There are people who are very good at a thing, and there are people around whom other people become good at that thing, and everybody has met both and almost nobody has a word for the difference.
The first is legible. You can watch them work, you can measure the output, and their excellence is located in them — it goes home with them at night.
The second is stranger and harder to talk about, because nothing about them is impressive on inspection. They are often not the best in the room. But the room is different when they are in it, and it stays different for a while after they leave, and the effect is not something they appear to be doing.
The usual response to this is mystification — presence, energy, some word that closes the conversation. Which is a shame, because the difference is not mysterious. It is a difference in what you measure. One is a property of a person and the other is a property of what happens near them, and only the second can be seen from outside.
Madhya 16 states the second as doctrine, and it is the corpus's own answer to its worst problem.
The villagers of Kulīna-grāma ask Caitanya the same question in successive years: what is the duty of a Vaiṣṇava, and who is a first-class one?1
Year one, the answer is a practice.
"A person who is always chanting the holy name of the Lord is to be considered a first-class Vaiṣṇava."2
That is a perfectly good answer and it is where most traditions stop. It is checkable in the weak sense — you can see whether somebody chants — and it locates the quality in the person's own conduct.
Year two, asked again, the answer moves.
"A first-class Vaiṣṇava is he whose very presence makes others chant the holy name of Kṛṣṇa."3
Read the shift. The criterion has left the person entirely. It is no longer about what he does; it is about what happens to other people standing near him.
And from that the three grades follow: Vaiṣṇava, Vaiṣṇavatara, Vaiṣṇavatama — good, better, best, sorted by transmissible effect.
To see the weight of it you have to hold it against everything else this corpus uses to decide anything.
Eligibility rests on an interior state "He alone, and no one else, can understand." Authenticity is settled by an expert declaring genuine-looking symptoms to be a mere semblance. Efficacy is tested such that an absent result proves the practitioner's offence and never the method's limit.4
Three decisive tests, none of which can return a negative, and all of which are adjudicated by somebody's declaration about an invisible thing.
And here is a criterion that is none of that. Do others chant when he is present? is observable by anybody, checkable without a certified interpreter, and capable of coming out negative. You do not need to be told; you can stand in the room.
This is the corpus supplying, in one sentence, the exact instrument its own epistemology otherwise lacks — and it is the corpus's own counter-position, not an outsider's objection. That is what makes the sentence more important than its length suggests.
And then it is not used.
The disciplinary chapters decide their hardest cases on interior states nobody can inspect: a man of impeccable conduct refused because "within his heart he was an impersonalist"; a man of outrageous conduct exempted because "his body is not material."5 Neither ruling asks the Kulīna-grāma question. Neither asks what happens to people standing near either man.
Nor is it used where it would have been decisive. Madhya 25 concludes that the six philosophical systems cannot reach the Absolute Truth and that one must follow the mahājanas — which is unusable without a criterion for identifying one.6 The corpus has a criterion. It is nine chapters away and it is not deployed.
The corpus's one checkable test sits unused beside its most consequential judgements, and the honest reading is not that it was suppressed. It is that a criterion which distributes assessment to everybody in the room is of no use to a system that needs particular people to be able to settle particular questions.
An observable test does not produce authority. It produces agreement, which is a different and less governable thing.
And the corpus does use the mechanism — just never as an assessment.
When Vallabha Bhaṭṭa refuses to concede, nobody defeats him. The Lord withdraws his attention, and the town stops listening.10 That is the effect criterion operating at full strength and in reverse: presence produces chanting, and removal of presence produces silence, and everybody in the community can see both.
So the corpus knows exactly how the mechanism works. It deploys it as an instrument of enforcement and declines to adopt it as a standard of assessment — which are opposite uses of one fact. Enforcement keeps the effect in the hands of the person generating it. Assessment hands it to everybody watching.
That is the clearest evidence available that the choice was a choice. A tradition that had simply not thought of the criterion would not be able to run it backwards.
The delivery is as interesting as the doctrine, and the corpus does not flag it.
They ask the same question in successive years and get a different answer. Not a corrected answer — the first is never withdrawn — but a deeper one, given to people who have spent a year with the first.
Which is a specific pedagogical claim: the answer a question can carry depends on how long the asker has been holding it. The year-one answer is not a simplification for beginners. It is the true answer at the resolution those hearers could act on, and the year-two answer would have been inert if delivered first, because make others chant is not something anybody can attempt.
That is also why the second cannot be turned into an instruction. You can be told to chant. You cannot be told to be someone whose presence makes others chant — the moment it becomes a target it becomes a performance, and a performance of that particular quality is exactly what it is not.
So the corpus has produced a criterion that is observable from outside and unusable from inside. It is a diagnostic and not a practice, and everything that goes wrong with it downstream comes from people trying to use it as the second.
One caution about the pedagogy, because it is easy to over-claim. This is the only place in the range read where the corpus is shown giving a deeper answer to a repeated question. Everywhere else its teaching scenes are single occasions — a day on a riverbank, a week of silence, one long instruction — and the year-gap structure appears once. That is a single instance, and a single instance is an observation rather than a method. It is described here as what happened at Kulīna-grāma and not as how this teacher taught.
Worth being clear about the price, because it is not obviously worth paying and the corpus may have declined it for reasons.
It is a popularity-shaped measure. Others chant near him rewards the charismatic, the socially warm and the well-placed, and it will systematically underrate the solitary practitioner, the difficult teacher, and anybody working with a hostile or indifferent group. A tradition running on this alone would promote its most affable members.
And it is defeasible by circumstance. The same man produces the effect in a village of devotees and none at all in a hostile town — which the corpus itself demonstrates, at length, in the chapters where its protagonist's presence empties temples and fills streets depending entirely on where he is standing.
So the interior criterion is not simply a failure of nerve. An unobservable test is immune to circumstance and to charm, which are real problems, and the price of that immunity is that it cannot fail.
The corpus holds both and uses one. That is the finding, and it is worse than not having the second at all — because the existence of the better instrument is what makes the choice a choice.
Layer A supplies the two answers, the year's gap, and the three grades, compactly.7
Layer B expands the three grades into a taxonomy of qualification for initiating others and for institutional position.8
The direction is worth marking. The verse grades by an effect on bystanders. The commentary grades for a role. Those are different operations: one is a description of what is observed, the other is a selection criterion for office — and a criterion built for observation behaves badly the moment it is used for appointment, because the thing being measured becomes the thing being sought.
Nothing in Layer A makes the three grades an appointment scheme.
You are assessing somebody and you keep watching them. Watch the people near them instead. It is the only version of the question that produces an answer you can check, and it will frequently disagree with your assessment of the person, which is the point.
You want to become the kind of person whose presence changes a room. You cannot aim at this, and the aiming produces the counterfeit. What you can aim at is the year-one answer — the practice itself, done consistently — and the corpus's own ordering says the second arrives on top of the first, or not at all.
You are asked something and you give the accurate answer for where the asker is. Expect to be asked again, and expect the second answer to be different without the first being wrong. Most teaching failures are a right answer delivered at the wrong resolution, and the fix is a year rather than a rephrasing.
You have an observable measure and a private judgement that disagree. Notice which one gives you authority. The observable one distributes the assessment to everybody present; the private one keeps it. That difference, and not accuracy, decides which measure institutions actually adopt.
The evidence is a two-year exchange with the same interlocutors, stated plainly, with the escalation visible in the text rather than inferred.9
The tension the source doesn't close: the corpus has one checkable criterion and decides its hardest cases on uncheckable ones. Both are Layer A, nine chapters apart, and no passage in the range read acknowledges the choice.
A second one. Follow the mahājanas is offered as the conclusion of the whole Madhya-līlā and requires a way of identifying one. This criterion is exactly that and is not deployed there.
Open: whether the effect criterion survives a hostile audience. The corpus's own narrative supplies cases where presence produces nothing, and the criterion as stated has no clause for it.
This page is the direct counter-instrument to Rati-ābhāsa, and the two should be read as a pair. That doctrine says visible signs cannot be trusted and appoints an expert; this one makes a visible sign decisive and appoints nobody. They are the corpus's two answers to how would anybody know, and the tradition runs on the first.
Against Refusing the King the contrast is a judgement rather than a doctrine. A king conceded to be a great devotee is banned on category; nobody asks what happens to people standing near him. The corpus had an instrument that would have answered the question and used a gate instead.
And it supplies the missing measure for The Six Philosophies and the Mahājana Path, whose conclusion is unusable without one. The corpus's own criterion and the chapter that needs it never meet.
Best Closer Is Not Best Sales Manager — that page's principle is that personal excellence and the capacity to produce excellence in others are different skills that overlap less than anybody expects. Phil Jackson was not the best player on the Bulls; he was who the best player listened to. Promote the top closer and the team's output falls.
The Kulīna-grāma escalation is that distinction arrived at from the other end, and the pair produces something neither has alone.
The business version is a warning about a selection error. It assumes the two measures exist and says: do not use the first to fill a role that needs the second. Its unit is the organisation and its worry is misallocation.
The corpus's version makes the second measure a ranking of persons rather than a fit for a role. Vaiṣṇavatama is not a job. It is a claim about what somebody is, graded by what happens near them — which means the corpus has done what the business literature explicitly warns against, in reverse: it has taken an effect-on-others measure and made it an assessment of the individual.
Which gives the finding. An effect measure is reliable as a description and treacherous as a rank, and for the same reason in both directions: the effect belongs to a pairing rather than to a person. The closer's numbers travel with him; the manager's do not, and neither does the Vaiṣṇava's. Move any of them to a different room and the measurement changes without the person changing at all.
That is why the business literature keeps the measure attached to a role and the corpus does not — and it predicts exactly the failure mode a tradition using this criterion would develop, which is that its most highly graded members would be the ones who never had to work a difficult room.
Sharpest implication. A tradition with one checkable criterion and several unfalsifiable ones will use the unfalsifiable ones, and not from cowardice — an observable measure distributes assessment to everybody in the room, which is precisely what an authority structure cannot afford. The better instrument is left in the drawer for the same reason it is better.
Generative questions.