Reb Feivel is sitting at home in Tarnopol, deep in his Talmud, when a gang of children starts making a racket outside.
So he leans out of the window and improvises the first story that occurs to him. Run to the synagogue — there is a sea monster there, "a creature with five feet, three eyes, and a beard like that of a goat, only it's green!"
The children scatter. He goes back to his book, smiling into his beard.
Shortly he hears running feet. Jews, heading for the synagogue. Haven't you heard? There's a sea monster — five legs, three eyes, a beard like a goat's, only it's green!
He laughs and sits back down.
Then a great crowd — men, women and children — all running. Right in front of the synagogue there's a sea monster. Five legs, three eyes, green beard.
And in the crowd he sees the rabbi.
*"'Lord of the world!' he exclaimed. 'If the rabbi himself is running with them surely there must be something happening. Where there's smoke there's fire!' Without further thought Reb Feivel grabbed his hat, left his house, and also began running. 'Who can tell?' he muttered to himself as he ran, all out of breath, toward the synagogue."*1
The first thing to notice is that the description does not drift.
Five feet. Three eyes. A goat's beard, only green. It comes back to him from the children, from the Jews in the street, from the crowd — word for word, three times.
That is not how rumours are usually depicted. The comic convention is distortion: the story grows and mutates until it is unrecognisable. This one is transmitted with perfect fidelity.
And the fidelity is what makes it convincing. A story that changes reveals that people are embellishing. A story that arrives back at you exactly as you sent it looks like a report of something observed, because that is what accurate transmission normally indicates.
Feivel's own invention returns to him wearing the signature of independent corroboration, and the signature is genuine — the details really did survive intact. They survived because there was nothing to observe, and an account with no referent has nothing to correct it toward.
The fable is structured as four visits to the same window, and each one is a different kind of evidence.
The children. Feivel knows they were told, because he told them. No information.
A few Jews running. These are people he did not tell. The story has crossed from his mouth into circulation — but they might have got it from the children.
A great crowd, men women and children. Volume. And volume is doing something specific: at this scale, he can no longer trace how anyone knows.
The rabbi. A person whose judgement Feivel defers to, in motion, in public.
Three of those are quantity and the fourth is quality, and it is the fourth that moves him. Nothing about the story changed between wave three and wave four; only the standing of one participant did.
"If the rabbi himself is running with them surely there must be something happening."
That inference is the whole page.
Feivel is not reasoning about sea monsters. He is reasoning about the rabbi — a careful, learned man who would not run through the streets for nothing. And that reasoning is sound. In almost every situation it is correct, and Feivel has decades of experience confirming it.
What it cannot survive is the possibility that the rabbi is running for exactly the same reason as everyone else: because other people are.
So the fable is not about credulity. It is about a chain of deferrals in which nobody is the origin, and the origin has forgotten he is the origin. Each person in the street is behaving reasonably, drawing a correct inference from the behaviour of people who are themselves drawing correct inferences, and the whole town ends up at the synagogue.
This is the part that makes the fable more than a joke about crowds, and it is worth stating flatly.
Feivel had certain knowledge. Not a suspicion, not a competing account — he made it up, at the window, minutes earlier, out of nothing, to get rid of some children. There is no epistemic position stronger than being the author.
And it is overridden.
"Who can tell?" he mutters, running. A man who could tell. Who is, in fact, the only person in Tarnopol who could tell.
That is a far harder claim than the usual social-proof story. Ordinary conformity experiments show people abandoning uncertain private judgements under group pressure — the line looks slightly shorter, and everyone else says longer. This shows a certainty being abandoned, and the mechanism is that the certainty gets outweighed rather than forgotten.
Feivel does not conclude the monster is real. He concludes something weaker and more corrosive: there must be something happening. His own knowledge is preserved intact and reclassified as possibly incomplete — maybe he invented a monster and a monster also turned up. And once your own certainty has been demoted to one input among many, the crowd wins on volume.
You started something and now it is coming back to you.
A number you estimated in a meeting that is now in three decks. A phrase you improvised that has become the way people describe the project. A concern you raised offhand that is now, apparently, a known issue.
The first time you hear it repeated you will feel something pleasant — a small proprietary warmth. That is the smile into the beard.
The second time you will feel something else, and it is worth catching: a faint upward revision of your own confidence. You said the number was about forty. Three people have now said forty. Forty is starting to feel more like forty than it did when you said it.
Nothing has been added. The only new information in the room is your own estimate, returned.
So build the one habit that defends against this, and it has to be built in advance because it cannot be reconstructed later. Write down where things came from. Not elaborately — forty, my guess, Tuesday, no data. The provenance is the thing that gets lost first, and it is the only thing that would have saved Feivel.
And the second, harder half. Watch what senior people's agreement does to you. When the person whose judgement you most respect starts using your number, you will treat that as confirmation. It is not. They got it from you, and they have no more information than you did — and they are, at that moment, doing to you exactly what the rabbi did to Feivel.
The fable's placement matters. Greene prints it in the margin of Mesmer's salon, in a law about creating a following.
Every case in the body of Law 27 has a clean division: an operator who knows, and followers who do not. Borri knows there is no philosopher's stone. Mesmer knows what his assistants are doing. The whole technique is described from a position outside the belief.
Reb Feivel dissolves that division, and does it in eight lines. The manufacturer of the belief joins the crowd chasing it, and the mechanism that catches him is the same one he used — a story arriving from many mouths, apparently independently, endorsed by someone credible.
Which is the finding the chapter has no room for. Greene's contagion model is explicitly lateral — "one believer infecting the other"2 — and lateral transmission has no way of distinguishing members from authors. The crowd removes the operator's distance as reliably as it removes anyone else's, and Law 27 contains no instruction about that, because it is written from a vantage the fable says does not exist.
It is the second complete Ausubel fable in Cluster G — the first is the Chelm cobbler at Law 26 — and both do the same job: they carry, in the margin, the objection the chapter needs to be false.
The cobbler shows a scapegoating that works with nothing concealed, against a law premised on concealment. Feivel shows a manufactured belief capturing its manufacturer, against a law premised on an operator who stands outside.
Neither is answered. Both are better attributed than the text they decorate — named collection, named editor, named year — and both are funnier and more precise than the argument they interrupt.
Attribution: complete. A Treasury of Jewish Folklore, Nathan Ausubel ed., 1948, quoted in full.1 The same collection supplies the Chelm cobbler at Law 26 and the cheese-dividing fable at Law 20.
🚩 It is a joke and is being read as an argument. Real interpretive risk, flagged rather than hidden. Yiddish folk humour of this kind is making fun of reasoning, not advancing a theory of information cascades. What survives is the structure, which is coherent whether or not anyone in Tarnopol meant it as one.
Tension with the chapter's operator-outside premise. Every body case in Law 27 assumes a knowing operator. This fable's operator ends up running.
Open question. Feivel's inference about the rabbi is sound — a careful man does not run for nothing. Is there any version of deference to expertise that survives the possibility that the expert is deferring too, and does that require knowing the provenance of every claim you hold?
Greene's Preface argues that the reader can occupy a position outside the game — studying the laws, applying them deliberately, being "more systematic" than the unconscious practitioners.
The vault's reading of that Preface already identifies the problem: the book's own closing law tells the reader to disregard the book, and its opening argues that anyone claiming to stand outside the power game is its most dangerous player.
Reb Feivel is the third instance of that structure and the cheapest. Eight lines, in a margin, showing that the position the whole book is written from — outside, knowing, unmoved — is not available even to the person who authored the belief.
And it converges with the law's own Reversal from an unexpected direction. Greene warns that the crowd may turn on you.3 The fable warns that the crowd may absorb you, which is worse and has no exit, because there is no moment at which you are attacked and therefore no moment at which you notice.
Psychology — The Question You Actually Answered
The mechanism is the swap. Standing in front of a used car you cannot evaluate, the mind quietly answers do I trust the seller instead of is this a good car, and returns the substitute answer with the confidence the original would have earned — without your noticing the exchange.
Reb Feivel is the strongest case of that mechanism available anywhere, because the hard question was not hard for him.
Is there a monster is normally unanswerable from a window; you would have to go and look. For Feivel it is answerable with total certainty, at zero cost, from memory. He made it up.
And the swap happens anyway. The question his mind processes is would the rabbi run for nothing, which is a question about a person rather than about a synagogue — and it returns a fast, confident, correct answer that has nothing to do with the monster.
What that adds to the substitution account is a limit case nobody would have predicted. The usual model has substitution triggered by difficulty: the mind swaps because the real question is beyond it. Feivel's is triggered by social salience — a highly legible cue about a trusted person arrives, and the mind processes the available question rather than the relevant one, even when the relevant one is already answered.
Which means the substitution is not a shortcut taken under cognitive load. It is the default route, and private certainty does not automatically win against it — it has to be actively held, and holding it requires remembering where your own knowledge came from.
History — The Homefront Supply-Collapse & Rumor Ecology
Petrograd, winter 1916: white bread has vanished from the capital of one of the greatest wheat-growing countries on earth. The grain exists — a traveller who has just crossed Siberia describes it as overflowing with food — and the empire cannot get a loaf into the city. What fills the gap is a rumour ecology in which nothing can be sourced and everything is circulating.
Tarnopol is that ecology in miniature, and running the two together isolates the variable that decides whether a false story stops.
In a small town with a single origin, the correction exists and is one person. Feivel could have leaned out and said I made it up. The story was stoppable for as long as he remembered he had started it.
What breaks is not the correction's availability but its custody. By the fourth wave the account has passed through enough mouths that provenance has been stripped, and it is stripped for Feivel too — he is no longer receiving his own invention, he is receiving what everyone is saying, and those are different objects even though the words are identical.
Petrograd is that condition at civic scale and permanently: no claim has an origin, so no claim can be retracted by anyone, and a population that cannot source anything eventually stops trying and acts on dread.
Which gives the general form. A rumour is killable only while somebody still holds its provenance, and provenance is the first thing transmission destroys — faster than content, which is why the monster's five feet and three eyes survive perfectly while the fact that Feivel invented them does not survive four repetitions in his own head.
Sharpest implication. Feivel is not deceived. He knows — with the strongest certainty available to anyone, authorial certainty — and he runs anyway, because the crowd does not defeat his knowledge, it demotes it: there must be something happening, maybe he invented a monster and a monster also turned up. Once private certainty has been reclassified as one input among many, volume wins by arithmetic. Which means the defence is not confidence and not scepticism. It is provenance — and provenance is the first thing that gets stripped, including from the person who supplied it.
Generative questions