Han Fei's fable is four sentences long: "There was in Sung a man, who tilled a field in which there stood the trunk of a tree. Once a hare, while running fast, rushed against the trunk, broke its neck, and died. Thereupon the man cast his plough aside and watched that tree, hoping that he would get another hare. Yet he never caught another hare and was himself ridiculed by the people of Sung."1
The joke depends on the man being obviously foolish, and it is worth resisting that for a moment, because the reasoning he performed is the reasoning that produces every useful rule anyone has.
He observed an event. He formed a hypothesis with a mechanism in it — hares run fast, this trunk is in their path, therefore hares die here. He acted on it. That is not superstition; it is the ordinary procedure by which experience becomes practice.
What is wrong is the sample size. And the sample size problem has a specific property that makes it invisible from inside: a single confirming instance and a reliable pattern look identical at the moment of observation. He did not have less evidence than he thought. He had exactly what he thought, and no way to know what it was worth.
The line that carries the fable is "he cast his plough aside."
He did not watch the tree in his spare time. He abandoned a working practice — reliable, low-yield, requiring labour — for a speculative one, and the abandonment is the whole loss. Watching a stump costs nothing if you are also ploughing.
So the failure is not credulity about hares. It is a portfolio decision made on one data point, and it went in the direction that always tempts: out of the reliable thing and into the exceptional one.
Why that direction? Because the hare was free. It arrived with no work attached. And once a return has arrived with no work attached, ordinary work is not merely less attractive — it is reclassified. The plough stops looking like the source of the harvest and starts looking like an inefficiency, a thing you do because you have not yet found the tree.
A windfall does not only mislead you about the future. It retroactively devalues the method you already had. That is the sharper claim, and it is the same mechanism as Law 47's account of good fortune, stated more precisely by a marginal from the third century B.C. See Good Luck Is More Dangerous Than Bad Luck.
The fable is an analogy and Han Fei says so immediately: "Now supposing somebody wanted to govern the people of the present age with the policies of the early kings, he would be doing exactly the same thing as that man who watched the tree."1
The tree stump is the sage-kings. The hare is the fact that their policies once worked. And the man's error is not believing the hare was real — it was — but believing that a result obtained under one set of conditions will recur because it once did.
The passage opens with the positive doctrine: "The sage neither seeks to follow the ways of the ancients nor establishes any fixed standard for all times but examines the things of his age and then prepares to deal with them."1
Note the symmetry. Two rejections, in both directions of time. Not the ancients — no authority from the past. And not "any fixed standard for all times" — no authority projected into the future either. What remains is assessment of present conditions and a response built for them.
That is Law 48's thesis, arrived at twenty-two centuries earlier and stated with more precision than the chapter that quotes it.
🚩 And here is the problem with Greene's use of the passage, which he does not raise.
Han Fei is the culminating theorist of Chinese Legalism, and the Legalist programme is fa: codified law, published so that everyone can read it, applied uniformly regardless of rank, deliberately impersonal, and designed specifically so that a ruler cannot govern by whim, by relationship, or by improvised judgment. It is among the most rule-bound political philosophies of antiquity.
The man who rejects "any fixed standard for all times" is the architect of a system whose entire virtue is fixity.
The resolution is available and it is the useful part. Han Fei rejects standards drawn from another age, not standards as such. The fixity he requires is a system built for present conditions, published, and revised when conditions change. What he is attacking is inherited authority — the idea that a rule is binding because the early kings used it.
Which is a considerably more workable doctrine than Greene's, and the difference matters: no law is fixed2 and no law from elsewhere is authoritative here are not the same instruction. The first leaves you with nothing to operate. The second leaves you with a system you built and are willing to rebuild.
"And was himself ridiculed by the people of Sung."1
Not ruined, not attacked. Laughed at.
⚠ That is the same sanction Law 48's Keys identify for rigidity twelve pages later: the powerful who are locked in the past "look comical—they are overripe fruit", and "predictability makes us appear comical."3 Han Fei and Greene agree on the enforcement mechanism, and the agreement is worth marking because the mechanism is doing more work than either author claims for it.
Ridicule is what a community applies to someone who is operating on a model everyone else has already discarded. It requires no organisation, no cost, and no adversary. And it is informative — being laughed at by people who are not your enemies is one of the few reliable external signals that your model has fallen out of contact with the situation.
The man of Sung had that signal available. The fable does not say he acted on it.
One thing worth recording about how the fable persuades. In Warring States literature the people of Song are the stock fools — Mencius has a man of Song who pulls up his seedlings to help them grow, and there are others. Naming the man's origin does a share of the argumentative work before the anecdote begins.
That is not a defect in Han Fei's reasoning; the argument stands without it. But it is worth noticing that the passage's rhetorical force partly rides on an existing prejudice, and that a reader in 250 B.C. knew the man was a fool from the second word.
Which is a small instance of the thing the passage is about. The audience's prior classification arrives before the evidence and determines what the evidence is taken to show — which is what happened to Artabanus, to Tomyris, and to the Communist units in Manchuria, and it is Law 48's most reliable mechanism, appearing here in the machinery of a fable rather than in its content.
The marginal sits in a chapter that is the forty-eighth of forty-eight fixed standards, in a book whose Preface calls them "timeless and definitive."
Han Fei's sentence — "neither seeks to follow the ways of the ancients nor establishes any fixed standard for all times" — is the sharpest statement of that problem anywhere in The 48 Laws of Power, and it arrives in the margin, unattached to any comment, of the chapter whose closing paragraph performs the same self-cancellation at length.
Greene evidently knows. The placement is not accidental — the Han Fei passage is on the same pages as the go material and four pages from the instruction to throw out the books. See Throw Out the Laws.
What the marginal supplies that the closing paragraph does not is a criterion. Greene says discard the laws others preach. Han Fei says discard rules whose authority comes from having worked in different conditions — which tells you which rules to discard and, by implication, which to keep and how to know when to revisit them. One is a gesture; the other is a test.
Something worked once, and it worked without the usual effort.
Before you conclude anything, count. Not the size of the result — the number of times it has happened. One is a hare. The confidence you feel is not tracking the sample; a single vivid confirming instance produces roughly the same certainty as a long reliable pattern, and you cannot tell them apart by how sure you feel.
Then look at what you are about to stop doing. That is where the cost is. Nobody is harmed by watching a stump; they are harmed by putting the plough down, and the plough is whatever reliable low-yield thing was actually feeding you. If the new idea can be run alongside it, run it alongside. If it requires abandonment, it needs far more than one instance.
Watch for the reclassification. The tell is not enthusiasm about the new thing — it is a change in how the old thing feels. When ordinary work starts to feel like a failure of imagination, that judgment did not come from evidence about the work. It came from having seen a return arrive for free.
And for rules you have inherited rather than found: apply Han Fei's test rather than Greene's. Do not ask whether the rule is fixed — ask what conditions it was built for, and whether those conditions are the ones you are in. A rule that fails that test should go regardless of its pedigree. A rule that passes it should stay, and should be re-examined when the conditions move, which is a schedule rather than a mood.
Evidence quality: attributed to Han-Fei-Tzu, Chinese philosopher, third century B.C. — correct by author and period, and the passage is a genuine and well-known one from the Han Feizi. 🚩 SECONDARY WITHOUT PRIMARY: no chapter or translator named. [POPULAR SOURCE] for the framing.
Tension: Greene quotes a Legalist against fixed standards. Legalism's central programme is codified, published, uniformly applied law. The tension resolves — Han Fei rejects inherited standards rather than standards as such — and the resolution produces a materially different doctrine from the one the chapter is arguing for.
Open questions: Han Fei's criterion requires knowing what conditions a rule was built for, which for most inherited rules is not recorded. What do you do with a rule whose original conditions are unrecoverable? The fable assumes the early kings' circumstances are known to be different; often the only available evidence is that the rule is old.
This is the second consecutive law in which a marginal supplies a more rigorous version of the chapter's own argument than the body text manages — Law 47 had the Kenkō tree-climber, Law 48 has this. [n=2], and both are pre-modern practitioner or philosophical texts against Greene's synthesis.
Against the corpus, the passage collides with the book's structural claim. Forty-eight laws presented as timeless, with a third-century-B.C. warning against timeless standards printed in the margin of the last one. Greene's closing paragraph makes the same move explicitly, which suggests the placement is deliberate rather than careless — but the marginal is the more useful statement because it supplies a criterion for discarding rather than an injunction to discard.
The convergence is with Greene's better recurring position on precedent: Mastery and The 33 Strategies of War are both consistently sceptical of doctrine inherited from conditions that no longer obtain, and both recommend building from present observation. Law 48's rhetorical version — no law is fixed — is the weakest formulation of a position he holds more carefully elsewhere.
The Availability Heuristic and Fear — the cognitive material explains why one hare outweighs a thousand hareless days, which the fable states as foolishness and does not mechanise. Availability holds that judged probability tracks the ease of recalling an instance rather than the frequency of the class, and a vivid, singular, personally witnessed event is maximally available while the absence of events is not represented at all. The man of Sung has one retrievable image and no counter-image, because nothing happening generates no memory to weigh against it. The insight neither frame produces alone: the asymmetry is not between one instance and many, it is between an instance and a non-instance, which means the correction cannot come from thinking harder about the evidence — the evidence against is structurally absent from recall. It also predicts where this failure clusters. Rare, vivid, high-payoff events are exactly the ones whose base rates are unrecoverable from experience, which is why the trap is specific to windfalls and why the fable is about a free hare rather than about a good harvest. And it yields the only workable remedy: count occasions rather than recall instances, because the count includes the days nothing happened and memory never will.
MacArthur Burns the Precedents — the historical case is Han Fei's doctrine executed rather than described, and the pairing exposes what the fable leaves out. Discarding inherited procedure is not merely a matter of deciding the old rules do not apply; it requires the discarder to possess enough standing, enough information about present conditions, and enough authority to impose the replacement — and those are extraordinary and rarely co-occurring. The insight the pairing produces: Han Fei's sage is a structural position, not a disposition. Examine the things of his age and then prepare to deal with them presumes someone with the access to examine and the power to prepare, and for anyone else the inherited rule is not a stump but the only available account of what has worked. That gives the fable a boundary it does not state. The man of Sung is mocked for abandoning the plough on one hare; a subordinate who abandons standing procedure on one anomaly is doing the same thing and will be treated the same way. The doctrine is sound and its application is licensed by position, which is why it reads as wisdom from a ruler and as insubordination from anyone below one.
The transferable claim is the reclassification effect: a costless return does not merely raise expectations about the future, it lowers the perceived value of the working method that preceded it. That is checkable — the prediction is that the abandonment of a reliable practice follows windfalls rather than failures — and it is a better early warning than any assessment of the new opportunity itself.
The open edge is Han Fei's criterion. Discard rules whose authority derives from conditions that no longer obtain requires knowing what those conditions were, and for most inherited practice that record does not exist. Without it the test collapses back into Greene's version, which is a mood.