The Tibetan empire bought a religion (and it outlived the empire, 775)

Tibet before the monasteries

The strongest power on the Inner Asian frontier

In the winter of 763 a Tibetan army rode into Chang'an. The Tang emperor Daizong had already fled east; the Tibetan commanders installed a puppet emperor, held the largest city in the world for roughly a fortnight, stripped it, and withdrew under no compulsion. Christopher Beckwith's history of the empire treats the episode not as a border raid but as the demonstration of what the Tibetan state had become in a century and a half — the dominant military power of Inner Asia, contesting the Silk Road oases simultaneously with the Tang, the Turkic khaganates and the Abbasid caliphate.³ At its greatest extent the empire of the Yarlung dynasty held the Tarim basin garrisons, the Hexi corridor including Dunhuang, tribute relationships in Nepal and the Himalayan foothills, and a frontier that reached the Pamirs.

This is the fact that determines everything else in this record. The Tibet that took delivery of Indian Buddhism in the second half of the eighth century was not a periphery being civilised. It was a conquering state with a surplus, choosing what to buy. Matthew Kapstein's account of the assimilation makes the point structurally: conversion in Tibet was a matter of imperial policy argued between factions at court, not of a small culture being overwhelmed by a large one.¹ The emperor who made the decision, Khri Srong lde brtsan — Trisong Detsen in the conventional English rendering — came to the throne in 755 at the age of thirteen and reigned until about 797. The sack of Chang'an fell in his eighth regnal year. The consecration of Samye fell in his twenty-fourth.

A state that already had writing

Tibet in 750 was already a literate administrative state and had been for roughly a century. The script had been adapted from an Indic model in the reign of Srong btsan sgam po (d. 649), traditionally by the minister Thon mi Sam bho ṭa, and by the eighth century it was carrying a working chancery: a year-by-year record of the court's movements, campaigns, censuses and oath ceremonies. That record survives. The Old Tibetan Annals, recovered from the sealed library cave at Dunhuang and edited by Brandon Dotson, is Tibet's first history, and it is an administrative document rather than a religious one — it notes where the summer council met, which minister was appointed, which region was surveyed for tax.⁷ The empire also legislated in stone. Hugh Richardson's corpus of the imperial pillar inscriptions preserves treaties, land grants, edicts and oaths cut into monoliths that still stand.⁶

So the Tibetans did not receive writing from Buddhism, and they did not receive record-keeping, law, or centralised administration. What they had was a script optimised for Tibetan, an aristocratic clan politics of sworn alliance, and a religion of the throne. What they did not have was a monastery, an ordination lineage, a translated scripture, an abstract vocabulary for mind and causation, or any institution that could hold property in its own name across generations.

The machinery that was in place is worth listing, because the standard account of this transmission tends to hand Buddhism the credit for all of it. By the middle of the eighth century the Tibetan state operated a periodic census of households and pasture, a graded system of military units drawn from those households, a penal and compensatory law code administered by the great ministers, a postal relay along the imperial roads, and a summer council at which the emperor and the clan heads renewed the arrangements of the previous year. The Annals record these operations in a bureaucratic shorthand so laconic that Dotson's edition needs a commentary the length of the text to make it legible.⁷ The society described was stratified — an aristocracy of clans, a large body of taxable commoners, and a servile population — and it was expansionist by design, since the surplus that fed the court came substantially from campaign, tribute and the tolls of the Silk Road oases the empire held.

The religion of the tomb

The royal religion of pre-Buddhist Tibet is difficult to describe because its practitioners left no doctrinal literature and its later opponents wrote most of the surviving accounts. What can be reconstructed comes from the Dunhuang manuscripts, the pillar inscriptions, and Chinese observation. Ariane Macdonald's long study of the Pelliot tibétain documents — a foundational and much-argued-with reconstruction of the royal cult — assembled the evidence for a coherent imperial religion centred on the sacred person of the btsan po, the emperor, whose descent from the sky and whose ritual security guaranteed the state's fortune.¹⁰ Rolf Stein's synthesis remains the standard general account of the ritual world it belonged to.⁹

The visible core of that religion was funerary. The kings were buried in earth mounds in the Yarlung valley at 'Phyong rgyas, in tumuli still visible today, and the burial was not a single event but a programme running for years. It involved:

• Animal sacrifice on a large scale — horses, yaks and sheep killed at the tomb, with the animals conceived as conveyances and companions for the dead
• Retainer burial, in which chosen companions of the dead king were assigned to accompany him; the practice is attested into the period around 800
• A class of ritual specialists — the gshen and the bon (a title for a function, not yet the name of a religion) — who performed the divination, the sacrifice and the recitation of the origin myths
• Sworn oath ceremonies, sealed by sacrifice, that bound the aristocratic clans to the throne and were the actual constitutional mechanism of the empire
• A cosmology of local powers — mountain gods, lake gods, earth lords and the class of beings later fixed in Tibetan as sa bdag, klu and btsan — whose good will was a practical requirement for building anything

Two of these deserve more than a line. The oath ceremony was the empire's constitution in the strict sense: there was no other instrument that made a clan chief in Zhangzhung obey a court in the Yarlung valley. Tang observers described the Tibetan rite in terms their own ritual vocabulary found shocking — a periodic sacrifice of animals, and at longer intervals a greater ceremony, at which the parties swore before the gods, the mountains and the emperor's person. The point of the blood was jurisdictional: it made the oath enforceable by powers that could reach a man anywhere on the plateau. And the tomb programme was the state's largest public works project, absorbing labour, livestock and precious goods on a scale nothing else in Tibet matched. To end it was not a liturgical adjustment. It was the retirement of a national industry.

Michael Walter's study of the political and religious culture of the early empire argues that this complex was not a folk substrate underneath the state but the state's own operating theology; the emperor's ritual status was the mechanism by which clans that had no reason to obey each other obeyed him.¹⁶ That is the thing Buddhism arrived to replace. Samten Karmay's studies of the myths, rituals and beliefs of this period repeatedly make the same point from the other direction: what looks like Tibetan folk religion in later centuries is very often the wreckage of a displaced state cult.¹³

The road from Nālandā to the Yarlung valley

Śāntarakṣita, and the epidemic that sent him home

The first specialist the court brought in was a scholastic. Śāntarakṣita was a Madhyamaka philosopher associated with Nālandā, the great monastic university on the Gangetic plain, and author of the Tattvasaṃgraha — a systematic doxographic survey of Indian philosophical positions written to refute them. He was, in other words, exactly the sort of figure a state acquires when it wants an institution rather than a devotion. David Snellgrove's history of Indo-Tibetan Buddhism describes the transaction in institutional terms: what Tibet was importing from the mahāvihāras was a complete package — Vinaya, ordination lineage, curriculum, and a scholastic method for adjudicating disputes.¹²

The earliest narrative source for what happened next is the dBa' bzhed, the "Testament of the dBa' clan," a manuscript of thirty-one folios that surfaced in Lhasa in 1997 and was published in translation by Pasang Wangdu and Hildegard Diemberger in 2000. It preserves the oldest recoverable version of the story the Tibetan court told about its own conversion.⁵ In that account Śāntarakṣita's first stay ends badly. Disease and disaster follow his arrival — the tradition specifies epidemic and a lightning strike on the palace at Marpori — and the anti-Buddhist party at court reads the omens correctly, by their own logic: the local gods are objecting. Śāntarakṣita is sent back to Nepal. Before going, he tells the emperor to send for a tantric specialist.

That structure is worth pausing on, because it is not a piece of hagiographic decoration. It is a precise statement of a real problem. Indian scholastic Buddhism had no answer to a mountain god. It could explain that the god was a deluded being trapped in saṃsāra, which was true within the system and useless at the construction site. The tantric ritual tradition could do something else: it could bind the god by oath and put it to work.

Padmasambhava: what the earliest documents actually say

Padmasambhava is the most famous person in this record and the least documented. The later Tibetan tradition knows him in extraordinary detail — eight manifestations, a lotus birth in the lake of Dhanakośa, a career of subjugating the deities of the entire plateau one by one, and a body of concealed "treasure" teachings scheduled for rediscovery across the following millennium. Almost none of that is early. The first full biography, the Zangs gling ma, was composed by Nyang ral Nyi ma 'od zer (1124–1192) roughly three and a half centuries after the events, and Lewis Doney's critical study of that text shows the figure being assembled in it rather than remembered.¹⁷

What the pre-eleventh-century record contains is much thinner and much stranger. In the dBa' bzhed Padmasambhava appears briefly, is credited with subduing the local powers and with a display of water-divining competence at Lhasa, arouses the suspicion of the court, and is sent away before the work is finished.⁵ In the Dunhuang manuscripts he is a tantric adept associated with the Vajrakīlaya rites — the ritual system of the consecrated dagger — and is not linked to Trisong Detsen at all. Sam van Schaik, who has worked directly on the Dunhuang tantric material, treats the historical Padmasambhava as a genuine but modest figure whose importance is overwhelmingly posthumous.² Doney's summary of the current position is that a historical Padmasambhava, questioned by an earlier generation of Tibetologists, is now "cautiously accepted" — which is a precise and honest formulation, and this record adopts it.¹⁷

The distinction matters editorially. The transmission Tibet actually received in the 770s was a ritual technology for dealing with a populated landscape, carried by one or more specialists from the northwestern tantric milieu. The Guru Rinpoche who now stands at the centre of the Nyingma tradition, whose image occupies more Tibetan wall space than any figure except the Buddha, is a construction of the twelfth to fourteenth centuries — and is, in his own way, a much larger cultural fact than the man.

It is also a distinction the atlas has to make carefully rather than smugly. "The later tradition invented him" is not the finding. The finding is that a real specialist did real work under imperial contract in the 770s; that the work was of a kind — binding local powers, consecrating ground — that leaves almost no documentary trace by its nature; that the earliest surviving Tibetan and Dunhuang accounts remember him as one figure among several rather than as the founder; and that a culture which had lost its empire spent the following centuries building, out of that figure, the founder it needed. Both the man and the construction are part of what was transmitted.

Samye, built as a model of the universe

Construction went ahead. The monastery of bSam yas — the name is usually glossed as "inconceivable" — was raised on the north bank of the Yarlung Tsangpo at Brag dmar, on a plan that was itself an argument. The central temple stood for Mount Meru, the axis of the Buddhist cosmos; four temples at the cardinal points stood for the four continents, with eight subcontinent chapels between them, sun and moon temples flanking the centre, four great stūpas at the corners, and a circular wall enclosing the whole as the ring of iron mountains at the world's edge. The Indian model most often named is Odantapurī in Magadha. The main temple was built in three storeys in three styles — Indian below, Chinese in the middle, Tibetan above — a piece of architectural diplomacy that states the empire's position between two civilisations without saying a word.

The dates are contested, and the record should say so plainly rather than pick one and pretend.

The 779 consecration is the date most historians work with, because it is anchored to something independent: the ordination of the first Tibetan monks, which the sources agree happened at Samye in that year, and to the imperial edict proclaiming Buddhism the religion of the state. Richardson's corpus preserves the pillar text recording that the "shrines of the Three Jewels were established by building temples at the centre and on the borders, Samye in Brag dmar and so on" — the flat administrative voice of an empire logging an act of policy.⁶

Building it required imports of a very ordinary kind alongside the doctrinal ones. Craftsmen came from Nepal and from the Tang territories; the three-storey stylistic scheme is not a legend but a description of who was on site. Timber had to be brought to a valley that has very little of it. The labour was corvée labour, drawn from the same households the empire drew soldiers from, and the consecration ceremonies were provisioned out of the same stores that provisioned campaigns. The mandala plan, in other words, was not a symbol laid over a building; it was the building, and the empire paid for the cosmos in grain, timber and unpaid work. Every later Tibetan monastery of the classic type is a descendant of that plan, and the plan is the reason a visitor to Samye today is standing inside a diagram of the universe as the eighth-century Indian tradition understood it.

Photograph by Lark Ascending. Samye Monastery, Dranang, Tibet. Public domain via Wikimedia Commons. · Public Domain

The debate that decided which Buddhism Tibet would keep

By the 790s Tibet had a problem of surplus. It was importing Buddhism from India across the Himalaya and simultaneously from Tang China through the occupied Hexi corridor, where the Tibetan administration of Dunhuang had inherited a functioning Chinese Buddhist establishment. The two arrived with incompatible accounts of how a person becomes enlightened. The Indian scholastic position, argued by Śāntarakṣita's student Kamalaśīla, was gradualist: enlightenment is the terminus of a long, structured cultivation of ethics, concentration and analytic insight. The Chan position, argued by the master Moheyan, was that conceptual effort is itself the obstruction, and that awakening is immediate.

The tradition remembers this as a formal debate held at Samye around 792–794, presided over by the emperor, ending with Kamalaśīla's victory, Moheyan's expulsion, and a standing imperial preference for Indian Buddhism thereafter. That account is the foundation of a millennium of Tibetan self-understanding: it is why Tibetan Buddhism looks to Nālandā rather than to Chang'an.

It is also the account modern scholarship handles most carefully. Paul Demiéville's Le concile de Lhasa of 1952 published a Chinese text from Dunhuang, purporting to be a record made by Moheyan's disciple Wangxi, in which the Chinese side wins.⁸ Jacob Dalton, summarising the state of the question, puts the resulting problem exactly: the Chinese work "purports to be a word-for-word record of the debate written by Wangxi, a direct disciple of Moheyan," and "its version of events differs radically from those of the various Tibetan sources; in this version, Moheyan wins the debate." The discovery, Dalton writes, "has led some scholars to doubt the very existence of the debate, suggesting that instead it should be viewed as indicative of an ongoing controversy through a series of only indirect encounters between Chinese and Indian factions at the Tibetan royal court. That said, it remains that all available sources agree that a debate of some kind did take place."¹⁸

What is not in doubt is the outcome at the level of policy. After the 790s the Tibetan court's patronage runs to India. Chinese Chan material continues to circulate — it is present in the Dunhuang manuscripts, and van Schaik has traced its quiet persistence inside later Tibetan meditation traditions — but it does so without institutional sponsorship, and the men who taught it lost their position at court.²

What changed and what was replaced

A language engineered to hold Sanskrit

The most consequential thing the Tibetan empire did with Buddhism was not build monasteries. It was rebuild the Tibetan language. Translating the Indian canon meant finding Tibetan equivalents for a technical vocabulary — dharma, skandha, pratītyasamutpāda, śūnyatā — that Tibetan had no words for, and doing it consistently enough that a reader in Amdo and a reader in Ngari would understand the same text the same way. The empire treated this as a matter for legislation.

Two royal decrees survive in substance. The sGra sbyor bam po gnyis pa, the "Two-Volume Lexicon," records a translation standard promulgated under Trisong Detsen and reissued in a final redaction sanctioned by imperial decree in 814, under his successor. Alongside it stands the Mahāvyutpatti, a fixed Sanskrit–Tibetan glossary of several thousand terms. Cristina Scherrer-Schaub's diplomatic analysis of these documents shows them functioning as genuine administrative instruments — decrees with legal form, not scholarly prefaces.¹⁵ Their provisions are startlingly modern in kind:

1. Established equivalents are fixed and may not be varied by individual translators
2. Word order should follow Tibetan syntax rather than calque the Sanskrit, except where meaning requires otherwise
3. Terms with no settled equivalent must be referred to the central authority, not improvised locally
4. Proper names and certain mantric material are to be transliterated rather than translated
5. Tantras and mantras are not to be translated or circulated without authorisation

The fifth provision is the one that shows what kind of document this is. A decree that restricts the circulation of tantric material is not a scholarly convention; it is a state managing a technology it considers dangerous. The court had imported both a scholastic tradition and an esoteric ritual tradition, and it was prepared to publish the first and licence the second. Scherrer-Schaub's reading of the decrees' diplomatic form — their protocols, their attestation clauses, their chain of authorisation — establishes that these were promulgated with the same machinery as a land grant or a treaty.¹⁵ Language policy in imperial Tibet was administered, not merely agreed.

The result was a purpose-built scholarly register: a Tibetan capable of carrying Indian philosophy with such regularity that scholars today routinely reconstruct lost Sanskrit originals from the Tibetan. The canon it eventually produced — the bKa' 'gyur of the Buddha's word and the bsTan 'gyur of the commentaries, together well over three hundred volumes — is one of the largest translation projects any premodern state undertook, and its terminological spine was laid down by decree in the imperial period.

Seven men and the institution they began

In 779, at Samye, Śāntarakṣita ordained the first Tibetan monks. The tradition calls them the sad mi mi bdun, the "seven men tested" — a trial cohort, drawn from aristocratic families, ordained to find out whether Tibetans could sustain monastic discipline at all. The number is small and the caution is real: monasticism was an untested import, and celibacy in particular was a direct affront to a clan society organised around lineage and inheritance.

The institution those seven began was novel in Tibet in a specific structural way. It was a corporate body that owned property, that recruited by ordination rather than birth, that could not be inherited, and that outlived every individual in it. Nothing in the pre-Buddhist Tibetan settlement worked like that. Kapstein's central argument is that the assimilation of Buddhism should be read as this kind of institutional and cultural sedimentation across centuries rather than as a conversion event — a slow accumulation of practices, memories and contested claims about what had happened.¹

The state then had to pay for it. The pillar inscriptions record the mechanism: monastic communities were endowed with assigned households whose labour and produce supported the monks, and those households were removed from the ordinary tax and corvée base. By the reign of Khri gtsug lde brtsan (Ralpachen, r. c. 815–838) the tradition reports an allocation of seven households to each monk. The Karchung pillar, raised under Khri lde srong brtsan around the turn of the ninth century, goes further and binds the future: it records the oaths of king, ministers and clans to protect the endowments in perpetuity, and its provisions include the clause that those who have gone forth into religion "may not be given as slaves to others and may not be suppressed by tax."⁶ Read from the throne, that is a guarantee of religious liberty. Read from the estate of an aristocratic clan whose taxable population has just been reassigned, it is something else.

The end of the tomb religion

What was displaced is easier to see in the negative. The imperial funerary complex — the tumuli at 'Phyong rgyas, the sacrificial programme, the retainer burial, the specialists who conducted it — does not survive the Buddhist settlement. The last royal tombs of the dynasty are the end of a tradition of monumental burial that had run for two centuries. Retainer burial and large-scale animal sacrifice, attested into the period around 800, do not continue. The ritual specialists of the royal cult lost their function at court and their livelihood with it, and because they wrote nothing that survives, they are visible in the record almost exclusively through the polemics of the people who replaced them.

Some of the complex was not destroyed but absorbed. The local deities Padmasambhava is remembered as binding were not abolished; they were sworn in. The mountain gods, earth lords and klu of the pre-Buddhist landscape reappear inside Tibetan Buddhism as oath-bound protectors — dam can, "those with a vow" — with their own liturgies, their own offerings, and a continuing practical role in Tibetan life. Rolf Stein's synthesis describes this stratigraphy across Tibetan religion with more care than any single-volume account since: what looks like syncretism is often a defeated cult retained under new management.⁹

Anonymous scribe, imperial-period Tibetan. Old Tibetan Chronicle, Pelliot tibétain 1287, first page. Bibliothèque nationale de France, Paris. Public domain via Wikimedia Commons. · Public Domain

Bon: a rival invented after the fact

The most persistent misunderstanding about this transmission concerns Bon. In the popular account, Bon is the indigenous shamanic religion of Tibet, which Buddhism arrived and suppressed. Specialist scholarship does not support that shape. Per Kværne's survey of the field distinguishes sharply between the ritual functionaries called bon in the imperial documents and the organised religion called Bon that appears from the eleventh century onward with its own canon, its own founder-figure in gShen rab mi bo, its own monasteries and its own doctrinal system — a system which is, structurally, closely parallel to Buddhism.¹⁴ Karmay's work on the Bon historical sources shows the tradition constructing a past for itself with the same tools, and often the same literary forms, that the Buddhist schools were using.¹³

The honest version is stranger than the popular one and worse for the losers. The imperial ritual specialists were not persecuted into an underground church that later resurfaced. They were made redundant, and their practices were partly absorbed and partly forgotten, and the organised Bon religion that arose two centuries later was a new formation that claimed them as ancestors. The people who actually served the royal cult in 750 have no descendants who speak for them. They have a name that was taken up by someone else.

The new cultural equipment

Against that loss, the list of what Tibet acquired between 755 and 842 is very long:

• A monastic institution capable of holding property, training specialists and reproducing itself independently of any dynasty
• A translated canon and a standardised technical vocabulary, fixed by decree in 814
• A scholastic method — the Indian debate-and-commentary tradition — that became the core of Tibetan education and remains so
• A tantric ritual system, including the Vajrakīlaya rites, that supplied Tibet with a technology for managing local powers
• An iconographic and artistic language, running from Indian and Central Asian models to what became the thangka tradition
• A historiography: the dBa' bzhed and its descendants are Buddhist accounts of the Tibetan past, and they set the shape of Tibetan self-understanding for a thousand years
• A durable international position — Tibet became, and for centuries remained, the reference civilisation for Buddhism across Inner Asia and Mongolia

What the cost was

The empire bought the monasteries with the empire

The bill arrives first as arithmetic. An empire that assigns households to the support of monks removes those households from the fisc that pays for garrisons. An empire that exempts a growing clerical population from tax and corvée narrows its own base while its frontier commitments stay constant. An empire that grants land in perpetuity, and binds its successors by pillar oath not to revoke the grant, has permanently alienated part of its revenue. Tibet did all three, at scale, in the space of two generations, while holding the Tarim garrisons and a live frontier with the Tang and the Uyghurs.

No imperial-period budget survives, so the scale has to be reasoned rather than counted, and this record is explicit that it is reasoning. The seven-households-per-monk figure comes from the later Tibetan histories rather than from an inscription, and it should be treated as an order of magnitude rather than a line item. But the order of magnitude is the point. A monastic population in the hundreds, at that rate of support, removes households in the thousands from tax and corvée; a population in the thousands removes tens of thousands. The empire that was carrying this was simultaneously garrisoning the Tarim oases, holding the Hexi corridor against Tang counter-attack and Uyghur pressure, and maintaining a frontier that ran from Kashmir to the Ordos. Beckwith's account of the empire's Inner Asian position makes plain how expensive that position was even without a clerical establishment attached to it.³

The aristocratic clans read this correctly. The conflict at court between the pro-Buddhist and anti-Buddhist parties, which the dBa' bzhed narrates as a religious struggle, was also and perhaps primarily a fiscal and constitutional one: whether an institution answerable to the throne and to India could be endowed out of resources the clans regarded as theirs.⁵ The regent Ma zhang grom pa skyes, remembered as the leader of the anti-Buddhist faction under the young Trisong Detsen, was removed by the pro-Buddhist party; the tradition says he was tricked into a tomb and entombed alive. Whether that happened is unrecoverable, and this record does not assert it. That the story was told, and told approvingly, is itself evidence about the temperature of the dispute.

The community that was expelled

The Samye debate had losers, and they were Buddhists. Whatever its actual form — single formal disputation, or the long controversy Dalton describes — the settlement of the 790s ended Chinese Chan Buddhism's institutional standing in Tibet. Moheyan and his students left. A living transmission lineage, with its own texts, its own teachers and a constituency inside the Tibetan empire's Chinese-speaking territories, was closed off by imperial decision.

The cost of that is not measured in bodies. It is measured in what Tibet did not get: a Buddhism in continuous conversation with East Asia, which is what Korea and Japan received and built on. Tibet chose India, and the choice was made once, at court, in the 790s, and never seriously revisited. Demiéville's edition of the Chinese record is the surviving voice of the party that lost — and it is worth noting that it survived only because it was sealed in a cave at Dunhuang and forgotten for nine hundred years.⁸

Violence that came with the ritual, not only with the state

Jacob Dalton's The Taming of the Demons takes up the part of this inheritance that Tibetan Buddhist apologetics has been least willing to hold. Working from Dunhuang manuscripts of the ninth and tenth centuries, Dalton traces mythic and ritual themes of demon-subjugation, blood sacrifice and "liberation" — the ritual killing of a being said to be released to a better rebirth — from the tantric literature into local Tibetan practice during and after the imperial collapse.⁴ The subjugation myth that legitimates Padmasambhava's work at Samye, in which a hostile deity is bound and converted by force, is not merely a metaphor for cultural change; it is a template that had ritual and sometimes physical applications.

Dalton's further argument is subtler and more damning: from the late tenth century onward the Tibetan tradition treated the post-imperial period as a violent "other" against which a purified Buddhism defined itself — a founding story in which the excesses belong to a dark age that has been left behind.⁴ The atlas notes this because it is a general mechanism. Cultural transmissions are narrated afterwards by their winners, and the narration is part of the transmission.

838, 842, and four centuries of fragmentation

The dynasty did not survive its own religious policy. Khri gtsug lde brtsan (Ralpachen), the most lavish patron of the monasteries and the king under whom the seven-household allocation is reported, was assassinated in 838. His brother Khri 'U'i dum brtan — remembered as Glang dar ma, Langdarma — reigned four years and was assassinated in 842 by a monk, Lha lung dpal gyi rdo rje, in an act the tradition treats as justified tyrannicide. The empire dissolved immediately into a succession war between rival claimants and never reassembled. Central Tibetan political unity was not restored for four hundred years.

The traditional account makes Langdarma a persecutor and the collapse a punishment. The specialist assessment is more careful. Yamaguchi Zuihō's study of the empire's formation is the most prominent statement of the revisionist case — that the anti-Buddhist portrait of Langdarma is a later construction, and that what actually happened was less a persecution of the doctrine than a withdrawal of state funding from a monastic establishment the treasury could no longer carry.¹¹ Van Schaik's narrative history takes a similar line: the monasteries lost their endowments and their monks dispersed, which from inside the tradition is indistinguishable from persecution and from outside looks like insolvency.² The distinction matters, and the record refuses to collapse it: a state can destroy an institution by ceasing to pay for it, and the people whose lives it was are entitled to call that destruction.

Either way the effect is documented. Organised monastic Buddhism in central Tibet effectively ceased for roughly a century and a half. The ordination lineage had to be reimported after 978 through the eastern route from Amdo and, from the west, through the kings of Guge in Ngari, who financed a second wave of translation and brought Atiśa from Vikramaśīla in 1042. Ronald Davidson's study of that revival treats the eleventh and twelfth centuries as a genuine second founding — a rebirth of Tibetan culture organised around tantric Buddhism and around the new translation lineages, not simply a resumption.¹⁹

The bill

Set out plainly, this transmission's costs were:

• An imperial dynasty, bankrupted in part by its own endowment policy, ending in two royal assassinations (838, 842) and a succession war
• Four centuries of political fragmentation on the Tibetan plateau, from 842 to the Mongol-backed Sakya administration of the 1240s
• A displaced royal cult: the funerary complex, retainer burial and sacrificial programme of the imperial religion ended, its specialists made redundant, its literature never written down and now unrecoverable
• A rival Buddhist tradition expelled — the Chinese Chan community removed from Tibetan institutional life after the 790s, its record surviving only by accident at Dunhuang
• A century and a half of monastic collapse after 842, in which the ordination lineage in central Tibet lapsed and had to be reimported
• A subjugation theology whose ritual applications, as Dalton documents, were not always figurative

Against those, the transmission's cost is not of the same order as the atlas's worst cases. There is no plague, no famine, no mass enslavement, no demographic collapse attributable to it. What there is instead is a state that spent itself into ruin acquiring a religion, and a religion that survived the ruin so completely that it became the thing Tibet is.

There is one more coincidence, and it belongs at the end. The Indian monastic institutions that supplied Tibet with its ordination lineage, its canon and its teachers — Nālandā, Odantapurī, Vikramaśīla — were destroyed in the campaigns of Bakhtiyār Khaljī and his successors between roughly 1193 and 1200, in the same decades that Tibet's second diffusion was consolidating the tradition it had received from them. Organised Buddhism ended on the Gangetic plain. It continued on the plateau, in Tibetan, in a technical vocabulary fixed by an imperial decree of 814, preserved by an institution whose founding dynasty had been dead for three hundred and fifty years.

The senders did not survive. The gift did.