There is a thing you have and a word for it, and there is a different word you use when somebody official is asking.
The word you use with people who know you is accurate and would be useless in a form. The word you use on the form is a category — one of the available ones, close enough, and it does the thing the accurate word cannot, which is get processed.
Everybody who has ever had a condition that does not fit a box has done this. And everybody who has done it knows the peculiar feeling of hearing yourself say the box word out loud: it is not a lie, exactly, and it is not what you mean, and it is what will work.
Madhya-līlā 18 has the corpus's protagonist do this about his own divine ecstasy, to soldiers, with four men's lives on it.
Caitanya collapses on the road to Prayāga. Breathing stopped, foam at the mouth.1
Ten Pāṭhān cavalrymen ride up and read the scene: a sannyāsī who must have been carrying gold, four companions standing over him, and an obvious inference. They conclude the four poisoned him with dhuturā and robbed him, and arrest them for execution.
The Sanoḍiyā brāhmaṇa offers the defence: "This sannyāsī sometimes falls unconscious due to the influence of a disease."2
And then Caitanya regains consciousness and says it himself, to the soldiers:
"Due to epilepsy, I sometimes fall unconscious. Out of their mercy, these four men maintain Me."3
Before anything else, the arrest deserves defending, because the corpus records something unusual: an opponent doing sound work.
The inference the Pāṭhāns draw is this. A sannyāsī of visible standing is travelling. Sannyāsīs on that road carry donations. He is lying unconscious with foam at his mouth — which is exactly the presentation of dhuturā poisoning, the standard method for robbing travellers in that period and the very charge levelled in a different chapter against a young brāhmaṇa.10 Four men are standing over him. There is no gold on him.
Every premise is true and the conclusion is wrong.
The corpus does not mock them. It does not present them as brutal, credulous or hostile; it presents them as a patrol reading a scene competently, arriving at the likeliest explanation, and acting on it — which is what a patrol is for.
That matters for what follows. The epilepsy sentence is not spoken to fools. It is spoken to people whose reasoning is good and whose priors do not include a category for what is actually happening, and no true account would have moved them, because the true account is not in the space of hypotheses they can weigh.
Which is the general situation of anybody whose condition falls outside the categories of the people assessing it. The competence of the assessors makes it worse rather than better, because a competent assessor is precisely the one who will not accept an explanation with no evidential support.
Take the sentence at full weight, because the corpus does not soften it and neither should a reader.
This is the condition the entire Antya-līlā is about. Sixty-two chapters of sāttvika-vikāra, of joints separated by six inches with only skin bridging them, of a body found in a locked room and recovered at the Siṁha-dvāra — the corpus's principal evidence for who this man is.
And here he supplies a naturalistic diagnosis of it, in the first person, to strangers, because the naturalistic account is the one that keeps his companions alive.
Three things follow that the corpus does not draw.
He knows the naturalistic reading exists. It is not that outsiders might misread him; he has the alternative account ready, correctly labelled, and can deploy it under pressure.
He judges it more credible than the true one, to that audience, at that moment — which is a judgement about his own evidence, made by the person best placed to make it.
And the corpus records it without embarrassment. No gloss, no explanation that he was speaking in a lower register, no note that the word was a concession. The sentence stands.
This is the strongest cross-domain material in the whole ingest, and it needs no editorialising because the corpus has already done the work.
The same physical condition — collapse, rigidity, foam, stopped breath, deformation of the joints — is given three names by three parties.4
Divine ecstasy, by the devotees, throughout: the eight transformations, ardha-bāhya, the technical vocabulary of an entire tradition.
Demonic possession, by the fisherman who nets him from the sea, touches him, and catches it: "as soon as I touched it, a ghost entered my heart… if I die, who will take care of my wife and children?" — and who cannot recognise him: "I have seen the Lord many times, but this is not He. This body is very deformed."
Epilepsy, by Caitanya himself, to soldiers.
Three framings, three social positions, one set of symptoms. The corpus records all three and adjudicates none.
That is a remarkable thing for a devotional text to leave standing. It could have suppressed the fisherman, or explained the epilepsy sentence away, and it does neither — and what results is a piece of evidence that no partisan account should contain.
And the three framings are not equally available to their holders, which is the part that makes the arrangement stable.
The devotees' reading requires years of context and a technical vocabulary. It is not a conclusion anybody reaches by looking; it is a conclusion reached by having been taught what to call the things one is looking at, over a long period, by people who already held it.
The fisherman's requires none of that. He has one category for a body that transmits a state on contact, and he applies it in about four seconds. And he is not stupid — his description of the deformation is a clinical match for the devotees' own account in another chapter, which means he observed accurately and classified differently.
And the soldiers' is supplied to them by the subject.
Which produces the finding this page rests on. The three accounts are not competing hypotheses about one event; they are three different amounts of prior context applied to identical data, and no possible additional observation would move any of them, because the disagreement was never about what happened to the body.
The epilepsy sentence is not an isolated concession. It is one of three techniques the corpus records, and together they show something the text never states about itself.5
Self-limitation. He stops the Nṛsiṁha mood mid-street, in the middle of an involuntary state, because "causing fear among people is an offense."
Concealment. "The Lord did not always manifest His ecstasy. Being afraid of a great assembly of people, the Lord kept His ecstasy concealed."
Renaming. Due to epilepsy, I sometimes fall unconscious.
Suppress, hide, relabel. Three modes, one problem: what the ecstasy looks like to people who do not share the frame.
Which establishes something the corpus's own doctrine of involuntary transport does not have room for. He is managing the presentation of a state he is described as not controlling — choosing when it shows, to whom, and under what name — and the management is competent, situational and repeated.
The corpus supplies all three instances and never assembles them, which is why the finding is available to a reader and was not available to the participants.
Two readings of the sentence should also be ruled out, because both are tempting and neither survives the text.
It is not a confession. Nothing in the passage suggests he privately believes the naturalistic account; he says it to soldiers, under duress, about a matter of four lives. A statement made in that setting is evidence about the setting.
And it is not a lie in the ordinary sense either. Epilepsy is a description of the observable event that is accurate as far as observation goes — the collapse, the foam, the unconsciousness are all real and all present. What he has done is supply the correct account of the symptoms while withholding the account of the cause.
Which is the ordinary form of every such statement anybody has ever made on a form. It is why the sentence is uncomfortable rather than damning: there is no false proposition in it.
The corpus's own criterion would clear it without difficulty. The direction of benefit is entirely away from himself — four men were about to be executed — and by this tradition's own test, that settles it.6
Layer A gives the collapse, the arrest, the brāhmaṇa's defence, and the sentence in the first person.7
Layer B does the least intrusive thing available: it reads the episode as the Lord protecting his devotees, which is what the text says and requires no addition.8
Worth marking as a place where the commentary declines an opening. A purport tradition with an interest in the ecstasy's evidential value had every reason to gloss the epilepsy sentence — to explain that he spoke in the manner of ordinary men, or that the word was used loosely — and it does not.
The sentence is left exactly as the verse has it. That is a point in Layer B's favour, and this build has recorded few, and it is worth recording where it is earned.
You have a condition that does not fit the available categories. You will use the box word, and it is not dishonest. What matters is knowing which conversations get the accurate account and which get the processable one, and not confusing the two in your own head.
You are watching somebody be assessed by people who do not share their frame. The assessment will be made in the assessors' vocabulary whatever anybody does, and the only question is whether the person gets to choose which of the available words gets used.
You are collecting evidence about somebody's inner state. Notice who is doing the naming. Three parties named this condition three things, and each name is accurate to what its speaker could observe from where they stood — which means the disagreement is not about the facts and cannot be settled by more of them.
Somebody manages when their condition shows. That is a capability, and it sits awkwardly beside any account of the condition as involuntary. Both can be true — most involuntary states have a controllable margin — and the margin is where all the social information is.
The evidence is a short, high-stakes scene with the crucial sentence spoken by its subject in the first person, and two independent parallel framings elsewhere in the corpus.9
The tension the source doesn't close: the condition is involuntary transport and its presentation is managed. Suppressed in one chapter, concealed in another, renamed in a third — and the corpus's doctrine has no place for a state that can be timed.
A second one. Three parties name one condition three things, and the corpus adjudicates none of them. A partisan text has preserved the two readings that damage it.
Open: what he privately held. The sentence is spoken under duress about four men's lives, and nothing in the range read reports a view expressed in any other setting.
This page and The Fisherman and the Ghost hold two of the three framings, and the pair is the ingest's strongest evidential material. One outsider reads the body as possessed; its own subject names it a disease; the community reads it as divine. No party is lying and no party can check the others.
Against The Nṛsiṁha Mood and the Fear Offence it is the same capability in a different mode. There an involuntary state is halted on an ethical judgement; here it is relabelled on a tactical one. Both require a controlling margin the doctrine does not describe.
And it sits directly beside The Pāṭhān Conversion from the Koran — the same chapter, the same road, the same soldiers. One episode has him arguing a scripture he does not hold; the other has him naming a disease he does not have, and both are done to keep other people safe.
Emotional versus Narrative Memory — that page separates what a person can recount about an experience from what the experience actually was: the narrative is revisable, communicable and shaped by the frame it is assembled in; the experience is not, and the two can come apart completely without anybody lying.
The three framings are that separation at its most extreme, and the pairing shows why no amount of further evidence would settle them.
All three parties are giving accurate narratives of what they could observe. The devotees observe a state they have technical vocabulary for and years of context on. The fisherman observes a deformed body that transmits something on contact, and has one available category for that. The soldiers observe a collapse, and are given the category that fits.
None of them has access to the experience, including — and this is the part the corpus's own material forces — including, at the moment of speaking, its subject, who is describing an episode he was unconscious for.
Which gives the finding neither page reaches. Where a state is unavailable to introspection, every account of it is a narrative constructed by somebody from outside, including the subject's own — and the disagreement between accounts is therefore not resolvable by asking anybody, however sincere.
That is why this corpus's decision to record all three is worth more than its doctrine. A tradition that preserved only the flattering framing would have left less evidence than one that preserved the fisherman.
Sharpest implication. A partisan text has recorded its central evidence being named a disease by its own subject and a possession by a fisherman, and adjudicates neither. That is not a weakness in the corpus; it is the single most useful thing in it — because a source that preserves the readings that damage it is a source whose other reports become worth more, not less.
Generative questions.