Two men are sentenced to death by the sultan of Persia. One of them knows the sultan loves his stallion, and offers a deal: give me a year and I will teach the horse to fly. The sultan, taken with the idea of riding the only flying horse in the world, agrees.
The other prisoner is unimpressed.
*"You know horses don't fly. What made you come up with a crazy idea like that? You're only postponing the inevitable."*1
And the reply, which is the whole marginal:
"Not so," said the first prisoner. "I have actually given myself four chances for freedom. First, the sultan might die during the year. Second, I might die. Third, the horse might die. And fourth … I might teach the horse to fly!"2
The Craft of Power, R.G.H. Siu, 1979.
The sceptic's argument is sound: horses do not fly, and the year ends in the same execution.
He is reasoning about the promise. The prisoner is reasoning about the interval.
That is the distinction the marginal exists to make, and it is sharper than the surrounding chapter's version of it. The chapter says be patient, time is your ally. This says something specific and mechanical:
A delay is not a postponement of a fixed outcome. It is an exposure to everything that can happen in the meantime.
Three of the four chances have nothing to do with the prisoner at all. The sultan's death, his own, the horse's — he has not improved his position, he has purchased contact with a set of events he does not control, and one of those events is enough.
⚠ Note the honesty of including "second, I might die." He counts a chance that is not a gain for him, which is what makes the list a real enumeration of outcomes rather than a piece of consolation. The regime he is under ends, one way or another, in four possible manners.
The chapter's own case for waiting rests on the operator being skilful — Fouché reading tides, Ieyasu foreseeing disaster. This marginal makes no such requirement.
The prisoner has no forecast. He does not claim the sultan is ill or the horse is old. His argument is purely about the arithmetic of open time, and it works precisely because he cannot predict anything: the more that is uncertain, the more the interval is worth.
Which inverts the usual relationship between knowledge and patience. You would think waiting was for people who can see what is coming. This says waiting is most valuable to the person who can see nothing, because their position cannot get worse and the distribution of futures is wide.
⚠ And the precondition is the one Law 35 keeps omitting and this marginal supplies by construction: the technique only makes sense when the certain outcome is bad. A man not under sentence of death should not trade a known position for an interval. The flying horse is a strategy for people with nothing, and the chapter's other exemplars are all people with everything.
The transaction repays attention, because the prisoner does not ask for mercy.
He identifies something the sultan wants — "knowing how much the sultan loved his stallion" — and constructs an offer around it. The sultan does not grant a reprieve; he buys a service, and the reprieve is the delivery schedule.
That is a considerably better structure than pleading, for the reason established at Law 34's three outward strategies: a request states an asymmetry and a transaction does not. The prisoner is briefly a supplier rather than a condemned man, and the year is a contract term.
And the offer is unfalsifiable for exactly as long as it needs to be. Nobody can demonstrate mid-year that the horse will not learn to fly. The deliverable is remote, spectacular, and unverifiable until the deadline — which is the same structure the corpus recorded across Law 32's whole cluster: Bragadino's alchemy, Hartzell's Drake estate, the fantasy that requires distance. ⚠ The prisoner is running a con, and the chapter files him under timing.
⚠ The attribution is The Craft of Power, R.G.H. Siu, 1979 — a management book, not a Persian source.
🚩 SECONDARY WITHOUT PRIMARY · 🚩 TRANSLATION DISTORTION — this is a widely travelled folk tale with versions attached to Nasreddin Hodja, to Sufi teaching collections and elsewhere, and the "sultan of Persia" framing is a Western management-literature setting, not a located source. No original is named.
🚩 The parable is a constructed illustration, not evidence. It demonstrates the logic of the four chances and demonstrates nothing about whether the reasoning works — no version of the story reports the outcome, which is convenient for a story arguing that the interval is worth having.
Hold the argument, not the anecdote. The enumeration is genuinely instructive; the setting is decorative and the provenance is a 1979 business book.
You are facing something with a bad and apparently fixed outcome, and every option you can see is a version of accepting it.
Stop evaluating the promise and start counting the interval. The prisoner's method is a specific exercise: list every event that would change your situation, including the ones you have no influence over and the ones you would not enjoy. Sultan dies, horse dies, you die. Most people list only their own moves and conclude there is nothing to do.
Then ask what the interval costs. This is the part the marginal skips and it is where the technique fails in ordinary life. The prisoner pays nothing — his alternative is immediate death, so the year is free. You are usually paying something for the delay: a worse position later, an option that expires, credibility spent on a promise you cannot keep. Count it honestly, because the whole argument turns on the certain outcome being bad enough.
And construct a transaction, not a plea. The reprieve worked because it arrived attached to something the sultan wanted. Find the thing your counterpart wants that you can plausibly be the supplier of, and the extension becomes a delivery schedule rather than a favour. ⚠ Note what this requires: the deliverable must be remote enough that nobody can check on it early, which is also precisely what makes the technique dishonest when the delivery was never intended.
The honest version of the whole thing is short: when the certain outcome is bad, buy time at any price you can actually afford, and be clear with yourself about which of the four chances you are relying on.
Strongest evidence. A clean logical enumeration — four disjoint routes out of a certain outcome, three of them outside the actor's control — that makes a point about intervals which does not depend on the story being true.
Tension — the marginal argues for delay from weakness and the chapter's exemplars all delay from strength. Fouché and Ieyasu had reserves; the prisoner has a sentence.
Tension — the reasoning is a con. The deliverable is unverifiable until the deadline, which is Law 32's fantasy structure, and Greene files it under timing without comment.
🚩 [POPULAR SOURCE] · 🚩 SECONDARY WITHOUT PRIMARY · 🚩 TRANSLATION DISTORTION — a travelling folk tale given a Persian setting in a 1979 American management book, with no original named. 🚩 A constructed illustration reported without an outcome.
Open questions. Does the argument survive when the interval has a price? The prisoner's case is the limiting one where delay costs nothing, and every real instance has a cost — the corpus has no case testing where the break-even sits, and the chapter's advice to play for time never mentions that time is bought.
This is the third marginal in Law 35 that is better argued than the chapter around it, and the build's tally of that finding is now seven.
Against Fouché the pairing is genuinely illuminating. Fouché's 1795 escape is the flying horse executed for real: the arrest order was issued, he did not answer it, and "too much time had passed." He did not beat the charge and he did not need to — one of the four chances came in. ⚠ The chapter prints the parable and the case within twenty lines and does not connect them.
Against Law 32's cluster the tension is sharper. Fantasy requires an oppressive reality and its cases — Bragadino, Hartzell, the Drake estate — are all remote, spectacular, unverifiable deliverables sold to people in an unsatisfactory present. The flying horse is that exact structure, sold upward to a sultan. Greene files identical mechanics under two laws and treats one as fraud and the other as wisdom.
Option pricing — the value of the interval itself. Loss Aversion establishes that outcomes are valued against a reference point rather than absolutely, and that the same distribution is judged very differently depending on where the actor believes they currently stand.
That is why the sceptic and the prisoner reach opposite conclusions from identical facts.
The insight neither produces alone: both men agree the horse will not fly. They disagree about the reference point. The sceptic prices the year against a neutral baseline, from which a doomed year looks like nothing gained; the prisoner prices it against certain death, from which any nonzero probability is pure gain and there is no downside to weigh, because the worst case is already the default. So the marginal is not a lesson about optimism or nerve — it is a demonstration that the reference point determines whether an interval is an asset or an expense, and that people in bad positions systematically mis-price it upward to a neutral baseline they no longer occupy. The prisoner's real skill is having correctly identified where he was standing, which is the one thing the condemned reliably get wrong.
Contemplative practice — acting without a forecast. Attainment as Improvisational Freedom holds that effective response comes from availability to what actually arrives rather than from executing against a prediction, and treats the demand for a plan as itself a constraint.
The prisoner has no forecast and that is the source of his position, not a defect in it.
What the pairing produces: Greene's chapter requires the operator to be a "detective of the right moment" — to read tides, sniff trends, see two steps ahead. The flying horse needs none of that and is stronger for it. Three of the four chances are events the prisoner cannot foresee, influence or even hope for specifically, and his argument works because the future is unspecified: the wider the distribution, the more the interval is worth. Which inverts the chapter's premise about knowledge and patience — prediction narrows the set of futures you are prepared to accept, and the person with no prediction is the one still available to all of them. ⚠ That is a real fork in Law 35, not a synthesis: the chapter says wait until you can see; the marginal says buy time precisely because you cannot.
Sharpest implication. You're only postponing the inevitable is sound reasoning about the promise and irrelevant to the interval. A delay is not a postponement of a fixed outcome, it is exposure to everything that can happen in the meantime — and three of the prisoner's four chances have nothing to do with him. Which inverts the usual link between knowledge and patience: the interval is worth most to the person who can predict nothing, because the distribution is widest. And the reference point does all the work — the sceptic prices the year from neutral, the prisoner from certain death, and the condemned reliably make the sceptic's mistake.
Generative questions.