Something is happening on a riverbank and nobody has been told. Three groups of practitioners, in three places, each notice it — and each notices a different thing.
The aghorīs, sitting in cremation grounds, feel a presence: "there is some presence … there is something which is happening surrounding us."1
The kaulas feel an energy rise: "there is a kuṇḍalinī getting raised … invited."2
And the nāthas notice it in their breathing: "the nāthas saw their inner breath reversing into the ⚠️[kāla]-kumbhaka."3
One event. Three detections. Each group registers it through the faculty its own practice trains — the aghorī through presence in a burning ground, the kaula through the coiled energy, the nātha through breath.
The kaula detection comes with an explanation of what makes it strange, and the explanation is the useful part:
"Normally when they … invite, they get into the … energy forms, or sit in front of the energy forms. Normally they raise their energies — kuṇḍalinī — they raise their energies from the kuṇḍalinī just to make sure they establish a proper connectivity with the ⚠️[devas]. This is what actually they were able to feel without any kind of invocation."2
So the ordinary sequence is: sit down, invoke, raise. Deliberately, with a procedure, toward a chosen contact.
What happened instead was the third step without the first two. "They were able to feel that uninvited … kuṇḍalinī energy is getting spiked up."3
That is a claim about causation rather than about intensity. The energy is described as arriving under somebody else's operation — a man in absorption on a riverbank some distance away — and reaching practitioners who had performed nothing.
The nātha detection is the one this page is named for, and one word in it is doing more work than the rest of the sentence:
"…the nāthas saw their inner breath reversing into the ⚠️[kāla]-kumbhaka. We call it kāla-kumbhaka — kumbhaka is actually a process. So in that process, normally what happen[s] is: it is a breathing process similar to actually vijñāna."3
Kumbhaka in ordinary usage is retention — the breath held still between an in-breath and an out-breath. Holding.
Reversing is not holding. The word implies that the inner breath has a direction and that the direction inverts, which is a claim about the two movements exchanging roles rather than about air being kept in the lungs.
⚠️ Nothing in the passage develops this, and this page does not build a mechanism on it. What is recorded is that the word chosen commits to more than retention does, and that the material does not appear to notice.
Then the classification, which is the reason the passage matters beyond the story:
"This is one of the particular svarūpas, or … in [the] 112 ways. This is one way of it — one way of the dhāraṇā which we do it, kumbhaka."4
A dhāraṇā here is one of the hundred and twelve practices of the Vijñāna Bhairava Tantra — each a distinct doorway into the absolute.
This is the only place in thirty-four sources where a member of the 112 is named. Everywhere else the number is a total: there are a hundred and twelve ways, a hundred and eight disclosed, four withheld. Here one of them is identified, given a name, and attached to a lineage.
And the register is right, which is worth crediting. A dhāraṇā in that text is not a technique with an outcome; it is a place to be, an aperture in ordinary experience. The phrase used at the second mention is the stillness of the kumbhaka — not the practice of it — and that is the correct way to speak about one.
A gloss two minutes earlier governs everything that follows, and it is a self-correction made without prompting:
"…the Avadhūta … sat … in the nirvikalpa samādhi — unmovable, unbreathing, dissolved into the womb of the void. See — unbreathing is somewhat like breathing in a very, very subtle way."5
He says unbreathing, hears it, and corrects it in the same breath.
That matters for the whole page. Breathless in this material means attenuated, not arrested, and the correction is made once, about the samādhi, and governs the kumbhaka passages by proximity. ⚠️ No claim of breath cessation is carried here, and no physiological claim is made anywhere on this page.
And the frame is named too: these things belong "to the haṭha yoga sampradāya … when it comes to the nātha, haṭha yoga is part of it."6 Which is why the breath is the nāthas' instrument and not somebody else's — the detection follows the training.
The second statement comes twelve minutes later, at the initiation, and it adds the claim the state is famous for:
"He revealed [the] stillness of the ⚠️[kaula] kumbhaka — the breathless state beyond the death — which is part of the Vijñāna ⚠️[Bhairava] Tantra."7
Beyond the death is not decoration, and the episode's own architecture explains it. This is a recording about kāla, time. Its central image is a wheel of time — a ring of blue-black flame described as a cycle of dissolution and rebirth.8 Its central figure is defined against time.
In that frame, a state in which breathing stops being a movement is a state outside the cycle, because the breath is the most immediate thing that counts. It is the body's clock. What the wheel is at cosmic scale, the retention is at the scale of one body — the interval where the counter stops.
⚠️ The qualifier differs between the two statements: kāla at 7:45 and kaula at 21:45. Neither is reconstructed into the other, and the difference is not trivial. Kāla-kumbhaka would name retention as a suspension of time, which fits everything above. Kaula-kumbhaka would name it as the Kaula lineage's form of retention, which fits the passage's structure, where each lineage receives its own technique. Both readings are live and both are left standing.
The three detections are the sharpest thing in the range, and they are sharp because of what they predict.
Nobody chose how to notice. The aghorīs did not decide to sense a presence; the kaulas did not decide to feel the coiled energy; the nāthas did not decide to observe their breathing. Each group registered the event in the only channel it had.
And each channel is exactly what that lineage's practice builds. Aghora work is done in burning grounds, where what you train is the capacity to be somewhere and notice what is there. Kaula work runs on raising energy, so a rise is what registers. Nātha work is haṭha yoga, and haṭha yoga is breath — so the nāthas felt it in the one place they have spent years learning to read.
Now set that beside what happens later the same night. At the initiation, each lineage receives a disclosure in its own idiom — the kaulas get the coiled energy's route, the nāthas get the breathless state, the aghorīs get fearlessness in the presence of decay.9
The detection prefigures the disclosure exactly. What each group could feel beforehand is what each group was shown afterwards, and the passage never connects the two.
Which makes the whole night more coherent than it first looks, and also more circular. If a practitioner can only detect through what they have trained, and can only receive through the same channel, then the initiation delivers to each lineage the thing that lineage was already equipped to notice — and no crossing over is possible, because there is nothing in a nātha that could take delivery of an aghorī's disclosure.
⚠️ [n=2 SUGGESTIVE] — two moments, one episode. The decisive test would be a passage anywhere in which a practitioner detects or receives through a faculty their own lineage does not train.
There is no procedure here. The state is named twice, classified once, and never taught — kumbhaka is actually a process is the whole of the method given.
What the passage supports is a diagnostic and a caution.
Step one — find out what your own channel is. Whatever practice you have been doing for years has trained one faculty far past the others. The question is not what you believe but what you notice first — a shift in the room, a rise in the body, a change in the breathing. That is your channel, and on this account it is also the only one you can receive through.
Step two — expect it to fire uninvited, and do not read that as achievement. The whole point of the kaula detection is that it happened with no invocation and no procedure. Something moved and they registered it. Nothing in the passage suggests the practitioners had done anything to deserve it, and a rise you did not cause is not evidence about you.
Step three — do not take breathless literally. The correction is in the source: unbreathing is somewhat like breathing in a very, very subtle way. ⚠️ Any practice that aims at stopping the breath is aiming at something the material itself has ruled out of the description.
Step four — treat the classification as unusable until it is located. One of the 112 is a claim about which door this is. Without a verse there is no way to look it up, compare it, or check whether the identification is the speaker's or the tradition's.
⚠️ No count, no duration, no ratio, no posture and no marker is given anywhere in this range, and this page reproduces no breath practice.
BD 6:13–8:32 (the samādhi gloss, the three detections, the classification) and BD 21:45–22:08 (the second statement). [PARAPHRASED] — auto-generated captions.
Unreconstructed: the qualifier — kāla at 7:45 and kaula at 21:45, neither adopted over the other; Bhairava at 22:08 (as viganva, vigana); the category-word at 6:34 (as badis), in the clause placing this material in the haṭha yoga sampradāya; devas at 7:23.
Trust calibration. The reasoning here is narrative with a technical classification dropped into it, and the two parts have very different footing. The narrative — three lineages detecting one event — is transmitted and uncheckable. The classification is different in kind: this is one of the 112 dhāraṇās is the single most checkable claim this material makes about the text it names most often, because a verse number would settle it. 🚩 The weakest link is that the one checkable claim is delivered in the one register that makes it uncheckable. No verse is given here or anywhere in thirty-four sources, so a reader has to decide on the speaker's general reliability rather than on this claim's merits — and the identification is probably right, which is precisely what makes the omission frustrating rather than damning.
🚩 SINGLE SOURCE. 🚩 NO LOCATOR for the Vijñāna Bhairava, here or anywhere.
Density dropped from high to medium on 2026-08-25. The cited ranges were read in full and total about two and a half minutes — the kumbhaka is named twice, in passing, inside an episode about something else, and no procedure exists to carry. Restored from inside them: the three simultaneous detections at 6:34 and 7:08, the uninvited contrast with the normal invocation sequence at 7:23, and the haṭha-yoga placement at 6:34 that explains why the breath is the nāthas' channel. The previous version carried the two kumbhaka statements without the scene that produces the first one.
Open. Kāla or kaula cannot be settled from the caption, and the two make it a different state.
Two accounts of what the retention is sit twelve minutes apart, and the qualifier is not the only thing that differs between them.
At 7:45 it is a symptom. Nobody is practising it. The nāthas notice their breath reversing, involuntarily, because something is happening elsewhere — and the state arrives as evidence of an external event rather than as an attainment.
At 21:45 it is a disclosure. It is what the nāthas are shown at the initiation, revealed deliberately by a figure who has chosen to reveal it, and described as the breathless state beyond the death.
Follow the first and the state comes over you when conditions elsewhere are right. Follow the second and it is transmitted to you by somebody who has it. Neither passage refers to the other, and the same state is a weather report in one and a gift in the other.
A second strain runs through the word reversing. Everything else in both passages describes stillness — the stillness of the kumbhaka, breathless, a stopping. Reversal is not stillness. A thing that has inverted its direction is moving, and moving in a specified way, which is a considerably more particular claim than a thing that has stopped.
The material uses both vocabularies about one state without registering the difference. And the one word that would resolve it — whether the reversal is the onset of the stillness or is itself what the stillness consists of — is never said.
Metsuke and Perceptual Attention — the swordsman's discipline of where to put the eyes, built on the claim that most failure is a failure of perception rather than of effort, and that narrow fixed focus creates a specific blind spot: everything outside the beam disappears. The training is to hold the whole field while staying responsive to any point in it.
The three detections are that finding with the practitioners separated out. Each lineage has a beam, and each one saw what fell inside it. The aghorī's beam is presence in a place; the kaula's is energy in a body; the nātha's is breath. One event, three beams, three reports — and none of the three is wrong, because none of them is reporting the event. They are reporting what their beam contains.
What the swordsman's account adds is the part the tradition does not have, and it is the sharper half. Metsuke treats the mode as selectable. Hard focus and soft focus are both trained, deliberately, and the skill being taught is knowing which one the moment requires. A practitioner who can only do one is described there as having a costly blind spot.
Nothing in this material allows that. A nātha detects through breath because a nātha is a nātha, and there is no suggestion anywhere that a practitioner might learn to detect through somebody else's channel, or should. The lineage supplies the beam and the beam is not adjustable.
Which turns the pair into a real question rather than an illustration. If perception is trainable and modes are selectable, then a lineage-locked channel is a limitation the tradition has not noticed it has. If the channel really is fixed, then the initiation that follows — where each lineage receives in its own idiom — is not generosity toward different temperaments. It is the only delivery that could have worked. The two accounts disagree about whether that is a design or a constraint, and the disagreement is checkable by anybody who has trained in two traditions.
Sharpest implication. The three lineages detected the same event and each felt only what its own practice had built the equipment for — and hours later each was shown a disclosure in exactly that idiom. So the training does not prepare a practitioner to receive; it decides what they are capable of receiving, and the night's apparent generosity in giving three groups three teachings turns out to be the only arrangement that could have delivered anything at all. Nothing here says that, and the two halves sit fourteen minutes apart in one recording.
Generative questions.