Chao ruled Han from 358 to 333 BC. One afternoon he got drunk and fell asleep in the palace gardens.
The court crown-keeper — "whose sole task was to look after the ruler's head apparel" — walked past and saw his master asleep without a coat. It was getting cold. He put his own coat over him and left.1
When Chao woke and found the coat, he asked who had done it.
He then had the coat-keeper punished for neglecting his duties.
And he had the crown-keeper beheaded.1
The pairing is the whole scene, and it is worth setting out plainly.
The coat-keeper failed to do his job. His master was cold and uncovered and it was his responsibility. He is punished — proportionately, in the ordinary way, for negligence.
The crown-keeper did something kind, effective and outside his remit. He is executed.
So doing nothing when you should have acted is a smaller offence than acting when you should not have.
That is the scene's actual content, and it is a claim about how bureaucracies price error. Omission inside your boundary is a performance problem. Action outside your boundary is a category problem — and category problems are treated as threats rather than as failures.
"Do not overstep your bounds. Do what you are assigned to do, to the best of your abilities, and never do more. To think that by doing more you are doing better is a common blunder. It is never good to seem to be trying too hard—it is as if you were covering up some deficiency. Fulfilling a task that has not been asked of you just makes people suspicious. If you are a crown-keeper, be a crown-keeper. Save your excess energy for when you are not in the court."2
Two mechanisms are named and they are different.
Suspicion. An unrequested act raises the question of why — and the answers available to an observer are all worse than the true one. He wants to be noticed. He is positioning. He thinks the coat-keeper is failing and wants it seen.
And the effort signal, which connects directly to the chapter's diagnostic questions: "it is as if you were covering up some deficiency." Visible extra effort reads as compensation, and compensation implies something requiring it.
Greene's reading is about the courtier's exposure. The scene also contains a claim about the ruler that is worth separating.
Chao's court runs on strictly bounded roles. There is a person for the crown and a person for the coat, and the boundary between them is precise enough that crossing it is capital.
That looks absurd, and it is doing something. A ruler surrounded by people who each have exactly one duty can tell, instantly and without investigation, who is responsible for anything. The system is optimised for accountability legibility — and every act that blurs a boundary degrades it.
So the execution is not merely a tyrant's caprice about a coat. It is the enforcement of the property that makes the whole arrangement work, which is why the disproportion is the point rather than a flaw.
That does not make it defensible. It makes it intelligible, and the chapter offers neither.
The instruction — do what you are assigned and never more — is straightforwardly bad advice in most modern contexts, where initiative is rewarded and rigid role adherence is the mark of someone who will not be promoted.
But the underlying mechanism is real, and it is worth extracting from the terrible prescription.
Unrequested action is read as a claim. When you do something outside your remit, you have implicitly asserted one of three things: that it needed doing, that the person responsible was not doing it, or that you are the right person to do it.
All three are judgements about other people, and they arrive without the other people having agreed to be judged.
The crown-keeper's coat says, silently: the coat-keeper is failing at his job. He never said it. The act said it, and the coat-keeper was punished, which means the act was correct — and being correct is exactly what made it dangerous.
So the rule is not never do more. It is: an unrequested act is an assessment, and assessments have to be delivered like assessments — with the person concerned, and not by demonstration in front of the master.
You can see something that needs doing, it is not yours, and doing it would take five minutes.
Ask the question the scene turns on before you do it. Whose failure does this fix?
If the answer is nobody's — a genuine gap, unassigned — do it, and say so lightly afterwards. This is the case where initiative is cheap and safe.
If the answer is a person, you are about to make a public statement about their performance, and the fact that you are being helpful rather than critical does not change what it communicates. The crown-keeper's coat was kindness and it was also evidence.
There, do the slower thing. Tell them. Your thing is uncovered, do you want me to pick it up or will you? Ten seconds, and it converts an assessment into a courtesy.
And notice the two readings your extra effort will get, because you do not control which one lands. Generosity, or positioning. In an environment where people are competing for a superior's attention, unrequested visible work is read as positioning by default — and the more visible the beneficiary, the stronger that reading. The crown-keeper covered the ruler, in the garden, where it would be found. That is the worst possible venue for a favour, whatever his motive.
The strongest material is the paired punishment — negligence penalised, initiative executed — which is a precise and unusual data point about how bounded systems price the two kinds of error.
Tension: Greene draws a career rule from a capital punishment and does not distinguish environments where role-crossing is fatal from those where it is expected.
Tension: the scene's real mechanism is unstated. Suspicion is named; that an unrequested act is an implicit assessment of whoever's job it was is what actually happened, and the coat-keeper's punishment proves it.
Tension: the ruler's side is not examined at all. A system of single-duty roles is doing something — accountability legibility — and the chapter treats the execution as pure caprice while drawing a lesson from it as though it were rational.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — no text, chronicle or translator for a fourth-century BC episode; the regnal dates are given and nothing else is checkable. The story is a well-known Legalist parable and its didactic origin is not mentioned. [POPULAR SOURCE]
Open questions. Was this a real event or a parable — the account has the shape of a teaching story, and reading it as history rather than as doctrine changes what it evidences. And: what happened to the role? A system that executes a man for covering his king has a recruitment problem for the next crown-keeper, and the scene does not follow through.
Against Fra Filippo Lippi two scenes later, Law 24 contains a direct contradiction it never acknowledges. Lippi, a slave, does something entirely outside his assigned function — draws a portrait on a wall in chains — and it buys his freedom. The crown-keeper does something outside his function and is beheaded. Greene draws opposite lessons from the same structural act and supplies no rule for telling them apart.
The available distinction is that Lippi's act displayed a capability nobody else had, while the crown-keeper's performed a duty that belonged to somebody present. One creates value that cannot be attributed to anyone else's failure; the other cannot avoid doing so.
Set beside Never Be the Bearer of Bad News the two rules share a structure. The bearer is contaminated by information he did not create; the crown-keeper is destroyed by an act he intended kindly. In both, the courtier's fate is decided by what his presence signifies rather than by what he did or meant.
And against Practice Nonchalance, this scene supplies the rule's dark version. That rule says hide your effort because visible strain looks like a workaholic rather than a genius. Here visible extra effort is read as "covering up some deficiency" — a much harsher inference, and the one that gets people removed rather than merely underrated.
Behavioral mechanics — Master-Servant Influence and Behavioral Mirroring
That page's operating principle is that behaviour cascades downward — a master's conduct propagates through those who serve him, so the retinue reproduces the centre.
Chao's court is that principle enforced with an axe, and the pairing shows what the enforcement is for.
What the pair produces: rigid role boundaries are how a cascading system stays legible, and the disproportionate punishment is aimed at the boundary rather than at the man. If conduct propagates downward, then a court where roles blur is a court where the propagation cannot be traced — nobody knows whose behaviour originated what.
Executing the crown-keeper is not about the coat. It is a demonstration that the boundaries are real, delivered to everyone watching, at the cost of one man who happened to cross one kindly.
Which explains why the kindness made it worse rather than better. A malicious boundary-crossing could be punished as malice, leaving the boundary itself unexamined. A benevolent one raises the question of whether the boundaries should exist at all — and that is the question the system cannot allow, so the response has to be severe enough to close it.
Psychology — Moral Agency & Categorical Responsibility
That page works the explanatory abyss under blame — how categorical assignments of responsibility paper over causal complexity, and what that tidiness costs.
Chao's court is categorical responsibility built into an institution: one man, one duty, complete assignment.
What the pair produces: categorical responsibility buys instant attribution and pays for it by making the world unsalvageable at the margins. The system works beautifully for its purpose — Chao knows exactly who to punish, immediately, without inquiry.
The cost is that nothing outside a category can be attended to by anyone. A cold sleeping king in a garden is not in anybody's remit but the coat-keeper's, and the coat-keeper is absent, so the correct behaviour of the entire court is to let him lie there.
And that is the general failure of tidy responsibility rather than a quirk of Han. Every gap between categories becomes nobody's, and the more precisely the categories are drawn, the more gaps there are. The crown-keeper's crime was noticing one — which is the behaviour such systems most need and are structurally obliged to punish.
Sharpest implication. The coat-keeper was punished for negligence and the crown-keeper was executed for initiative — so failing inside your boundary cost less than succeeding outside it.
And the kindness is what made it dangerous, because a benevolent boundary-crossing raises the question of whether the boundary should exist, and that question could not be permitted an answer.
Generative questions.