A monk steps into a small stone cell cut into a cliff face. Behind him, the entrance is walled up — actually bricked over, leaving a gap just large enough to pass a bowl through. He will be in there for years. Once a week, or once a month, an acolyte climbs to the gap and pushes food in. No conversation, no company, no daylight to speak of. He is, in Dukes's flat phrasing, completely a-social.1 What he is doing in the dark has a name: anatmabhavacarya — the practice of the experience of self-vacuity, the deliberate, sustained rehearsal of having no self.2 This was not punishment and not a stunt. It was the extreme end of a method the Yogacara schools built to study consciousness by removing nearly everything consciousness usually leans on.
This is solitary retreat taken to its structural limit. Dukes records that as a result of the need to both study and practice the profound principles discovered within the Yogacara schools, many schools developed practices enabling monks to live alone for a number of years in small, restricted areas.3 Some practitioners were walled up in caves for a number of years, their food brought by acolytes weekly or monthly.4 The discipline rehearsed inside was anatmabhavacarya — literally the carya (practice, conduct) of anatma-bhava (the state of no-self), the experiential cultivation of self-vacuity rather than its mere intellectual grasp.5
Two features grew especially strong in this stream, and this is the part with long consequences. There was a noticeably greater emphasis on self-confession of fault, and on the transference of one's merit to others.6 Both features were eventually incorporated into the liturgies of nearly all the different schools and sects.7 A practice forged in absolute isolation seeded two of the most widespread communal rites in later Buddhism.
Why wall someone up to study no-self? Because the ordinary sense of being a someone is propped up by a thousand small confirmations, and most of them are social. Other people call your name, react to your face, hold a mirror up to your story all day long. Remove the people, the conversation, the changing scenery, the schedule, even reliable daylight, and you have pulled out the props one by one. What is left is a consciousness with almost nothing to reflect off of — which is exactly the laboratory condition for testing whether there is a self in there at all, or only the running process that the props kept lit. Anatmabhavacarya is subtraction as method: not arguing the self away but starving it of everything it feeds on and watching what remains.
Now the strange turn. You would expect total isolation to breed total self-absorption — alone in the dark with nothing but your own mind. Instead this stream produced its opposite: an intensified turn outward. Self-confession of fault is the self refusing to defend itself, naming its own failures instead of guarding its image.8 Merit-transference is the self giving away the one thing it most wants to hoard — its spiritual credit, dedicated to others.9 Both are the no-self insight made into a gesture. If there is no fixed self to protect, then concealing fault is pointless and hoarding merit is incoherent. The walled-up monk, having loosened the self in the dark, performs its loosening: confesses what a defended ego would hide, gives away what a grasping ego would keep. The cave teaches no-self; confession and merit-transfer are no-self in action.
This page anchors the operational extreme of several vault structures. Yogacara Consciousness as Primary Reality gives the theory of mind this retreat was built to test in the body; the wall is where the eight-consciousness model went to be lived rather than argued. Death-in-Life Practice names the broader pattern of rehearsing ego-dissolution while alive; immurement is its most literal form — a tomb the practitioner walks out of. And Anatta supplies the doctrine; anatmabhavacarya is its dedicated training protocol, the gym where the doctrine becomes an experience.
Track the single most consequential claim in the passage: confession-of-fault and merit-transference, born in walled-up solitary practice, were eventually incorporated into the liturgies of nearly all the different schools and sects.10 This is a transmission story hiding in one sentence. Something done alone, by a few extremists, in stone cells, becomes something done together, by everyone, in temples. The mechanism is worth seeing. The walled-up monks were the prestige practitioners — the ones who had gone furthest into the Yogacara method. What the prestige practitioners did privately set the standard for what ordinary practice should aim at. So the gestures that expressed their no-self realization got formalized, then ritualized, then handed to the whole community as liturgy — confession recitals, merit-dedication verses closing a ceremony. The radical individual practice was domesticated into a shared rite. There is a quiet loss buried in the gain: in the cave, confession and merit-transfer were the spontaneous overflow of a loosened self; as liturgy, recited on schedule by all, they risk becoming exactly the kind of formalized rule that empties out — performed because the order says so, the original self-vacuity that made them meaningful nowhere in the room. The same passage that records the practice's spread quietly records the seed of its possible ossification.
You will not be bricked into a mountain. But you can borrow the structure. Pick a stretch of real solitude — a day, a weekend, longer if you can — and strip the props deliberately. No phone, no talking, no audience for your inner narration. Notice, in the first hours, how loud the self gets when nobody is reflecting it back: the rehearsed conversations, the imagined defenses, the running highlight reel. Do not fight it. Let the silence keep pulling the props out. Then, when the noise thins, try the two gestures the cave taught. First, confession: name one fault you normally hide even from yourself, plainly, with no excuse attached — feel the self decline to defend itself and notice it does not die. Second, transfer: take whatever small good came of this solitude and dedicate it, silently, to someone specific who is struggling — feel the self give away what it wanted to keep. The gap in the wall through which the food came is the image to hold: even sealed in, you stay porous to others. The point of going in alone was never to become more alone.
The source is a single endnote (Note 13) in a practitioner-scholar's appendix, compact and confident.11 [PLAUSIBLE — needs corroboration] The historical core is sound: walled-up long retreat is well attested across Tibetan and Chinese esoteric traditions, and confession (Skt. desana) and merit-dedication (parinamana) are genuinely near-universal in Buddhist liturgy. What Dukes asserts and does not prove is the causal claim — that these two rites entered nearly all schools specifically because of the walled-up Yogacara retreat stream. [CONTESTED] Both confession and merit-transfer have independent roots elsewhere in Mahayana devotion and the bodhisattva vow, and a critic could argue Dukes is reading a single tributary as the whole river. Preserved as a genuine tension: the practices and the retreat clearly co-existed in this stream; whether the retreat originated their liturgical spread, or merely intensified what was arising on several fronts, the note does not settle. Open question: did the gestures flow outward from the cave to the community, as Dukes implies, or were the walled-up monks simply performing, at maximum intensity, devotional moves already shared by the lay and monastic mainstream?
Dukes writes this from inside a living esoteric lineage — he notes plainly that yogic physical trainings were, and still are, used in Tibetan, Chinese, and Japanese esoteric monasteries12 — and his decisive sources are his own temple's privately-held manuscripts. So a one-sentence causal claim about liturgical origins ("both features were eventually incorporated into the liturgies of nearly all the different schools") carries the confidence of a practitioner who treats his tradition's internal account as settled, where an outside historian would want corroboration across schools that did not share his lineage. The tension is not that he is wrong — the practices and the retreat genuinely belong together — but that the strength of the causal arrow rests on insider conviction more than on cited cross-school evidence. Where he converges with the vault's critical-historical voices is the underlying insight both would endorse: that liturgy is sedimented practice, that today's recited rite was yesterday's lived breakthrough, and that the gap between them is where ossification creeps in. Dukes the practitioner and Samuel the historian would tell that sedimentation story in different registers, but neither would mistake the recited verse for the realization that first produced it.
Sensory Deprivation and Self-Dissolution — Modern sensory-deprivation research finds that when you remove a mind's inputs — float tanks, darkness, prolonged solitude — the sense of a bounded self frequently destabilizes, sometimes into peace, sometimes into hallucination and panic. The walled-up cave is sensory deprivation run for years, and it predicts both outcomes the lab finds: the failure-sign of isolation thickening the ego (the panic-and-hallucination branch) and the goal of self-vacuity (the dissolution branch). The structural parallel is exact — subtract the props, watch the self wobble — but the contemplative case adds something the deprivation literature lacks: a direction. The lab tends to treat self-dissolution as an effect to be measured; anatmabhavacarya treats it as a skill to be cultivated, with confession and merit-transfer as the techniques that steer the wobble toward release rather than terror. The joint insight neither domain reaches alone: psychology explains why immurement destabilizes the self (input-starvation), and the Buddhist frame explains why some emerge liberated and others broken — the difference is whether the practitioner had a method for what to do with the dissolving self, or only endured its dissolution.
Strategic Self-Disclosure / Confession — In the influence literature, confession is a lever: disclosing a fault disarms an opponent, builds trust, lowers the listener's guard, and can be deployed tactically to win compliance. Anatmabhavacarya's confession-of-fault is its mirror image and its inversion. Same outward act — naming your own fault aloud — opposite engine. Tactical confession is the self using disclosure to gain something; the cave's confession is the self declining to defend itself with nothing to gain, because there is held to be no fixed self worth defending. Setting them side by side exposes the tell that separates them: tactical confession always angles toward an advantage, contemplative confession toward release of advantage. The insight neither domain produces alone: behavioral-mechanics shows that confession is a tool that can be weaponized, and the Buddhist frame shows that the same gesture, drained of the self that would weaponize it, becomes liberation instead of leverage — which means the diagnostic for telling sincere confession from manipulation is precisely whether a self stands behind it angling for return.
Sharpest implication: the most radical solitude the tradition built did not turn the practitioner inward toward himself but outward toward others — confession that drops self-defense and merit-transfer that gives the self's treasure away — which suggests that going all the way into no-self comes out the far side as service, not withdrawal.
Generative questions: