Two frogs live in the same pool. Summer comes, the pool dries, and they set out together to find somewhere else.
On the way they pass a deep well, full of water. One frog says: let's go down, it will give us shelter and food.
The other replies:
*"But suppose the water should fail us, how can we get out again from so great a depth?"*1
Moral: Do nothing without a regard to the consequences.
Sixty-odd words, in the margin of Law 29, and it does something the whole chapter around it does not manage.
Look at what the cautious frog actually asks.
He does not dispute that the well has water. He does not say the plan will fail. He grants the whole thing — good water, shelter, food, exactly as advertised — and asks a question about a later state in which the plan has already worked and then stopped working.
That is a different question from the one careful people usually ask. The ordinary caution is will this succeed? The frog's is what is my position if it succeeds and then conditions change?
And the sting is in the phrase from so great a depth. The very property that makes the well attractive — it is deep, so it has water when the shallow pool did not — is the property that makes it inescapable. The asset and the trap are the same feature. You cannot have the water without the depth.
Greene's law asks you to plan all the way to the end and to take account of "all the possible consequences, obstacles, and twists of fortune."2 That is a completeness standard: see more, further, in more detail.
The frog is not applying a completeness standard. He is applying a single test, and it is one specific question:
Can I get out?
This is enormously cheaper than the chapter's method and it catches a different class of error. Completeness fails on unknown unknowns — you cannot enumerate the twists of fortune you have not thought of, and the Athenians proved you will not. Reversibility does not require enumeration. You do not need to know what will go wrong in the well. You only need to know that if anything does, you are eighty feet down with wet walls.
Greene's chapter never states this test. It is present in the margin, in a fable he printed, and the Keys go a different way entirely — toward foresight as a quantity you accumulate, with the gods as the limit case.
Turn back to Law 28's epigraph and read it again.
The Two Adventurers. A knight and his companion find a post by a torrent promising "that which has never been seen by any knight errant" to whoever crosses the water, takes up a stone elephant, and carries it "in one breath" to a mountain summit.
The companion reasons it through. The water may be deep and fast. The elephant could be carried four steps, but not to a summit in one breath — "unless it should be the dwarf figure of an elephant, fit only to be placed on the top of a stick, and then what honor would there be in such an adventure?" He concludes "there is some deception in this writing" and leaves.3
Same structure as the frogs. Two companions. One proposes, one calculates. Word for word, the calculating one is doing what the cautious frog does.
And the verdict is reversed.
In Law 28 he is "the reasoner," he departs, and the fable ends: "They who calculate with too great nicety every difficulty and obstacle which is likely to lie in their way, lose that time in hesitation, which the more daring seize."4
In Law 29 the calculating one is the wise one and the moral is do nothing without a regard to the consequences.
Two marginals, 185 lines apart, identical in structure, opposite in verdict, and Greene connects them nowhere.
Worth noting, because the two fables are not equally honest about their own evidence.
The frog's objection is never tested. Nobody descends. The story stops at the question.
The reasoner's objections are tested, and every one of them holds. The water is crossed — but by a man who "rushed with his eyes closed," which is not a refutation of the depth, it is a refusal to look at it. The elephant is carried, and it shrieks. And the payoff, the thing the inscription promised? It never arrives. What arrives is a crowd who "were awed by his presence" and proclaim him successor to a king "who had recently died."5
The inscription promised a sight never seen by any knight errant. He gets a throne, because of a bereavement, in a town he had not heard of that morning.
So in Law 28 the calculating companion is mocked by a story whose own narrative confirms him and delivers its reward by coincidence. In Law 29 the calculating frog is endorsed by a story that never tests him at all. Neither fable earns its moral, and they are stacked in opposite directions inside one book.
⚠ One attribution needs stating precisely, because the pair is easy to compress wrongly.
The Two Adventurers is Jean de La Fontaine, 1621–1695 (l.4072). The Two Frogs is Aesop, sixth century B.C. (l.4253). They are not the same author, and the structural inversion above holds between two different fabulists.
There is a same-author contradiction in this cluster, and it is a different pair. Law 28's other marginal, The Boy and the Nettle, is also Aesop, and its moral is "Do boldly what you do at all."
So: Aesop supplies Law 28's cleanest statement and Law 29's counter-statement, seventy-two lines apart, and the two instructions cannot both be followed. Do boldly what you do at all against do nothing without a regard to the consequences. One says the hesitation is the injury; the other says the hesitation is the whole of the wisdom.
That is the sharper finding, and it belongs to the Aesop pair rather than to the La Fontaine one.
Nothing.
The fable is printed between the Balboa narrative's two halves and never referred to. No line of the Keys picks up reversibility. The Reversal, which is where a contrary position would go, is spent declaring that the law has no reversal.
This matters more than it would for an ordinary uncited quotation, because the frog's test is better than the chapter's method on the chapter's own worked case. Athens did not fail a completeness test in some exotic way; the four things they missed — distance, defenders at home, enemies combining, forces stretched — are all versions of one question. If this goes badly, how do we get the fleet home from Sicily? They could not. That is the whole disaster in one line, and the line is available in the margin of the same chapter.
Someone offers you the deep well. Better terms, more room, real water — and everything about it is exactly as described.
Do not spend the afternoon listing what could go wrong. You will produce four items, they will all be plausible, and none of them will be the thing that happens.
Ask the one question instead. If I am in this and it turns, what does leaving cost?
Then make yourself answer it in units. Not it would be difficult. Months. Money. Which relationships end. What I would have to say publicly. Whether the thing I would go back to still exists by then.
Now look at the number and notice something. The feature that makes this attractive is probably the same feature generating that number — the depth is why there's water. Long contract, big client, the move to a new city, the co-founder who brings everything you lack. In every case the size of the upside is the depth of the well.
That does not mean don't. Frogs die in dry pools too, and the fable never says the well was the wrong choice — it says the question was worth asking before the descent.
It means: know the number before you go down, because the one thing you will not be able to do afterwards is find out cheaply.
Strongest evidence. The fable is complete, named, dated, and its criterion is stated in the frog's own words rather than reconstructed. Twenty-seventh complete marginal fable in this build, and one of the few that supplies a test rather than an observation.
Tension — the fable and the previous law's epigraph give opposite verdicts on identical behaviour. Documented above. 28↔29 is now on the do-not-merge tension-pair list.
Tension — Aesop contradicts himself across the two laws. Do boldly what you do at all against do nothing without a regard to the consequences, seventy-two lines apart, both printed as authorities.
Tension — the fable is never tested and its counterpart's objections are all confirmed. Neither moral is earned by its own narrative.
🚩 [POPULAR SOURCE] — the Aesop attribution is conventional and the text is a standard translation; no edition given.
Open questions. Is reversibility a genuine substitute for completeness, or only a cheap first filter that catches the common case and misses the catastrophic one? And: the frog's test assumes exit is a coherent notion. What does it become in situations where every option is irreversible, which is most of the interesting ones?
Greene wrote the opposite of this fable and put it in a book of his own, and the contradiction is explicit rather than inferred.
Strategy 4: The Death-Ground Strategy argues that having a way out is the problem. The fallback feels like security and is "the exact thing keeping you from finishing" — because part of you knows you could survive failing, no part of you fully commits, and half-commitment is why the thing never gets done. The prescription is to remove the exit deliberately.
Set that against the frog: how can we get out again from so great a depth?
These are not different emphases. They are contradictory instructions about the same decision. Death Ground says burn the way back so you will fight properly. The frog says never enter anything whose way back you have not priced. Greene prints one as his own doctrine and the other as an authority, and never puts them in the same paragraph.
The reconciliation that suggests itself — remove the exit after you have priced it, so the commitment is chosen rather than discovered — is a real position and neither text states it. It also quietly concedes the frog's point, because you cannot deliberately burn a bridge you never surveyed.
And note the same structure at the previous law: the chapter that tells you to enter with boldness prints, as its Authority, Gracián's "it would be safer to do nothing." Three laws' worth of marginals now argue against the bodies they decorate.
Behavioural economics — why the depth argues for staying. Sunk Cost Fallacy describes the $100 concert ticket that drags an exhausted person out of the house: the money is gone and irrelevant to tonight's choice, and it is nonetheless dispositive. The page's better observation is that this is not stupidity — it is "a sophisticated adaptation to a world where commitment matters," the mechanism that makes people honour commitments and stick with projects long enough to work.
Put that next to the well and something falls out that neither text has. The frog's test is run before the descent, and it has to be, because the descent installs the exact bias that will prevent it from being run afterwards. Every metre down is a cost paid, and paid costs argue for continuing. So the well is not merely hard to climb out of physically — the effort of getting in becomes an argument against getting out, and the argument grows with the depth.
Which means the frog's question is not one of a set of prudent checks you can perform at any stage. It is a one-time, non-renewable option that expires the moment you commit, and after that your own reasoning is working for the well. That is a stronger claim than either the fable or the sunk-cost literature makes, and it comes from putting them together.
Technology and economics — the well that a whole civilisation is in. Technology Lock-In and Path Dependence traces QWERTY: a layout designed in the 1870s to stop typewriters jamming, demonstrably worse than the alternatives, universal 150 years later because early adoption created an installed base which created incentive to match it which created more installed base. Switching would mean retraining billions. Nobody is trapped, and everybody is.
That is the frog's problem with the individual removed. No single actor descended a well. The depth accumulated from a great many separately reasonable steps, none of which was the descent.
The insight neither domain generates alone: the frog's test presumes a moment of entry — a discrete point where the question can be asked and the choice made. Path dependence describes irreversibility with no such moment. If reversibility is the criterion that actually protects you, then the dangerous commitments are not the deep wells at all. They are the ones with no threshold to stand at, where each step is small, obviously fine, and the depth is only visible in aggregate. Law 29's whole apparatus — plan to the end, before you begin — has no purchase on those, and neither does the fable. The commitments that will hold you are precisely the ones you were never offered.
Sharpest implication. Reversibility is a better test than foresight and costs almost nothing to run: you do not need to know what will go wrong, only what leaving costs if something does. And the feature that makes an opportunity worth taking is usually the same feature that makes it hard to leave — the well is attractive because it is deep. The question is only available once, before you go down, because afterwards the descent itself becomes the argument for staying.
Generative questions.