Chapter 3. Freezing on an island: the formation of Theravāda¶

In the middle of the third century BCE, Mahinda led a mission to Sri Lanka and founded the Mahāvihāra at Anurādhapura. For the first time, Buddhism left the subcontinent as an institutional whole. The island gave it a combination of circumstances it had never encountered at home: a single royal power, one island and one dynasty, a sole patron whose favor or disfavor was immediately felt; a limited scale, with too few monks and laypeople to support a hundred contending schools; and distance from the homeland, which delayed the passage of new intellectual currents across the sea and made them possible to intercept. This combination determined that island Buddhism would face a different question from its continental counterpart. The continent asked, "How can we win the competition?" The island asked, "How can we preserve intact what we already have?"

First, take stock of what Mahinda brought. What exactly did Mahinda bring? Tradition answers: the complete set, the three baskets approved by the Third Council, together with commentaries. Buddhaghosa later claimed merely to be translating back into Pāli the old Sinhala commentaries Mahinda had brought. A critical answer must distinguish layers. What can safely be said is that he brought a core oral corpus of the Vibhajjavāda line: the main body of the Vinaya; the main body of the four Nikāyas, the layer that can be corroborated against the Chinese Āgamas (for the comparative method, see Volume 1); the tradition of doctrinal matrices and early treatises; and the earlier layer of the Khuddaka Nikāya. The "present Pāli canon," however, is the product of another two or three centuries of supplementation and consolidation of this core on the island. By the tradition's own account, the Kathāvatthu was newly composed in Aśoka's age. The canon thus contains an officially acknowledged late work of its own. The Parivāra is generally judged a later island supplement. Even in the commentarial age, the reciters of the Dīgha and Majjhima Nikāyas had not agreed on which books the Khuddaka Nikāya should contain, as cited in the Interlude. Even Pāli itself is an ecclesiastical language that took shape in transmission, not the particular dialect Mahinda spoke. In a word, Mahinda brought the seed and the cultivation manual, not the tree we see today.¹

The island's answer was freezing, completed in three operations, each with a date. The first sealed the text. After famine and war in the first century BCE, the three baskets were written down at Aluvihāra. The Interlude recorded one side of this event, its character as a vote of no confidence. Here the other side must be recorded. To put the text in writing was to freeze it. One particular oral version became the authoritative text, and subsequent additions or deletions could proceed only under the name of commentary. What was sealed into palm leaves was precisely the version of that transmission described above after its supplementation and consolidation on the island. Who had the right to interpret this canon immediately became the next point of dispute, as recorded in Volume 1, Essay 9.²

The second operation sealed the chain. When Abhayagiri was founded around 29 BCE, the king granted the monastery to a monk expelled from the Mahāvihāra. The Mahāvihāra categorically refused to hold uposatha with it. In Vinaya terms, breaking off joint uposatha amounts to declaring that the other party does not belong to the same monastic community. Essay 2 recorded the procedural character of the definition of schism. The Mahāvihāra used this definition as an instrument. There was no need to debate doctrine; it was enough to sever mutual recognition of formal acts of the community. The third operation sealed interpretation. In the fifth century, Buddhaghosa reorganized the commentarial system under the rule that it must not depart from the Mahāvihāra's ancient commentaries. The commentarial layer, too, was fixed in writing. Together, the three operations gave island orthodoxy three locks: a lock on the text, a lock on the chain, and a lock on interpretation.

The three locks retain their separate dates: writing down the canon in the first century BCE, the lineage separation associated with Abhayagiri around 29 BCE, and Buddhaghosa’s organization of the commentaries in the fifth century CE. The circle is not a sequential timeline. “New local thought” on the incoming arrow means new thought arriving from the Indian mainland. “Uposatha severed” means the Mahāvihāra’s refusal to hold the ceremony jointly with Abhayagiri, cutting off mutual recognition in formal Saṅgha acts; it does not mean that all Uposatha ceremonies ceased. “Abhayagiri expelled” compresses the account of its founder’s expulsion from the Mahāvihāra and the exclusion of his community from shared ritual recognition.

The text that was fixed was the oral version after its further development on the island, not the present canon supposedly brought there complete by Mahinda. The chapter distinguishes the claim of an unbroken ordination lineage from its later interruption and reintroduction from Siam in the eighteenth century. The cost at the bottom is expressed in the unit “intellectual movement”: it does not deny changes of doctrinal interpretation. Chapter 3 gives the sixth-century Mahānāma’s reading of sabhāva as sat-bhāva as a substantive change that did not announce itself as a new movement. The return and cost are both retained as the author’s assessment.

This volume will not retell the two monasteries' confrontation over nearly a millennium. The chronicles' Mahāvihāra partisanship, Faxian's observation that Abhayagiri was actually larger than the Mahāvihāra, and the two monasteries as two strategies were treated in Volume 1, Essay 9, Chapter 1; this chapter draws directly on that treatment. Restated in terms of this volume's question, their dispute was over whether to lock the gate for good. Abhayagiri continued accepting new sūtras and teachings from the homeland; the arrival of Vaipulya sūtras on the island is one example. The Mahāvihāra not only refused them but twice used the king's hand to burn books and expel monks. This episode comes from chronicles of the Mahāvihāra line, the victors' own proud record, and its direction is credible. In the twelfth century, Parakkamabāhu I united the three schools by royal command. Abhayagiri's thousand-year Dharma lineage was incorporated under the Mahāvihāra's rules.³ Freezing's final victory was again a royal command, the island version of the Dharma-protecting king model. The contract from the previous chapter reached its end here.

The rules require an explicit marker: the following is this volume's conjecture. The Mahāvihāra's conservatism arose from a chosen strategy of identity, not inertia. This community's greatest, almost sole, claim was orthodoxy. It was not in the homeland and could not produce the authority of originating new intellectual currents. What it guarded was the title "the original Buddhism" and the two things carrying that title: a canon fixed in writing and an unbroken ordination lineage. For a claim of this kind, variation was self-destruction. Accept one new sūtra, and the claim to be “the original” is diminished. Hold a joint uposatha once, and the chain loses a measure of its exclusivity. Locking the gate was guarding orthodoxy, not fearing novelty. There are three pieces of corroboration. First, during the centuries when royal power persistently favored Abhayagiri, the Mahāvihāra preferred to lose influence rather than yield. Short-term interests could not move it, indicating that it was guarding something long-term. Second, its internal self-description consistently placed an unmixed transmission first among its honors, not victory in doctrine. Third, at the level of entries in the treatises, the same discipline left its traces. Among the fifty-two conditions, nine instances give two names to the same thing. The tradition preferred to keep duplicate names in the list rather than merge entries, simply because each name was attached to wording in a sūtra (Volume 1, Part II, Essay 14, Chapter 7). It is the roster that is guarded, not the provisions. Here that sentence has a tangible instance on the doctrinal side.

Essay 2 left a sentence behind: what a council guards is less the provisions than the roster. Freezing on the island made this into an institution. At the level of provisions, the schools' monastic codes were already largely alike (Essay 2, Chapter 3). The Mahāvihāra had no exclusive invention in Vinaya learning. Its exclusive possession was the chain: an ordination genealogy running directly from Mahinda to the Buddha's time, with not one link borrowed. This chain later became the basis of orthodoxy throughout the southern Buddhist world. When Burma, Thailand, and other countries reorganized their monastic communities, successive generations brought in ordination lineages from the island or its line. There is a later episode as well. The island's own bhikkhu line once died out and had to be brought back from Siam in the eighteenth century. This is the most striking footnote to the logic of the chain; Essay 14 will develop it alongside the bhikkhunī case. The chain also has its cost. It recognizes status alone, not attained fruits. Once the vote on direction recorded in the Interlude, taking the teaching as the foundation, converged with the logic of the chain, island orthodoxy had its complete three-piece set: text, commentary, and ordination lineage. Every item could be checked; none involved attainment. A community whose highest mission was to preserve authenticity removed everything that could not be checked from the definition of orthodoxy.

Record one return and one cost. The return is that the only complete surviving canon of early Buddhism in the world today is the Pāli canon. The texts used as a baseline in the first two volumes are precisely the product of these three operations. Its critics, including this book, are its beneficiaries. That sentence must appear in the main text. The cost is that, once the way in for new thought had been locked, the island produced no new intellectual movement for more than a thousand years. The unit of measurement in this sentence needs explanation. The three locks can secure things that can be named: entries, rosters, and the wording of commentaries. They cannot lock down readings. Volume 1, Part II, Essay 14, Chapters 1 and 7 record a substantive doctrinal shift. Sabhāva was first used to define a dhamma in the Paṭisambhidāmagga. In the hands of the sixth-century commentator Mahānāma, it was read as sat-bhāva, existence, fixing the ontological reading. The judgment is that not one item was added to the list, yet the meaning of a word had shifted a level. Ontologization did not occur in any announced doctrine. It occurred in the procedures of cataloging and the pressure of the format. This shift had no banner, no author, and no debate, so it remains invisible when the unit is a movement. Vitality turned to preservation, and preservation sometimes became a rejection of "mixture." Freezing is a tradeoff at the level of a civilization, neither a virtue nor a defect. What it gained and what it relinquished are both in plain view.

The continent moved in precisely the opposite direction. In the century or two when the island was closing its canon, manuscript culture in the homeland opened the gate to sūtras as wide as it would go. The next essay looks at what poured through that gate.

A closing note records the theoretical cross-section. On the doctrinal side, scholasticization gave Buddhism the shape of a system for the first time: lists of dharmas, streams of moments, dialogues rewritten as entries, and entries organized into a system. Questions left open in the previous essay's closing note now have counterparts in the developing systems. Three gaps come into view here. How can non-self be reconciled with memory and responsibility? The Vātsīputrīyas answered with the pudgala, but they were never the target of a concerted attack, nor did the question stay where it was. Across the two hundred fascicles of the Vibhāṣā, the Vātsīputrīyas occur only seventeen times, compared with ninety for the Dārṣṭāntikas and one hundred and twenty-five for the "Analysts" (分别论者). They are occasional visitors, not standing opponents. A systematic reckoning had to wait until Vasubandhu in the fourth to fifth centuries, and the question, too, had shifted: from whether a bearer was needed to how much a continuity without a bearer could bear (Volume 1, Part II, Essay 16, Chapter 5, and Essay 17, Chapter 6). What carries the transmission of karmic results? Each school offered an account of continuity; the eve of the seed theory had arrived. What is the Buddha? The Sthaviras and Mahāsāṃghikas took a half each, as the human teacher and the transcendent sage began to part ways.

The cost of systematization is already visible. Collections of sayings can tolerate ambiguity; systems cannot. A correction to the direction of this argument must be added, or the passage sounds as though the cost arose from saying too much and could be remedied by saying it better. Volume 1, Part II, Essay 14, Chapter 7 judged the matter directly. These empty places cannot be resolved within analysis because they are conditions of the instrument's existence, not its objects. To use a list, relations must already exist; to arrange a sequence, time must already exist; to establish a list, a boundary must already exist. The empty places are therefore not omissions but the shape of this method's limits. More diligent commentaries or cleverer scholars would leave them just as empty. Essay 15, Chapter 7 confirms the same point from the Sarvāstivāda side. The southern tradition leaves them outside the list, so the list cannot claim to have exhausted what exists. Sarvāstivāda takes them into the list, whereupon those included items turn around and demand another set of entries for themselves. Leaving them outside and taking them inside are the two sides of the same boundary. Note that the two lists are distinct. The three gaps discussed here concern how non-self can be reconciled with memory and responsibility, what carries karmic results, and what the Buddha is. This is not a one-to-one restatement of the preceding essay’s three questions; who judges awakening, listed there, is not separately developed in this paragraph. The three groups listed in Part II are how dharmas relate to one another, how dharmas exist in time, and how the boundary between dharmas and non-dharmas is drawn.

References¶