Somebody is telling you something you cannot possibly believe, and you are being polite about it, and then they open their shirt.
Whatever you thought was happening in the conversation stops. It is no longer an account you are being asked to accept or decline. There is now a body in front of you and a thing on it, and you have to look, and looking commits you to saying something about what you see.
That is a completely different kind of move from telling a story, and everybody who has ever been on either end of it knows the difference.
Ādi 17 has it, and the man doing it is the magistrate who banned the chanting.
The Kazi tells it himself. The night he went into a house and broke the drum, something came.
A figure with a man's body and a lion's face landed on his chest.1 It spoke:
"I shall immediately bifurcate your chest as you broke the mṛdaṅga."
Then he bares his chest and says: just see the marks of His nails on my heart.2
And the report does not stop with him. His orderlies — the men who carried out the raid with him — turn up afterwards with singed beards and blistered cheeks.3
This is the feature that makes it a page rather than a marvel.
The corpus is full of supernatural enforcement, and almost all of it is narrated by devotees. Somebody offends and something happens and the community reports it.
Here the enforcement is narrated by the opponent, in the first person, in his own defence, in front of a crowd — and the physical evidence is on his own body, offered by him.
That is the strongest available form of a supernatural claim. A hostile witness is more credible than a friendly one, and a hostile witness who produces marks on himself and invites inspection is more credible still.
Which is exactly why a partisan text would construct it this way, and why the page has to say both things at once. The narrative form here is optimised for credibility, and optimised-for-credibility is what an invented account and a true account both look like.
The text supports all three and settles none.
Enforcement. Nṛsiṁha — the man-lion incarnation, the tradition's tearing form, the one that appears when a devotee is persecuted — came, threatened, and marked him. The corpus's own reading, and it is internally coherent: this is precisely the situation Nṛsiṁha exists for.
Experience. A man who has just done something he half-knew was transgressive, in a town he governs and lives in, has a night of extreme terror. The phenomenology in the passage — a figure on the chest, a threat, immobility, waking marked — is one of the most consistently reported experiences in human record, across cultures that have no contact with each other.
Construction. The corpus needed the ban to be lifted with the magistrate's dignity intact and its own doctrine vindicated, and an opponent's testimony to divine intervention does both at once.
The corpus is not aware of the second and third. It reports the first as fact, because from inside there is nothing else it could be.
The nail marks are the load-bearing detail and it is worth being precise about what they accomplish.
Without them, this is a dream. A dream is entirely interior, and an opponent reporting a frightening dream can be read as an opponent who was frightened — which is a psychological fact and does not compel anybody.
With them, the claim becomes external and inspectable. Something is on his chest; he is inviting people to look; and whatever the marks are, they exist.
That is a move from testimony to evidence, and it is the same move the corpus makes at every point where it most needs to be believed — the bodily transformations reported by witnesses, the joints measured, the food gone from the plate.4 The tradition consistently prefers claims with a physical residue over claims about interior states, which is interesting in a corpus whose decisive doctrinal tests are all invisible.
Where the corpus's theology runs on unfalsifiable interiors, its evidence runs on bodies. Those are opposite instincts and they coexist without comment.
Worth pushing on why a tradition would split that way, because the split is consistent enough across this corpus to look like a policy rather than an accident.
The interior criteria govern who is entitled to what — eligibility, authenticity, the line between love and appetite. Those are questions a tradition has to answer for its own members, where a wrong answer lets the wrong person in or the right person out. An unfalsifiable criterion is, for that job, a feature: it keeps adjudication inside the community and out of reach of anyone applying an external standard.
The physical evidence governs whether outsiders should believe any of it. There the incentives reverse completely. An interior claim persuades nobody who is not already persuaded, so the corpus reaches for marks, transformations, empty plates and singed beards — the class of evidence a sceptic can in principle inspect.
So the split is not incoherence. It is two audiences, and the corpus is using the appropriate instrument for each: unfalsifiable criteria to govern insiders, inspectable phenomena to reach outsiders. Neither is chosen for truth-tracking, and the tradition never had to notice that, because no reader before the modern period was likely to hold both departments in view at once.
The singed beards do the same work at one remove. They are corroboration from a second set of people, they are visible, and they are the kind of injury that has ordinary causes — which is precisely why including them is either honest reporting or very good construction.
One further asymmetry worth naming. The Kazi's marks are on his chest and are offered by him. The orderlies' injuries are reported about them, by the narrative, and no orderly speaks. The strongest piece of testimony in the passage is first-person and the corroboration is not — which is the ordinary shape of how a story consolidates, and is worth marking in a passage whose whole force is that the witness is hostile.
Worth marking what the passage is not doing, because it sits in a chapter that contains something much worse.
This is not the polemical material. The Kazi is not disparaging his own tradition here; he is reporting an experience and asking people to look at his chest. He comes out of it as a frightened man being candid in public, and the text treats him with a kind of respect.
The disparaging material — where the text has him repudiate his own scripture — is elsewhere in the same conversation and is handled on its own page, with the flag in the body.5
One chapter, two operations on the same man, and the difference between them is worth keeping visible. Reporting a nightmare is something a person might do. Announcing that your scriptures are full of mistaken ideas is something an author does to a character.
Layer A gives the Kazi's account in his own voice, with the marks and the orderlies, and does not comment.
Layer B treats it as straightforward Nṛsiṁha intervention and connects it to the doctrine that offence against devotees is dangerous.6 That is the verse's own reading, applied.
The one thing worth flagging is what neither layer does: nobody in the text or the commentary examines the chest. The marks are offered for inspection and no inspection is reported. The corpus's most inspectable piece of evidence is described and never checked, which is a small thing and is the kind of small thing that separates a report from a record.
Somebody tells you something implausible and produces physical evidence. Notice the shift in what is being asked of you — you have moved from assessing a claim to assessing an object, and the two use completely different faculties. The object is usually real and rarely settles the claim.
You did something at work you half-knew was over the line, and you slept badly, and you woke with a feeling of having been visited. That is an extremely common sequence and the visitation is not the interesting part. What you half-knew is.
You are recording an event and a hostile witness has given you the account you needed. That is the best evidence you will get and it is also, structurally, exactly what you would have made up. Say who told you, say when, and say whether anyone checked — because the checking is what a later reader will look for and not find.
Somebody in authority acts against you and then reports being frightened. Do not press the advantage. The corpus's own next move, having received this testimony, is a conversation that leaves the man in his job, and that is why the episode ends.
The evidence is a first-person account from the adverse party, with an invitation to inspect a physical mark, plus a second-hand report of corroborating injuries to his subordinates.7 As religious testimony goes, this is an unusually strong evidential form.
The tension the source doesn't close: the marks are offered and never examined. The corpus's most checkable claim in the chapter is left unchecked by everyone present and by the commentary.
A second one, and it runs through the corpus. Its theology rests on interior states that cannot be inspected, and its evidence is consistently physical — bodies, marks, transformations, food. Two opposite epistemic instincts, both operating, never compared.
Open: what the marks were. The text asserts their existence and their cause and reports no examination, and this vault has no independent source for the episode. Nothing on this page reaches past the text on that point.
This page belongs with The Chand Kazi Confrontation, The Kinship Gambit and The Cow Argument and the Koran as the four faces of one evening. Read as a set, they show the corpus doing four different things to one man in a single chapter — defeating him tactically, being defused by him, receiving his testimony respectfully, and writing his capitulation for him. The four operations are not consistent with each other, and only splitting the pages makes that visible.
Against The Nṛsiṁha Mood and the Fear Offence there is a pairing the corpus does not draw. There, Caitanya enters the Nṛsiṁha mood in the street, frightens people, and stops himself on the grounds that causing fear is an offence. Here the same form appears at night and frightens a man badly enough to leave marks, and nobody raises the offence. The corpus has a rule about causing fear and applies it to its founder's public conduct and not to its God's nocturnal enforcement.
And it stands beside He Refuses the Identification as a study in what the corpus accepts as evidence. That page shows the identity claim being unfalsifiable by anything its subject says. This one shows the tradition's evidential weight falling, as it always does, on third-party reports of physical phenomena — which is the right instinct and is sourced, here as everywhere, from inside.
SUNDS: Sleep Paralysis or Brugada Syndrome? — that page describes a syndrome in which apparently healthy young men die in their sleep with terror on their faces, documented across the Philippines, Thailand, Japan and Laos, with no structural cause found at autopsy. The clinical account: awareness during REM while the body is paralysed, terror, and — in the predisposed — a fatal arrhythmia. The cultural account, in Tagalog bangungot, is that a nightmare attacks the person.
The Kazi's report has the same phenomenology, item for item: a figure arriving at night, positioned on the chest, a threat, and waking marked.
What the pair makes visible is not that the corpus is wrong. It is the relationship between the two accounts, which that page states carefully: the cultural interpretation is usually treated as folk explanation that does not alter the underlying medical reality — and the page notes that this treatment is itself contested.
Applied here, three things follow that neither page has alone.
The phenomenology is cross-cultural and the interpretation is local. A chest-pressing figure appears in traditions with no contact. Sixteenth-century Bengal supplies a lion-faced man because Nṛsiṁha is the available form for persecuted devotion; elsewhere the same posture is supplied by a demon, an old woman, a djinn. The experience is the constant and the face is the culture's.
The interpretation is not idle. The Kazi's reading of his own night did work: it gave him a reason to lift the ban that was not surrender to a crowd. Had he read it as indigestion, the political situation would have been unchanged and much worse for him. A folk explanation that resolves a standoff is not a failed medical explanation; it is doing a different job.
And the marks remain unaccounted for by either frame. Sleep paralysis does not produce nail marks. The corpus says a lion did. The honest position is that a man under acute stress, alone, in the dark, has hands — and that this is a hypothesis rather than a finding, offered here because a page that lists three readings should not quietly omit the fourth.
Sharpest implication. The strongest evidence in this chapter comes from the hostile party, on his own body, offered for inspection — and nobody inspects it. A tradition's most checkable claim tends to be the one nobody checks, because by the time it is offered everyone present already knows what they are going to conclude, and looking would only introduce risk.
Generative questions.