Two schools of monks stand in a Tibetan monastery courtyard near the end of the eighth century, and they have agreed on the rules of a fight. There will be no army. There will be a debate. Whichever side loses will admit its errors out loud, hand the winner a garland of flowers, and walk away in peace. That is the civilized contract. What actually happened is that after the verdict came down, one monk sliced his own flesh into pieces, two Tibetans crushed their own genitals, a Chinese teacher named Mem go set his head on fire and died of it, and the losing side started planning assassinations while accusing the winners of planning the same. Before the killing could begin, the defeated Chaan teacher hid his school's scriptures inside a rock so that if everyone died, the teaching would survive in the stone.1 This is the Great Debate. Dukes records it not as a footnote of doctrine but as the moment a whole lineage of Chinese meditation got legislated out of Tibet — and very nearly murdered out of it.
This page records the eighth-century contest, usually called the Council of Lhasa or the Samye Debate, in which the form of Buddhism that would take root in Tibet was decided by formal argument rather than gradual drift.2 The stakes were total. The winner's school would receive imperial sponsorship; the loser's would be cut off from patronage and, lacking it, fade. By the eighth century both Indian and Chinese Buddhism had penetrated Tibet about equally. Indian missionaries had been arriving since the mid-fifth century — the esoteric current through Padmasambhava, the orthodox Vinaya-based current through Santiraksita, who acted as a founding father and advisor to the Tibetan monarchs and came from an impeccable teaching lineage running back through Nagarjuna and Sariputra. Alongside this Indian Dharma, the Chinese Chaan (meditation) school had grown equally popular, with ordinations recorded at the same time as Santiraksita's pupils. Its main teacher was known as Hva San Mahayana.3
Santiraksita was alarmed. He considered the Chaan school heretical, and he knew China itself had already moved to suppress what it saw as radical deviations from its sponsored scholastic Chaan, exiling several prominent members. Shortly before he died, he sent word to his Indian pupil Kamalasila (circa 794) to come to Tibet and meet the chief Chaan teacher in open debate.4 So the contest is not Tibet versus China in any clean sense — it is an Indian orthodox lineage moving to expel a Chinese meditative lineage from a third country, with two rival Tibetan noble families standing behind the two sides, each angling for the throne.5
The two schools were sorted into two doors. The "sudden path" — Tun men p'ai in Chinese, Yugapad in Sanskrit — held that enlightenment arrives all at once, whole, not assembled piece by piece. The "gradual path" — Tsien men p'ai, Karamavrittya in Sanskrit — held that you climb to it by stages, building practice on practice.6 Kamalasila's school was the gradual. Hva San's was the sudden.
This is the same fault line that already runs through the vault's Bodhidharma material, and that continuity is the point. The sudden/gradual split did not originate at Samye; it is the Tibetan replay of the Chinese Northern/Southern division — the gradual Northern School of Shen Hsiu against the sudden Southern School descending from Hui Neng, the line the Japanese samurai later took up for its tactical and psychological benefits (see Bodhidharma: The Revisionist Biography).7 Dukes's deflating note on that earlier split — "in reality, there is little difference between them, and both names came after Hui Neng" — hangs over the Tibetan rerun like a warning. The two doors that men were willing to die in front of may have opened onto nearly the same room. The garland-of-flowers contract assumed the difference was small enough to settle with a bouquet. The suicides proved the participants did not experience it that way at all.
The word "Yugapad" carries the whole sudden claim in one term: simultaneity, everything-at-once. "Karamavrittya" carries the gradual one: a turning-of-actions, a sequence. Both sides walked in so confident of victory that they signed the loser's-forfeit clause without hesitation. Confidence on both sides is itself a clue — when each party is certain it will win a contest of ideas, the contest may be measuring something other than the ideas.
Here Dukes turns historian against his own story. The records of this period, he says, are very fragmented. We do not even know for sure whether the head Chaan teacher in Tibet was a Chinese man or a Tibetan who had trained in China and gone by his Chinese title — there were certainly both kinds of teacher present.8 Later accounts of the debate vary depending on which side's record you read, and in each one, that side claims to have won. The verdict, in other words, is not in the sources; only the partisan memory of the verdict survives, doubled and contradictory.9
And then the line that should make any reader sit up: the later Gelug-Pa school — the Dalai Lama's sect — is said to have considerably edited the various official accounts of these events.10 The winning tradition wrote the history of its own win, generations downstream, and the editing hand is visible. But Dukes adds the saving clause: several alternative records still exist. The Chinese teacher's instinct to hide texts in a rock turns out to have a documentary twin — the truth of the debate also got hidden, in alternative manuscripts the editors did not reach. This is the texture the vault keeps insisting on: a contested institutional event leaves not one clean record but a stack of mutually accusing ones, plus an editor's fingerprints, plus a few survivors in the cracks.
This page gives the vault its first fully-staged example of doctrine settled by tournament rather than by transmission. It hands Bodhidharma: The Revisionist Biography the downstream chapter that page only gestures at — the sudden/gradual split, named there, here gets a date, a courtyard, and a body count. It supplies Dzogchen as Preserved Chinese Chaan its precondition: the Chaan school had to be defeated and de-patronized before it could go underground and survive as a sub-current of the Nyingmapa. It feeds Wandering Acarya Transmission Mobility a hard case — what happens when the wandering teacher arrives and the host country holds a referendum on whether to keep him. And the editing claim plugs straight into the source-criticism spine running through Bodhidharma Dating Controversy, Source by Source: the winner edits the record is a transferable suspicion, applicable anywhere a single sect's account survived its rivals'.
Read the contract and the aftermath side by side and the whole episode opens up. The agreement was elegant: loser acknowledges his errors, presents a garland of flowers, departs in peace.11 It is a debating-society ritual, the kind of thing that assumes the disagreement is about positions a reasonable person could hold or drop. A flower is a soft object. You give it when the wager was never really existential.
Then the verdict lands, and the response is the opposite of soft. Nan sa mi cuts his flesh into pieces. Two Tibetans crush their genitals. Mem go burns his own head off. Members of the Chinese party decide to kill members of the Indian; the Chinese claim the Indians had already plotted the same.12 The gap between the contract and the conduct is the real datum. These men had agreed in advance that losing meant a bouquet and a polite exit — and when they lost, several of them destroyed their own bodies and the rest reached for weapons. That gap tells you the debate was never about which path to enlightenment was philosophically superior. It was about which lineage got to exist. The flower was a fiction everyone signed and no one believed. Dukes does not editorialize the contradiction; he lays the two facts next to each other and lets the seam show. The self-mutilations are also, grimly, on-theme — a sudden-path school, defeated, answers with the most sudden, total, all-at-once acts a body can perform.
You are reading two histories of the same event, one in each hand, and they disagree about who won. Resist the urge to average them. Instead, ask the question Dukes's method trains: who held the pen last? You notice that one account belongs to the school that went on to imperial power, the Gelug-Pa, and that this account has been "considerably edited." So you stop treating it as the baseline and start treating it as the prosecution's closing statement. You set it beside the alternative records — the ones the editors missed, the dPao gtsug p'ren ba manuscript with its detail of texts hidden in a rock — and you read the official version for what it needed to suppress, not for what it asserts. The suicides survive in the record because they were too vivid to scrub; you weight them heavily precisely because they are not flattering to anyone. Last, you hold the verdict itself as unknown. The sources tell you a debate happened, that both sides forfeited everything on it, that one current was driven underground after. They do not reliably tell you who the judge ruled for. You write "outcome contested; both sides claim victory" and you do not pretend to more. The history that survives is the history someone wanted; the history you can defend is the part no one would have invented.
Evidence base: Dukes's Note 42a, "Tibet, Bodhidharma, and Chaan," and the "Great Debate" section following it, which draw on Tibetan manuscript sources including the dPao gtsug p'ren ba and which Dukes explicitly cross-references to G. Tucci's work in the field — work he elsewhere complains the mainstream Bodhidharma scholarship ignored entirely.13 The event is historically attested in Tibetan and Chinese records; Dukes's framing of it as the suppression of a Chaan lineage descended from Bodhidharma is the load-bearing interpretive layer, and it serves his larger thesis.
[CONTESTED] on the verdict itself: the outcome is irrecoverable from partisan sources, and the dominant account was edited by the victorious Gelug-Pa. Standard Tibetan tradition holds that Kamalasila's gradual school won; Dukes preserves the dispute rather than ratifying that outcome, noting each side claims the win.14
Open seam: whether the chief Chaan teacher was Chinese or a China-trained Tibetan is unresolved in the sources, and Dukes leaves it unresolved.8 The identity of "Hva San" is genuinely uncertain — the Tibetan manuscripts apply the title indiscriminately to many teachers, so the name may not pick out one man at all.15
Preserved tension: Dukes reads the Chaan school as the original tradition Tibet preserved and China later lost, which makes the losers of the debate the carriers of authenticity. The orthodox Tibetan account makes the winners the carriers of correct Dharma. These two valuations of the same event are not reconciled here and should not be.
Dukes is not a neutral umpire of this fight, and it shows in a particular tilt. His sympathies sit with the defeated Chaan school — he calls the Tibetan Chaan adherents possible preservers of "the ancient original tradition that was later lost or forgotten in China," and he closes his survey of the modern academic histories with the line that if the new trend is to dissolve Bodhidharma, "be thankful we still have living monks as teachers."16 That is a practitioner-scholar defending a living lineage against an academy he regards as outsiders to its spirit. The tilt is honest but it is a tilt: a writer whose own temple claims descent from this transmission will read the suppression of that transmission as a loss of truth rather than a settlement of error. Set him beside Geoffrey Samuel and the contrast sharpens — Samuel would treat the debate as a data point about how royal patronage shapes which religion survives, declining to crown either path as more authentic, where Dukes cannot help crowning the underdog. The convergence is real too: both men would agree the official record is an artifact of power, not a transcript of truth, and both would point at the Gelug-Pa editing as exhibit A. They diverge only on what to do with the doubt. Samuel rests in it. Dukes uses the doubt to make room for the tradition he already trusts. Where his warrant rests on Kongoryuji's privately-held lineage manuscripts — as it does whenever the chain back to Bodhidharma is asserted — no outside reader can audit the descent claim; that limitation rides along with this page wherever it touches the Akasagarbha-from-Hui-Neng succession.
Control the Framing, Win the Fight — the Great Debate is a textbook case of a contest decided less by the better argument than by who controlled the surrounding conditions, and the pairing exposes a mechanism neither domain sees alone. The behavioral-mechanics page describes how the party that sets the frame — the venue, the stakes, the rules of victory — often wins before the substantive exchange begins. Watch that mechanism operate here: Santiraksita's lineage set the terms (a debate, in Tibet, judged by the monarch, with imperial patronage as the prize) and imported a champion, Kamalasila, trained for exactly this. The Chinese Chaan school was made to defend itself on ground chosen by its opponent, with a verdict structure that allowed only one survivor. The history page supplies what the tactics page lacks — a real, lethal instance where framing determined which religion a country would practice for a thousand years, and a body count proving the participants knew the frame was the fight. The tactics page supplies what the history lacks — the reason the garland-of-flowers contract was strategically brilliant for the side that wrote it: a forfeit ritual that sounds magnanimous quietly encodes total defeat for the loser, dressing annihilation as etiquette. Together they yield a rule the history domain can carry forward: when a doctrinal dispute is settled by a single winner-take-all event, read the event's design first and the doctrines second, because the side that built the contest usually built it to win.
Cognitive Dissonance and Commitment Escalation — the self-mutilations after the verdict are a psychology problem the history page can date and locate. The psychology page describes how people who have sunk their whole identity into a position cannot simply absorb being wrong; the dissonance between "I staked everything on this" and "this lost" has to discharge somewhere, and the more public and irreversible the prior commitment, the more violent the discharge. That is precisely the shape of the aftermath. These monks had signed a contract making the debate existential, then lost. A man does not burn his own head off over a position he held lightly. The history page gives the psychology page a brutal natural experiment — commitment escalated to the maximum (lifelong monastic vows, public wager, lineage identity) producing self-destruction rather than belief-revision the instant the verdict went against it. The psychology page gives the history page a mechanism for why the aftermath looked like this rather than like quiet defeat: the suicides are not religious exotica, they are dissonance with nowhere to go. Together they suggest the diagnostic for any winner-take-all ideological contest — measure the violence of the losers' response and you have measured how completely they had fused self with stance going in.
Sharpest implication: if the verdict of the most consequential debate in Tibetan Buddhist history is unrecoverable because the winners edited the record, then "who won the Samye debate" is the wrong question — the answerable question is "what did each surviving account need to be true," and the editing itself becomes the primary evidence.
Generative questions: