You are properly angry, and you are in the right, and you have been building this sentence for six minutes.
And then you catch your own reflection in the dark window behind them — or you see the other person's shoulders come up — and something in you goes oh. Not that you were wrong. That you are frightening somebody, which was never any part of the plan and is happening anyway.
Almost everyone has that moment. And what makes it strange is that it does not come from calming down. You are still angry when it lands. Something else is running alongside the anger, and it has read the room, and it has overruled you.
Ādi 17 has that happen to Caitanya in the middle of a street, in the middle of an ecstatic state, and gives it a name.
He is in the mood of Nṛsiṁha — the man-lion, the tearing incarnation, the tradition's most terrifying divine form. He is running through the streets with a club in his hand.
People flee.
And he stops. The reason he gives is a single sentence: causing fear among people is an offence.1
That's the whole episode, and it is four lines long. It is also, structurally, one of the strangest things in the book.
Read it against the corpus's own account of these states and a problem opens immediately.
The same chapter-cluster has him say of his dancing: "Please do not think that I intentionally do it. I do it automatically."2 The whole evidential weight of the ecstatic material rests on it being involuntary — a man who chooses to dance has shown enthusiasm; a man who is made to dance has shown that something is acting on him. Voluntary ecstasy is, elsewhere in this corpus, an offence in itself.3
Here the state is stopped, deliberately, on a stated ground.
So the involuntariness has a governor. The thing that is not chosen can nevertheless be halted, by something that evaluates its effects and acts on the evaluation. The corpus asserts both, four chapters apart, and never places them together.
There are ways to reconcile it — the mood might be involuntary while the acting-out is not; the halting might itself be attributed to divine arrangement. The corpus offers neither. It simply records both claims in the same confident register.
The reason he gives is not what a reader of this corpus would predict.
The predictable reasons would be theological: the mood was misdirected, or the display was inappropriate to the setting, or majesty was being shown where sweetness belonged. Any of those would fit the tradition's usual grammar.
The reason given is the effect on bystanders. People were frightened, frightening people is an offence, so it stops.
That is an ethical judgement about ordinary third parties, made in the middle of an ecstatic state, by a figure whose community would have taken any amount of terrifying behaviour as evidence of divinity. Nobody in the street was going to file a complaint. The correction is entirely self-administered and entirely other-regarding.
For a corpus whose disciplinary chapters are famously severe about interior states and famously indifferent to consequences, this is a rare instance of the reverse: a rule grounded in what happened to people rather than in what was in somebody's heart.
Take the mechanism seriously and it is more interesting than the theology.
The states this corpus describes are, on its own account, absorbing. That is the point of them — attention is wholly occupied, the body does things the person is not directing, the surroundings drop away. He mistakes the sea for a river and a hill for Govardhana; a fisherman handling him catches the state like a contagion.4 Absorption is the register.
An absorbed state that can be interrupted by a moral observation is a compound thing. It requires something to be still running that is not absorbed: a channel that keeps monitoring the room, keeps registering other people's faces, and retains the authority to stop the whole apparatus.
The corpus does not theorise that channel and its practice, in places, works against it — the diagnostic signs of attainment are precisely the signs of absorption, and a devotee who stayed monitorial would show fewer of them.5
But it records the channel operating, once, cleanly, in a street in Navadvīpa. And the thing that got through the absorption was not a doctrine or a rule from a manual. It was a group of frightened strangers.
Notice what he was holding.
A club. Not a symbol, not a posture — an object, in his hand, while running toward people.
This is the corpus's most militant register arriving in a body, in public, with an implement. The book elsewhere keeps its martial vocabulary safely abstract: soldiers who are associates, weapons that are limbs, an enemy who is a category with no address.6 For four hundred pages nothing in that frame ever reaches a person.
Here it is four seconds from reaching several. And what stops it is not the doctrine, which had already licensed the frame, but a man noticing faces.
Layer A gives the episode in four lines and does not editorialise. It sits among the chapter's other youth narratives without special weight.
Layer B has a natural use for it and takes it: the purports read the episode as instruction in propriety — a devotee should not behave in ways that disturb ordinary people, and the Lord demonstrates the standard.7 That is a fair application and the verse supports it.
What thins out is the oddity. Read as an instruction in propriety, the episode is unremarkable. Read as what it is — an involuntary state halted by an ethical judgement, in a corpus that insists such states are not chosen — it is a small crack in the book's account of itself, and the commentary's framing seals it.
You are in the middle of being right and you notice the other person has gone still. That noticing is the whole content of this page. It does not require you to stop being right. It requires you to decide whether the frightening is a price you're willing to have charged to someone else.
You're in a state you value — absorbed, working, lifted — and something in the room needs your attention. Notice that the pull to protect the state is real and often correct. Notice also that a state which cannot be interrupted by another person's distress is not a higher state; it is just a louder one.
You're in a group that reads intensity as evidence. Ask what the group would do with someone who intensified and also watched the room. Usually the answer is that they'd be read as less far gone, which is to say as less advanced — and that is a scoring system with a specific and predictable failure.
You catch yourself mid-escalation and the thing that stops you is a face rather than an argument. Trust it. Arguments arrive after you've committed; faces arrive on time.
The evidence is four lines: the mood, the club, the flight of bystanders, and the stated reason.8 It is thin, and the page says so — what raises it is not the volume of material but the fact that a single sentence contradicts a doctrine the corpus states repeatedly elsewhere.
The tension the source doesn't close: ecstatic states are involuntary and this one was stopped. Both are asserted plainly. The corpus offers no account of the governor.
A second one. The corpus's disciplinary logic is overwhelmingly about interior states and is famously untroubled by consequences — a man is expelled to his death for talking with a woman, and the ruling is defended on grounds no observer can check. Here a rule is grounded entirely in an observable effect on ordinary people. The corpus has a consequentialist sentence and does not notice it is out of key.
Open: the episode is not developed. Nothing in the range read returns to causing fear among people is an offence, applies it to another case, or lists it among the offences the tradition enumerates formally. Whether it is a considered principle or a remark is not something the text settles.
The direct collision is with I Do Not Dance by My Own Will, and it should be read as one. That page has the ecstatic behaviour disclaimed as automatic; this one has it halted on a judgement. A corpus cannot have both without a theory of what is doing the halting, and it does not have one.
It pairs naturally with Devotees Angry at Their Own Ecstasy, which is the same phenomenon from the other side. There, ecstasy is resented because it interrupts service — the feeling is treated as an obstacle to the work. Here, ecstasy is stopped because it harms bystanders. Both pages record a devotional tradition ranking something above its own most prized states, and in both cases the something is other people. That convergence, across two chapters that have nothing else in common, is worth more than either instance.
Against The Army of Associates this is the one moment the martial register acquires a body and an implement, and the only moment it is called off.
Observer Self — Deikman's observer is the part that watches the argument while the argument is happening: chest tight, voice louder than intended, and something registering all of it, never itself observable. Whitfield treats it as the clinical apex — a capacity that only becomes stable after substantial healing has already happened lower down.
That last clause is the useful one, and it turns the Nṛsiṁha episode into a real question rather than an edifying anecdote.
The corpus has no developmental account of the governor. It records a man in maximal absorption, holding a club, who registers frightened faces and stops. On Deikman's and Whitfield's account that capacity is not free and not early: it is what remains available after a great deal of other work, and its instability under high arousal is precisely what makes most people unable to use it in the moment they need it. Almost nobody can observe themselves at full intensity. That is not a moral failing; it is a fact about the equipment.
So the two frames make opposite predictions about what the episode demonstrates. The devotional frame reads it as propriety — the Lord modelling correct behaviour. The clinical frame reads it as the hardest thing in the passage: an observer channel that stayed online at an arousal level where observer channels do not normally stay online, which is the single most impressive capacity anywhere in the chapter and is reported in four lines with no comment.
And the handshake gives a test neither has alone. If the observing capacity is developmental, it should improve — and if it is a property of the person's identity, it shouldn't need to. The corpus's later chapters supply the data and it does not read well for the second reading: in the Antya-līlā the same figure's absorption becomes total, he does not recognise the sea, his companions have to physically restrain him, and nobody reports the governor operating at all.9 Whatever was running in that street was not stable, and the tradition never asks where it went.
Sharpest implication. The measure of an intense state is not how far it takes you but whether anything can still interrupt it — and a tradition that grades attainment by depth of absorption is grading the one variable that predicts the governor failing. This corpus records the governor working once, in its protagonist's youth, and never again in the material read.
Generative questions.