Every institution has one, and everybody in it knows who it is.
There is a rule, and it is enforced. And there is a person to whom it does not apply — not by written exception, not by any category that could be extended, but because they are them. The senior researcher who does the thing everybody else is fired for. The founder who keeps the practice the compliance memo abolished. The one person whose version of the awkward behaviour is understood by all present to be different in kind.
And when somebody junior asks whether they can do it too, the answer is not no, because X. The answer is no, because you are not them — and there is no further reason available, because the exemption was never grounded in anything that could be tested against a second case.
Antya-līlā 5 has that arrangement stated outright, in open assembly, by the person who issued it, with the bar on generalising made explicit.
Rāmānanda Rāya — the man who supplies the corpus's highest teaching, on a riverbank, over the course of one day1 — takes two beautiful young dancing girls to a solitary place in his garden.
He massages oil into their bodies with his own hand. He bathes them, dresses them, decorates them.2
And Caitanya says so in open assembly, unprompted, without euphemism: "he naturally sees and touches the private parts of their bodies… His mind is as steady as wood or stone."3
The justification offered is an identity claim. Rāmānanda holds himself to be a maidservant of the gopīs; on that self-understanding this is a woman handling women, and the category that would make it a breach does not apply to him.
The corpus does not soften any of this. It reports the conduct in full and defends it in full, which is unusual and is why the passage matters.
Then the sentence that makes it a doctrine rather than an anecdote.
*"The authority for such acts is the prerogative of Rāmānanda Rāya alone, for I can understand that his body is not material but has been completely transformed into a spiritual entity. He alone, and no one else, can understand the position of his mind."*4
Three moves, and each closes a door.
The exemption is named. It attaches to a person, not to a class of person, and no criterion is offered by which anybody could join the class — because there is no class.
The ground is an interior state. His body is not material is a claim nobody can check from outside.
And the bar on generalising is explicit. He alone, and no one else, can understand the position of his mind — which is not merely a statement that the exemption is narrow. It is a statement that nobody, including Rāmānanda's peers and including anyone who might later be in his position, can assess the ground on which it was granted.
The precedent is issued and voided in the same breath. That is unusual enough to be worth naming: most institutions grant the exemption and leave the generalisation problem to be discovered later. This one closes it in advance.
The passage's real force is that the same chapter supplies a second ruling, and the corpus's own record supplies a third.5
| Person | Conduct | Ruling |
|---|---|---|
| Junior Haridāsa | asked an elderly saintly woman for rice, on another's instruction, in another's name | permanent expulsion, no appeal, one year, suicide |
| Rāmānanda Rāya | bathes, oils and dresses two young dancing girls daily | his prerogative alone; his body is not material |
| Caitanya himself | — | "if I even hear the name of a woman, I feel changes in My mind and body" |
Three standards, three people, one question, within a few chapters — and the corpus never sets them side by side.
This build's earlier reading was that conduct is never the measure and the operative variable is an invisible interior state. That is right as far as it goes, and Madhya 11 corrects it: category decides who is eligible to be judged on interior state at all, and interior state decides within the category.6 A king cannot be assessed on his heart because his office disqualifies him.
Which places Rāmānanda precisely. He is inside the category of the eligible — an intimate, a teacher, one of the three-and-a-half — and within it his interior state is vouched for by the only party whose vouching counts.
There is a party to this arrangement the corpus does not consider at all, and the omission should be stated rather than absorbed.
Two young women are being oiled, bathed and dressed by a man, daily, in a solitary place, and the entire question the corpus addresses is what this means about his interior condition.
Their consent is not discussed. Their view is not recorded. Whether the arrangement was good, tolerable or unbearable for them is not a question the passage contains, because the frame is entirely about whether a devotee's conduct is compatible with his standing.
That is not a modern complaint imported onto a sixteenth-century text; it is a structural feature of the criterion itself. A test that runs on the actor's interior state has no place to put anybody else's experience. Direction-of-benefit asks whose senses are being served, and the answer the corpus gives is the gopīs' — which routes the entire question past the two people in the room.
The corpus's own material shows it can attend to this when it chooses. It records what a woman's singing voice did to Caitanya, what Mādhavīdevī's standing was, what the prostitute at Haridāsa's door said and did.7 Here it does not, and the reason is that the passage is a jurisprudential defence rather than a narrative.
Fairness requires the strongest available reading, and there is one.
A rule about conduct with women, in a renunciate order, is a rule about a failure mode. It exists because the ordinary practitioner cannot be trusted to distinguish service from desire in himself — which is a realistic assessment and the reason the rule is categorical rather than case-by-case.
If that is what the rule is for, then a person in whom the failure mode is genuinely absent is not an exception to the rule's purpose. He is outside its scope, and applying it to him is a category error rather than a fairness question.
The trouble is entirely evidential. The corpus has no way of establishing that the failure mode is absent except the say-so of one man about another, and it says as much: he alone, and no one else, can understand the position of his mind.
So the doctrine is coherent and unverifiable at once — which is the shape of this corpus's decisive tests everywhere, and here it is applied to license rather than to condemn.8
The same chapter supplies a second instance of the shape, and it is worth setting beside the first because the circularity is even cleaner.
Hearing or describing the rāsa-līlā is said to destroy hṛd-roga-kāma — the disease of lust in the heart — outright.13 Which would make the corpus's most sexually explicit material a cure rather than a hazard, and it is offered exactly that way.
And then it is hedged, immediately and necessarily: the effect holds for those already transcendentally situated.
Read the two together and the structure is identical to the prerogative. The remedy works on people who no longer need it; the exemption applies to a man whose state cannot be checked. In both cases the qualifying condition is the thing at issue, and in both the qualification is announced in the same breath as the claim.
That is not a lapse in one place. It is how this corpus protects a doctrine it means seriously from being taken up by people it does not mean it for — and the cost is that neither claim can ever be tested, by anybody, including by the person it is about.
Layer A gives the conduct, the defence, the prerogative sentence and the bar on generalising, all in Caitanya's own voice in open assembly.9
Layer B is where the discomfort shows. The purports around this material press the exclusivity hard — the point being made is overwhelmingly that nobody else may do this, with warnings against sahajiyā imitation and against inferring permission from the example.10
Worth marking the direction, because it is a rare case of the commentary being more anxious than the verse. Layer A states the exemption confidently and once. Layer B returns to the fence around it repeatedly.
That is intelligible: a twentieth-century movement in rapid expansion, publishing this passage to a mass readership, had a practical problem the sixteenth-century verse did not — the verse was addressed to people who knew Rāmānanda Rāya, and the purport is addressed to people who do not.
You are being told a rule does not apply to somebody. Ask what the rule is protecting against, and whether the claim is that this person lacks the failure mode or that they are simply important. Those are different exemptions and only the first has an argument behind it.
You are granting the exemption. Say whether it generalises, at the moment you grant it. This corpus did, in the same sentence, and that is the single most responsible feature of the passage — because an exemption whose scope is left open will be cited by somebody with a worse case and you will have no principled way to refuse.
You are the one exempted. Notice that the ground is unverifiable and that you are the only witness. That is not a reason to decline; it is a reason to expect that nobody can tell your case from a worse one, and to behave as though the distinction has to be made visible in some other way.
An arrangement is being assessed entirely on one participant's interior condition. Find the people whose experience the frame has no slot for. They are usually in the room and usually not mentioned, and their absence from the analysis is a property of the criterion rather than an oversight.
The evidence is a short passage in which the conduct, the defence and the limiting clause are all stated by the corpus's own authority in open assembly — the strongest register available.11
The tension the source doesn't close: three standards on one question within a few chapters, and the corpus treats them as consistent. The harshest enforcement, the widest exemption, and a confession of personal vulnerability, and no passage sets them together.
A second one. The exemption's ground is explicitly unverifiable, and the corpus's disciplinary practice elsewhere runs on the same kind of unverifiable ground in the opposite direction. The same epistemology licenses Rāmānanda and condemns Rāmadāsa Viśvāsa.12
Open: what the two young women thought. The frame has no slot for it, and nothing in the range read supplies one.
This page and The Expulsion of Junior Haridāsa are the two halves of the corpus's answer to one question, and they should never be read apart. One man is exempted by name for conduct the tradition treats as unthinkable; another is destroyed for an errand. The corpus supplies both within three chapters and does not compare them.
Against Rāmadāsa Viśvāsa the pairing completes the logic. There, flawless external conduct is refused on an invisible interior state; here, outrageous external conduct is licensed on one. Conduct is not the measure in either direction, and the same instrument produces both rulings.
And it is the limiting case of Kāma and Prema: The Direction Test. That doctrine says identical acts are opposite categories depending on direction of benefit, with no content criterion whatever. This passage is what that looks like applied to the hardest available case, by the only person entitled to apply it.
Authority and Institutional Override — that page's mechanism is the capacity of a sufficiently senior figure to set aside a rule in a particular case, and the way that capacity behaves differently from the rule it overrides: it is discretionary, unappealable, and — critically — it does not create a rule of its own.
The prerogative doctrine is that mechanism with an unusual feature attached, and the feature is what makes the comparison worth having.
Overrides normally leave their scope undefined. The senior figure sets the rule aside, nothing is said about when this may happen again, and the ambiguity is functional: it preserves the rule nominally while leaving discretion intact, and everybody downstream is left to guess.
This override defines its own scope, to zero. He alone, and no one else is a statement that the override generates no precedent at all — which is more honest than the usual arrangement and, in one specific way, worse.
Because an undefined override at least leaves an argument available. Somebody in a similar position can say you did it for him. A zero-scope override removes even that — it is not that your case is weaker, it is that no case can be built at all, because the ground was declared inaccessible when it was granted.
Which gives the finding neither page reaches. An exemption's honesty and its contestability move in opposite directions. The clearer an authority is that a waiver sets no precedent, the more completely it has insulated its own discretion from challenge — and the version that looks most scrupulous, because it refuses to let anybody trade on the example, is precisely the version nobody can ever appeal to.
Sharpest implication. An exemption grounded in something only its holder can verify, and explicitly declared to generate no precedent, is the most defensible-sounding and least challengeable arrangement an institution can produce. It closes the generalisation problem in advance — and in closing it, it removes the only lever anybody would have had.
Generative questions.