Character Is a Risk, and We Insured It Away

Thoughts. What I notice while building.

The line comes up in every second conversation, usually after the second glass: there used to be personalities. In politics, on boards, in trade associations, in newsrooms. Now the same seats hold people who sound like they came off one production line. Friendly, prepared, unassailable, interchangeable.

The observation is right. The explanation that comes with it is wrong.

The usual explanation is that people got worse. More cowardly, smoother, more careerist. It is comfortable because it has culprits, and it is wrong because it explains nothing. People do not change inside forty years. Filters do.

My claim is narrower and less comfortable: character has not died out. It gets removed much earlier. By nobody who meant any harm. By twelve people who all wanted something reasonable.

The comparison is broken before it starts

Before any explanation can work, the comparison has to be clean. It never is.

When we say "back then", we mean three names. Maybe five. We put today's median up against the three outliers of 1972 who made it into memory. Sitting next to those three were three hundred nobody remembers, because they were exactly what we complain about today. They dropped out of the comparison, not out of history.

So the comparison is broken at the root. That does not make the matter harmless. It makes it precise.

The correct claim is not that the average fell. The correct claim is:

The variance shrank. The middle did not move. The tails disappeared, and both of them did.

That is a completely different sentence, and it is testable. A narrower field does not only mean fewer giants. It also means fewer disasters. Less megalomania on boards, fewer regional premiers with a drinking problem and a gun licence, fewer people who drive a company into the ground out of pure self-regard.

We bought insurance. It works. That is the problem.

We paid with the upper tail, because you cannot cut one end of a distribution on its own. Nobody ever voted on that price, because it was never on an agenda. It has never appeared in a report either. Prevented disasters can be counted. Prevented greatness cannot.

Every no counts. No yes does.

Now the mechanism. It is unspectacular, which is exactly why it works so well.

Almost every modern path upward runs on a rule that is never said out loud: nobody can say yes on their own. Almost anybody can say no on their own.

The local branch can stop you. The regional list can stop you. The caucus, the committee, HR, the board, compliance, the press office, one editor, one comment section. None of them can make you by themselves. Every single one can end you by themselves.

A selection running on effective unanimity does not pick the best. It picks the least objectionable. That is not a moral statement, it is arithmetic.

And here is the step that turns an observation into a law. Character is, by definition, the thing somebody can take issue with. A trait nobody can take issue with is not a trait.

Read the right-hand column again. Every one of those objections is legitimate. Every one would read well in the minutes. None of them is malice.

So now you get the seat.

The arithmetic behind it is one line. If p is the probability that any given gate objects to any given trait, a candidate gets through n gates with probability (1 − p)^n. Twelve gates at a rejection rate of one in four gives you three per cent. One in eight gives you twenty. A rate of zero gives you one hundred, and exactly one kind of person carries that rate: the one there is nothing to object to.

You do not need a conspiracy to remove character from a process. You need twelve reasonable people with a veto and one candidate with properties.

The amnesty is deleted, and nobody abolished it

A second filter has been added, and this one is genuinely new.

A biography used to come with a de facto statute of limitations. Not because people were more generous, but because looking things up was expensive. Somebody who said something stupid in 1983 relied on nobody bothering to drive to an archive, find a microfilm and pull a copy. Almost nobody ever did. The past was not forgiven. It was just too expensive.

That cost fell to zero. We did not raise the penalties. We made the search cheap. The result looks identical and has a completely different cause.

Which changes what is rational. If every utterance stays retrievable, and every utterance can eventually be used against you, then the optimal strategy is not to speak cleverly. The optimal strategy is to produce no evidence at all.

The optimal biography is the one with nothing in it.

And the people starting out at twenty-three today know this. Not as cynicism, but as ambient knowledge, as obvious as a helmet on a building site. They are not more cowardly than their parents. They are better informed about the cost.

The court never left

Norbert Elias supplied the best term I know for this behaviour, and he found it in a place nobody looks: at court.

His point in one sentence: the more densely people depend on each other, the further ahead they have to control their own behaviour. Not out of manners, but because in a dense web every impulse has consequences you can no longer see the end of. Elias calls this self-restraint, and it is how he describes the courtier: a person whose main instrument is the damping of his own affect.

Versailles did not defeat the aristocracy. Versailles disarmed it and paid it in ceremony. Anybody who still flared up there lost. Anybody who held himself in rose. Within two generations a warrior caste had turned into a society of observers studying each other at dinner.

This is where I start, and the next claim is mine rather than his: the court Elias describes had a few thousand members. Our density of interdependence is orders of magnitude higher. Everybody depends on everybody, permanently, verifiably, with a record. The courtier is therefore not a historical type any more. He is the default.

Conformity is not a character flaw of modernity. It is the price of density. Anyone who argues about it in moral terms is arguing about something else.

The new Baroque

Which brings us to the second thing people keep saying: that nobody cares about anything that is not directly around them.

That is also true. And here too, the moral explanation is the worst one available.

I call the condition the new Baroque, and I mean exactly one thing by it. The Baroque was the era with the highest expenditure on representation and the lowest individual influence on anything decided above you. Architecture was painted onto ceilings where no architecture existed. Marble was made out of stucco. Churches were built to pull your eyes upward. And at the same time, whether there would be a war was, for practically everybody in that room, a question of weather: it arrived, you dealt with it, you did not act on it.

When nothing upward answers any more, attention moves inward. Not out of selfishness. Out of economics.

Because attention is not a question of morals but of feedback. It flows to where the loop closes, where what you do comes back to you and you can tell that it came back. Your vote is one in sixty million. Your insulation is one in one. Your child's route to school is one in one. The argument in your team is one in one.

And the drawing says the part I find more important than the lament: the attention did not disappear. It was rerouted. The effort stayed exactly the same, it just goes into furnishing the room you are sitting in. That is why a society supposedly indifferent to everything simultaneously pours an unbelievable amount of energy into self-presentation, interiors, food, fitness, feed and family logistics. This is not shallowness. It is ceremony.

The Baroque knew this about itself, by the way. It painted its own emptiness, hung it on the wall and called it vanitas. We write posts about authenticity instead.

And now the loop closes

The two halves of this text are not side by side. They are a circle, and the circle is the real reason nothing corrects itself.

At the top are people who got through twelve vetoes. By construction, those are the people nothing sticks to. Nobody can attach themselves to somebody nothing sticks to. So no attachment forms. Without attachment there is no public willing to defend anyone after the first scandal. Without that public, every property is an uninsured risk. And nobody carries an uninsured risk.

So the only people who survive up there are the ones who do not need one.

Character is not a property somebody has. It is a risk somebody carries. As long as nobody is willing to carry it, nobody can have it.

That is the whole diagnosis. It needs no decline, no decadence, and not a single claim that people used to be better. It needs two ingredients: many vetoes and little cover.

And now you

So far this has been about politics and boards, which is the comfortable version, because it is far away. The uncomfortable one: the same machine runs at forty employees, and you operate it yourself.

It runs every time somebody in a hiring round says there were "no red flags". Every time a promotion is decided by consensus. Every time a candidate has to clear four conversations and any one of them can end it. Every time a feedback form says somebody "creates friction", and nobody asks against what.

Three things that actually change the outcome, and all three hurt:

Count your vetoes. Not the formal ones. The real ones. How many people in your organisation can block a hire without giving a reason? Every one of them multiplies. Eight stages at a generous ten per cent rejection rate lets fewer than half through, and it is always the same third that falls out.

Give somebody the right to say yes. One person allowed to hire alone, against consensus, with their name on it. Not because that person judges better, but because a process in which nobody can say yes alone has mathematically only one outcome.

Stand next to them. Whoever brought in the person with edges has to be visibly standing next to them at the first sign of trouble. Otherwise the yes was worthless, and next time nobody brings in anybody with edges again. Cover is the actual currency here. Not courage.

None of this is a call to hire loudmouths. The loudmouth is the other tail, and the same insurance policy cut him off for good reasons. The question is whether your process can still tell "uncomfortable" apart from "unacceptable" at all. Usually it cannot, because the only quantity it measures is the absence of objection.

We did not pick the wrong people. We built a process that lets exactly one kind through, and then wondered why nobody notices him.

Related

• Why Every Place Looks the Same. The same mechanism applied to rooms rather than people: optimisation cannot see what it cannot measure.
• It Has Never Been This Cheap to Be Wrong. Why the short feedback loop decides everything, and what happens when you build one yourself.
• Spider-Man in the Cockpit: When Success Is Only Administered. What becomes of a company that only secures what already exists.
• Brave New World. The literary version of the same calculation: stability costs you the tails.

The two drawn scenes are not pictures but programs: procedural ink, live in the browser, without a single photograph. In the distribution, the axis, the ticks and the centre mark are drawn from the same numbers in all three states, so the narrowing cannot be a camera move. In the court, the twenty-eight strokes reaching outward at the start are the exact same twenty-eight that form the chandelier at the end. The tremble in the lines is reseeded nine times a second, not sixty, because that is how hand-drawn animation lives.

Norbert Elias, "The Civilizing Process" (1939) and "The Court Society" (1969). The terms interdependence, affect control and self-restraint are his. The transfer to modern selection processes, and the claim that our density of interdependence exceeds that of Versailles, is mine and not his. The twelve gates in the interactive section are an assumption, not a measurement: the arithmetic shows how fast multiplicative filters bite, not how many committees sit in any particular process.