Worked Is Read from Where the Load Moved

What this is judged by is not whether things went well.

First, the frequency of explanation. Inside the convention, no explanation is required. Outside it, explanation is required. If the occasions for explaining are increasing, you are outside. If they are not, you are still inside.

One point is easy to misread here. Read a rising load as evidence that the direction is right, and a state that is merely inefficient reads the same way.

So second, whether there is somewhere the load has fallen. If departing increases explanation, something on the incoming side should have decreased in exchange. Premises no longer need confirming, price conversations get shorter, reasons for declining no longer need stating. If nothing has become lighter anywhere, departing is not doing any work.

What is being looked at, then, is not the total load but whether the load moved. If it moved, a position of your own has formed outside the convention. If it only grew, no position has formed yet.

Third, whether the audience for the explanation has changed. Right after departing, explanation increases toward the existing counterparts. As a position forms, the people being explained to change. Repeating the same explanation to the same people is not departing; it is generating friction on the inside.

On which convention to start with, nothing can be said from outside. The cost of doubting is paid only by the one who doubts, which is the very structure set out here. Let the side that bears nothing decide the order and the person bearing and the person deciding come apart — reproducing the distribution that has been describing.

Convergence has a payoff, and doubting everything in reach will exhaust you before it does anything else. Breaking convergence is not what is being recommended. What can be said is one thing only: that convergence has not passed through a check of its content. The remaining work is therefore not to doubt all of it, but to locate the places where something unchecked is carrying a heavy judgement.

That splits the conventions in your hands in two. One part concerns pricing, the range of what you hand over, and who you deliver to. Here the cost of doubting has a return. The other part is everything else. There, complying is cheaper and doubting leaves only depletion. The basis for the split is whether the convention bears on a judgement that is load-bearing for the business — and that basis can only sit on your own side.

On the individual side, common sense shows up as sensitivity to other people’s assessment; that is the subject of The mechanism of minding what others think. Putting into words the judgement you have placed outside the convention is worked through in How to build a worldview.

A going rate is the convergence described here, appearing in the form of a price. The analysis from the price side is in Stepping down from price competition.