In the ancient Chinese court of Wei there was a law: whoever rides secretly in the ruler's coach shall have his feet cut off.1
Mi Tzu-hsia's mother fell ill. He took the royal coach to visit her, pretending the ruler had given permission.
When the ruler found out, he said:
"How dutiful is Mi Tzu-hsia! For his mother's sake he even forgot that he was committing a crime making him liable to lose his feet!"1
Another time they walked in an orchard. Mi Tzu-hsia began eating a peach, couldn't finish it, and handed the ruler the other half.
"You love me so much that you would even forget your own saliva taste and let me eat the rest of the peach!"1
Years later, courtiers spread word that Mi Tzu-hsia was devious and arrogant. His reputation was damaged, and the ruler "came to see his actions in a new light."
"This fellow once rode in my coach under pretense of my order, and another time he gave me a half-eaten peach."1
Same coach. Same peach. Same man. He was punished for both.
Nothing about the acts moved. Not the facts, not the intent, not even the ruler's knowledge of them — he had known about the coach from the beginning and had chosen to be charmed.
What moved was the interpretive frame, and the frame was supplied by third parties who had never been in the coach or the orchard.
Greene's line: "The fate of his feet depended solely on the strength of his reputation."1
That word solely is doing a lot and it is, in this case, earned. There is no other variable in the story.
The mechanism sits in Law 5's Keys to Power, and it is the most philosophically serious passage in the chapter:
The people around us, even our closest friends, will always to some extent remain mysterious and unfathomable. Their characters have secret recesses that they never reveal. The unknowableness of other people could prove disturbing if we thought about it long enough, since it would make it impossible for us really to judge other people. So we prefer to ignore this fact, and to judge people on their appearances.2
Take the argument seriously, because it's better than it looks.
An act does not carry its own motive. He took the coach is compatible with devotion, with entitlement, with panic, with contempt. The act underdetermines the person, always, and no amount of additional observation closes the gap — you can watch someone take a coach a hundred times and still not know why.
So a judgement about a person is never a reading of their acts. It is a reading of their acts through a prior, and the prior is doing most of the work.
Reputation is that prior. It is not a summary of what someone has done; it is the interpretive key applied to what they do next, and to what they have already done.
Which is why the peach could be re-read. The evidence was never the thing deciding.
The property that makes this case worth a page rather than a paragraph is that the reinterpretation ran backwards.
Most accounts of reputation treat it as forward-facing: build a good one and future acts are received generously. That's true and it's the smaller half.
Mi Tzu-hsia's punishment was for acts committed years earlier, already adjudicated, already praised, and — critically — already forgiven in the only sense that matters, which is that the ruler had declined to enforce a law he could have enforced.
The coach offence had been settled. It became unsettled because the interpretive key changed, and when the key changes, the entire archive is re-read.
So reputation is not a buffer around your future conduct. It is a live index on your whole history, and the history is not safe. Every act you have ever committed that was received generously is stored, unaltered, awaiting a second reading — and the second reading requires no new facts.
They follow directly from the reach running backwards.
Nothing is ever actually forgiven, only currently read favourably. The ruler did not forgive the coach; he interpreted it as devotion. Interpretations are reversible in a way that pardons are not.
The favour you received in a good period becomes evidence against you in a bad one. The half-eaten peach was offered as intimacy and retained as insolence. Mi Tzu-hsia's closeness to the ruler — his greatest asset — turned out to be the archive of his conviction.
The courtiers are the invisible mechanism and Greene passes over them in eleven words: "envious fellow courtiers, spreading word that Mi Tzu-hsia was actually devious and arrogant."1
They did not produce evidence. There was none to produce, and none was needed.
What they supplied was a competing prior — and the ruler adopted it, then re-derived the same events from it. The re-derivation felt to him like discovery. This fellow once rode in my coach under pretense of my order is the sound of a man who believes he is now seeing clearly, when in fact he has only changed which lens he is looking through.
That is the part that transfers. The dangerous move against you is not an accusation, which can be answered. It is the successful installation of a frame, after which your own record convicts you and no one has to lie about anything.
Something you did months ago, which went down well at the time, has started being mentioned again in a different tone.
That is the signal, and it is much earlier and much more reliable than any direct criticism.
Do not defend the act. The act was never in dispute and defending it concedes that its meaning is now open. I took the coach because my mother was ill is a sentence that only gets said by someone whose frame has already shifted.
Instead, ask the diagnostic question: has anything factual changed, or only the tone? If the facts are identical and the reception has moved, the movement happened somewhere else — in a conversation you weren't in — and the problem is not the act. It is that a new prior has been installed and is now working through your archive.
Which means the repair is also not about the act. Arguing the coach cannot fix the frame that is re-reading the coach.
The available moves are frame-level. Give people a more compelling competing prior, which usually means visible present conduct rather than argument about past conduct. Or find whoever installed it, because a frame maintained by two or three people can sometimes be addressed at the source and one broadcast to a whole court cannot.
And the preventive note, which is the harder one. Your record of favours received is not neutral storage. Every act of closeness — the shared confidence, the special access, the half-eaten peach — is being held in a form that a shift in frame converts to evidence. That isn't an argument for distance. It is an argument for noticing that intimacy with a powerful person is an accumulating position that can be revalued without warning by people who were never present.
Strongest support. As close to a controlled experiment as this book contains: two specific acts, two recorded reactions, one changed variable, opposite outcomes. The ruler's own words are given for both readings, which makes the reinterpretation observable rather than inferred.
Tension — the ruler's agency disappears. Greene attributes the reversal to reputation, as though it were a force acting on a passive observer. The ruler chose to be charmed and later chose to be outraged, and the courtiers' whispers were an input to a decision rather than a determinant of one. The passive framing makes the mechanism seem more inexorable than the case shows.
Tension — against Greene's own Preface. Judge by Effects, Not Intentions instructs the reader to score conduct on outcomes and disregard accounts of motive. The Wei ruler did the opposite in both directions — he read motive into the coach twice, oppositely, from identical effects. Greene presents the case as a lesson about reputation and never notices it is a demonstration that his own Preface rule is almost never how judgement actually works.
Tension — "solely" is overstated. Greene says the fate of his feet depended solely on reputation. Years had passed, the ruler's own position may have changed, and a favourite's usefulness declines. The case is presented with one variable because it is told as an exemplum.
🚩 SINGLE SOURCE · 🚩 SECONDARY WITHOUT PRIMARY — the story derives from Han Feizi and is a well-known Chinese exemplum, cited by Greene without source, translator or date. [POPULAR SOURCE].
Open questions.
Greene's Ingratitude and the Burden of the Favor — built at Law 2 in this cluster — holds that a favour received accumulates as resentment in the receiver, on its own schedule, whether or not anyone mentions it.
Mi Tzu-hsia is the same asymmetry read from the giver's side, and the pair completes a circuit neither closes alone.
Law 2: the person you helped is quietly re-reading your generosity as condescension. Law 5: the person you were close to is being re-read by you, once someone supplies the frame.
So an asymmetric relationship is unstable in both directions simultaneously. The subordinate's gratitude is fermenting; the superior's affection is one whispered campaign from inversion. Neither party is doing anything wrong and neither is safe.
What the pair yields: the intimacy is the exposure. The peach was only usable as evidence because it had been shared. The favour was only corrosive because it had been given. In both chapters the thing that made the relationship valuable is the identical thing that made it dangerous, and Greene never puts the two halves together despite building them forty pages apart.
→ behavioral-mechanics · Doubt as a Weapon
The technique page holds that an insinuation cannot be refuted — denial reads as defensiveness, silence as confirmation — and that this makes doubt the weapon of the party with nothing to lose.
The courtiers are that technique, and Mi Tzu-hsia's case shows what it is actually operating on.
The mechanism page frames doubt as attacking a reputation. This case shows it attacking an archive. The courtiers did not damage Mi Tzu-hsia's standing in the abstract; they installed a reading under which his existing, known, previously-praised conduct became prosecutable.
Which is why he had no defence. There was no accusation to meet. Every fact the ruler cited was true, publicly known, and had been celebrated by the ruler himself.
The insight neither page produces alone: doubt's real payload is not a claim about you but a decoder for your history, and that is why it is unanswerable in a way ordinary accusation isn't. You can refute a false charge. You cannot refute a re-reading of true events, because there is nothing in it that is false.
→ history · Coriolanus and the Forum
The Law 4 case: a legend built in a man's absence, out of scars and rumour, destroyed by three speeches — because the construction was the Romans' and he repossessed it.
Set beside Mi Tzu-hsia, the two cover the complete space of how a reputation can fail, and the contrast is instructive.
Coriolanus supplied new information that was incompatible with the existing frame. The Romans updated because he gave them something to update on.
Mi Tzu-hsia supplied nothing at all. No new act, no speech, no error. The frame changed around a static record.
Which yields the uncomfortable general claim: only one of these two failure modes is under your control. Coriolanus could have stayed silent. Mi Tzu-hsia had no available conduct that would have prevented his — the whispers required no cooperation from him and used only material he had already generated in his best years.
That reframes Law 5's central instruction. Guard your reputation with your life implies a defensible perimeter. Against the Coriolanus failure there is one, and it is discipline. Against the Mi Tzu-hsia failure there isn't, because the attack surface is your past and the past cannot be guarded.
Sharpest implication. The coach offence had been settled — the ruler knew, and had chosen to be charmed, and had declined to enforce a law he could have enforced. It became unsettled years later with no new fact in evidence. Nothing is ever forgiven; it is only currently being read favourably, and interpretations, unlike pardons, are reversible at any time by people who were never present.
Generative questions.