Practice · SīlaVolume FourPart III · Three Vinaya SchoolsPaper 11 of 30

Two Sangha Crises and the Mechanism of Boundary-Renegotiation

中文主題：二僧團危機與邊界重議機制

Author: Shi Huijing 釋慧鏡

Date: Spring 2026

Version: 1.0

ID: SILA-P11

Abstract

When and how does a saṅgha boundary break, and when and how can it be renegotiated? The Vinaya literature preserves two paradigm cases that bear on this question — the five-point demand of Devadatta and the Council of Vesālī's hearing on the dasa-vatthūni — but the six surviving recensions (Pāli, Sifenlü / Dharmaguptaka, Wufenlü / Mahīśāsaka, Shisonglü / Sarvāstivāda, Mohesengqilü / Mahāsāṅghika, and the Mūlasarvāstivāda cluster) do not converge on a single narrative for either. This paper, instead of selecting any one recension as standard, reads the two crises side by side across recensions. The Council of Vesālī exemplifies procedural boundary-renegotiation: an eight-member committee (ubbāhikā), constituted by ñatti-dutiya-kamma with parity composition (four eastern, four western), tests the ten contested practices against pre-existing training-rules and returns a recognitive verdict — "not Dhamma, not Vinaya, not the Teacher's instruction" — restoring cross-regional consensus without legislating anything new. Devadatta's five-point demand, by contrast, exemplifies foundational-posture conflict: the irreducible tension between an ascetic absolutism that would universalise the dhutaṅga and the Buddha's permission-architecture refusal to do so terminates not in restored consensus but in Prātimokṣa-recital secession.

I. Introduction — Boundary and Its Mechanism of Renegotiation

In the late seventh century, writing from Śrīvijaya (in modern Sumatra) after twenty-five years of direct observation at Nālandā, the Chinese pilgrim Yijing opened his Nanhai jigui neifa zhuan with a methodological complaint that has not lost its force. On Chinese soil, he wrote, Vinaya-holding schools borrow from one another's recensions, and the chains of commentary and excerpt grow ever more entangled; in India, by contrast, plural traditions coexist without their principles contending. "If one verifies by the Vinaya text itself," he concluded, "it is not so; for judgments of light and heavy, a few lines of text suffice."¹ That posture — verification by the Vinaya text, in plural rather than in fusion — is the methodological starting-point of the present paper.

A contemporary Western reader encountering the two episodes treated here, "Devadatta" and "the Second Council", is likely to meet them as already-narrated stories whose contours are settled: an envious cousin, an apostate doctrine, an orthodox restoration. The contours feel familiar enough that revisiting them appears unnecessary. Yet when one verifies by the Vinaya text across the six surviving recensions, the unified story dissolves rather quickly. The five recensions that do enumerate Devadatta's five points do not enumerate the same five; the Mahāsāṅghika-vinaya (T1425) does not enumerate them at all. The ten Vesālī points are arranged in different orders across the four recensions that list them, while again T1425 does not produce a list. The apparently single narrative is a downstream composite, an aggregate impression formed by reading the recensions as if they were chapters in one book. They are not.

The thesis of this paper proceeds from that disaggregation. The two cases that the Vinaya literature preserves as paradigmatic moments of saṅgha-boundary stress are not the same kind of event, and they are not soluble by the same procedure. The Council of Vesālī is the paradigm of procedural boundary-renegotiation. Seven hundred monks convene, delegate the actual hearing to an eight-member committee (ubbāhikā) constituted by formal motion-and-second (ñatti-dutiya-kamma) with deliberate parity between the eastern and western parties, test the ten disputed practices item by item against the existing sikkhāpada corpus, and return a three-fold negative verdict — adhammika, avinayika, apagatasatthusāsana, "not Dhamma, not Vinaya, not the Teacher's instruction".² The boundary, on this account, does not break. It is reconfirmed by being renegotiated through machinery that was itself part of the Buddha's earlier dispensation. Devadatta's five-point demand is the paradigm of foundational-posture conflict. The dispute does not turn on whether a given training-rule applies in a given case; it turns on whether the Buddha himself permits the universalisation of an individual ascetic option (dhutaṅga) into a community-wide requirement. No committee can adjudicate that, because the committee would itself be operating within the very framework that is being contested. The Vesālī episode ends in consensus restored; the Devadatta episode ends in schism.

This bifurcation corresponds, with one explicit honesty-discipline qualification noted below in §V, to a two-fold classification advanced by Daoxuan (596–667) in his own Vinaya commentaries: fa-lun-seng-bheda, "the breaking of the dharma-wheel saṅgha", and karma-seng-bheda, "the breaking of the procedural saṅgha".³ The former is a conflict at the level of foundational doctrine; the latter, a conflict at the level of procedural acts performed within a single boundary. The two-fold classification names Devadatta as the paradigm case of fa-lun-seng-bheda by name. He describes karma-seng-bheda as a general typology without designating Vesālī as its paradigm. The application of Daoxuan's framework to Vesālī as the paradigm of karma-seng-bheda is therefore the present paper's interpretive completion of a typology Daoxuan leaves open — a point taken up explicitly in §V.

A second discipline informs the comparative work that follows. The mechanism of boundary-renegotiation is not legible within any single recension. If one reads only the Pāli Sattasati-kkhandhaka and Saṅghabheda-kkhandhaka, the absence of the five-point enumeration and of the ten-point list in T1425 reads as material loss rather than as structural choice. If one reads only the Dharmaguptaka-vinaya (T1428), the centrality of yāvajīvaṁ āraññikā ("for as long as life endures, forest-dwellers") as item one of Devadatta's demand in the Pāli will not be visible. Reading across the recensions, by contrast, makes visible a structure that none of them, taken alone, exhibits: a stable judgment-core with a variable periphery. The verdict — across the four enumerating recensions — judges the ten Vesālī points uniformly as non-permissible; salt (item one) and silver-and-gold (item ten) bracket the list at top and tail across all four; the middle items reorder. This stable-core / variable-periphery structure parallels the comparative finding of P10 in this same Part: that the eight garudhammas are stable as a block across recensions, while their placement varies. The two papers together describe a single methodological gesture in two different registers, synchronic and diachronic.

A third discipline is the indigenous reception axis. Comparative Vinaya scholarship is sometimes presumed to have begun with Western Buddhist studies in the twentieth century. It did not. Daoxuan and Yijing, working at the turn of the seventh and eighth centuries, performed comparative Vinaya work with explicit awareness of the multi-recensional Indian situation. Daoxuan's two-fold typology, mentioned above, is one outcome of that awareness; Yijing's methodological complaint, cited as epigraph, is another. The point is not nationalist. It is that the methodological resources for the comparative reading attempted here are partly continuous with materials that were already organised on East Asian soil more than a millennium before the founding work of modern Western comparative-Vinaya scholarship.

The structure of the paper is as follows. §II treats the Council of Vesālī as the paradigm of procedural boundary-renegotiation, identifying five architectural moves in the Sattasati procedure and locating the cross-four-recension judgment-core. §III treats Devadatta's five-point demand as the paradigm of foundational-posture conflict, presenting the cross-recension comparison of the five points and identifying two structural findings: the anti-eremitic narrative-framing posture inversion at T1450 fascicle 10 (at the redactional-posture layer, not the item-content layer; limited to this fascicle and not monolithic across the Mūlasarvāstivāda cluster), and the Mahāsāṅghika-vinaya's reframing of the crisis-genre itself. §IV stages the dialogue with modern scholarship through three figures — Bhikkhu Sujato, Charles S. Prebish, and Gregory Schopen — and notes the pivot between text-internal coherence and archaeological-historical reconstructibility. §V reads Daoxuan and Yijing as twin indigenous-reception anchors, with explicit attention to the asymmetry between Daoxuan's own attribution and the present paper's interpretive completion. §VI closes Part III of the volume and prepares the Mahāyāna turn into Part IV.

II. The Council of Vesālī — Procedural Boundary-Renegotiation as Paradigm

The Council of Vesālī is recorded in the Sattasati-kkhandhaka, chapter 12 of the Pāli Cullavagga, and is conventionally dated to roughly one century after the Buddha's parinibbāna.⁴ The trigger, in the Pāli account, is the visit of the elder Yasa Kākaṇḍakaputta to Vesālī, where he observes the local Vajjiputtaka monks accepting silver-and-gold offerings on uposatha days, water-filled bowl placed before them as receptacle. He raises the matter as unlawful. The Vajjiputtaka monks reply by formulating the dasa-vatthūni, ten points of practice that they hold to be permissible. Yasa seeks support; seven hundred monks, "neither more nor less" (satta bhikkhusatāni anūnāni anadhikāni), eventually convene at Vesālī to adjudicate.

2.1 The five architectural moves

Reading the Sattasati across the four enumerating recensions, the procedure that the council follows unfolds in five strict steps, each of which already lay within the Vinaya corpus established during the Buddha's lifetime.

The first move is the framing of the dispute as an adhikaraṇa, a formal Vinaya dispute requiring saṅgha-level adjudication. This framing is itself decisive: an adhikaraṇa triggers a procedure rather than an individual judgment, and the seven-hundred-monk assembly arises as the formal response to that frame.

The second move is delegation through ñatti-dutiya-kamma, the standard two-stage motion-and-second formal act of the saṅgha. Seven hundred is too large a body to argue point by point; the assembly therefore formally constitutes an eight-member committee — the ubbāhikā, "the drawn-out" or "the selected" — to carry the substantive deliberation on the assembly's behalf.

The third move is the parity composition of the committee. The eight members are constituted as four monks from the eastern party (Pāveyyaka) and four from the western (Pācīnaka) in deliberate symmetry. The parity is the procedural guarantor of neutrality: a committee that tilts toward either party would, by that fact, forfeit the legitimacy of its verdict.

The fourth move is the recognitive verdict. The committee tests the ten points one by one against the existing sikkhāpada corpus and returns a three-fold negative judgment: sabbāneva imāni bhikkhave vatthūni adhammikāni avinayikā apagatasatthusāsanāni, "all these practices, monks, are not Dhamma, not Vinaya, departed from the Teacher's instruction".⁵ The Dharmaguptaka-vinaya (T1428) preserves the same three-fold negation at fascicle 54: "these ten practices are not Dhamma, not Vinaya, not what the Buddha taught" (此十事非法非毘尼非佛所教).⁶ The verdict is recognitive, not legislative: the committee does not invent new rules; it recognises the unlawfulness of the disputed practices by reference to rules already in force.

The fifth move is the foundational principle on which the verdict rests. The principle was stated by Mahākassapa at the First Council, recorded in Pañcasatika-kkhandhaka (Cullavagga XI), as the transmission discipline of the whole tradition: apaññattaṁ nappaññapeyya, paññattaṁ na samucchindeyya, yathāpaññattesu sikkhāpadesu samādāya vatteyya — "what has not been laid down shall not be laid down; what has been laid down shall not be cut off; one should accept and live by the training-rules as they have been laid down".⁷ At the Second Council this formula is not re-cited; it is operationalised in the recognitive verdict. (One textual discipline is worth noting: the explicit formulation of the principle stands in Cullavagga XI, not in Cullavagga XII; the Sattasati's verdict enacts the principle's spirit without re-quoting its letter.)

What is structurally distinctive about all five moves is that each was already in place during the Buddha's lifetime. The Sattasati council does not invent procedure; it activates a mechanism of boundary-renegotiation that was already part of the dispensation. This — procedure under conditions of geographic dispersion and one generation's remove from the founder — is what "procedural boundary-renegotiation" means.

2.2 The cross-four-recension judgment-core

Across the four enumerating recensions — Pāli, Sifenlü / Dharmaguptaka (T1428), Wufenlü / Mahīśāsaka (T1421), and Shisonglü / Sarvāstivāda (T1435) — the verdict-core on the dasa-vatthūni is stable. Salt-storage stands at item one in all four; gold-and-silver stands at item ten in all four; all ten items are judged not permissible. The Mūlasarvāstivāda-vinaya kṣudraka-vastu (T1451) fascicle 40 supplies the Mūlasarvāstivāda anchor and is broadly consistent with the Dharmaguptaka enumeration, though with the same kind of minor middle-segment reordering observable across the four.⁸

The middle items (items two through nine) reorder across the recensions, and the precise wording of each item varies. This is the variable periphery against the stable core of the bracketing salt-1 and gold-silver-10 positions and the three-fold negative verdict. The pairing is not the product of late synthesis; each recension preserves the pairing internally, and the stability emerges from the comparison rather than being imposed on it.

2.3 The Mahāsāṅghika-vinaya as theoretical fifth pole

At this point a structural finding requires explicit naming. The Mahāsāṅghika-vinaya (T1425) does not enumerate the dasa-vatthūni. Its fascicle 33 records the Second Council narrative and treats the silver-and-gold prohibition motif within the tenth vagga (a procedural-miscellany section); but it does not produce the ten-point list, and its narrative architecture is organised not around an item-by-item adjudication but around the five purities (pañca-vidha-pariśuddhi, 五淨法), supplemented by a nine-fold ordinance sequence whose ninth rubric, "Dharma-conformity" (dharma-anudharma, 法隨順法), is uniquely Mahāsāṅghika.⁹ The same recension's narrative of the Devadatta crisis, in fascicles 7 and 26, similarly declines to enumerate the five points, focusing instead on Devadatta's claim "that the Prātimokṣa may be recited or not at will" — that is, Prātimokṣa-recital secession.

The methodological consequence is that T1425 should not be placed in a five-recension comparative table as the fifth data point. To do so would be to allow the table's implicit five-equal-column setting to absorb a divergence that is not, in fact, at the level of the data. T1425's divergence is at the level of the genre. It is a Mahāsāṅghika redactional decision about what a schism-narrative is for — not "what was lawful and what was not, item by item" but "what marks the threshold at which the Prātimokṣa-reciting community ceases to be one community". For comparative purposes, then, T1425 is best read as a fifth theoretical pole: an independent reframing of the crisis-genre that stands beside, and not within, the table of item-enumerating recensions.

This framing decision is itself a Mahāsāṅghika Vinaya-position. It is not gap or omission.

III. Devadatta's Five Points — Foundational-Posture Conflict as Internal Dynamics

The Council of Vesālī is a case of cross-regional procedural coordination one century after the parinibbāna. Devadatta's five-point demand is a case of intra-community foundational conflict during the Buddha's lifetime. The difference between the two cases is not chronological distance but the nature of the conflict. Vesālī's disputed points are matters of sikkhāpada-applicability that the existing rule-corpus can adjudicate by recognitive verdict; Devadatta's demand is a matter of the Buddha's posture on universalising dhutaṅga-asceticism, a matter that cannot be adjudicated by any committee operating within the framework that is itself contested.

3.1 The Pāli account and the Buddha's refusal

The Saṅghabheda-kkhandhaka (Cullavagga VII) records Devadatta's demand and the Buddha's response.¹⁰ The first item of the demand reads:

sādhu bhante bhikkhū yāvajīvaṁ āraññikā assu; yo gāmantaṁ osareyya vajjaṁ naṁ phuseyya.

"Good, venerable sir; let the monks be forest-dwellers for as long as life endures, and let blame attach to whoever moves into village outskirts."

Devadatta motivates the demand by appeal to the very vocabulary the Buddha himself uses to commend modest desire, contentment, paring-down, and the ascetic mode (dhutaṅga). The Buddha refuses:

na kho ahaṁ devadatta evaṁ vadāmi. yo icchati āraññiko hotu, yo icchati gāmante viharatu...

"I do not so speak, Devadatta. Let him who wishes be a forest-dweller; let him who wishes dwell in the village outskirts..."¹¹

The refusal makes visible the real axis of the conflict. The Buddha operates a permission-architecture: forest-dwelling is a permissible path, but it is not a required path. Devadatta argues a prescription-architecture: if forest-dwelling is a virtue, it must be required. The Buddha's refusal is not a refusal of forest-dwelling; it is a refusal of the universalisation of any single mode of practice into a community-wide requirement. That refusal is constitutive of the saṅgha's middle-way pluralism. To accede to Devadatta's demand would be to convert what the Vinaya preserves as permission into what it has consistently not preserved as prescription. The conflict cannot be resolved by ñatti-dutiya-kamma; the foundational posture is the point at issue.

3.2 The cross-recension table of the five points

Read across the recensions, the five-point demand exhibits substantial content variation, well beyond the kind of middle-segment reordering observable in the dasa-vatthūni:

Recension	Five points (notable items)	Treatment of forest-dwelling
Pāli Cullavagga VII 3.14.9	yāvajīvaṁ āraññikā enumerated as item one (with four further ascetic requirements)	listed as item one
Sifenlü / Dharmaguptaka (T1428)	five points without the forest-dwelling item; substituted by abstention from ghee, salt, and similar food restrictions	not listed (substituted by alimentary restrictions)
Shisonglü / Sarvāstivāda (T1435)	five points enumerated; forest-dwelling present but not as item one	listed; not first
Wufenlü / Mahīśāsaka (T1421)	five points enumerated; forest-dwelling present	listed
Mūlasarvāstivāda Saṅghabhedavastu (T1450), fascicle 10	five-item enumeration consistent in content with the other listing recensions; narrative-framing posture inverted (see §3.3)	enumeration listed; the framing is anti-eremitic at the redactional-posture layer
Mohesengqilü / Mahāsāṅghika (T1425), fascicles 7 and 26	no enumerated five points; the issue is reframed as "what was not laid down he laid down; what was laid down he opened up" plus Prātimokṣa-recital secession	this dimension does not apply

Four of the enumerating recensions — Pāli, Sifenlü, Wufenlü, Shisonglü — share, at the level of narrative-framing posture, a pro-eremitic consensus: forest-dwelling and dhutaṅga-asceticism are commended in the Buddha's own teaching, and the problem with Devadatta's demand lies in its push to universalise the option into a requirement, in contradiction of the Buddha's middle-way posture.

3.3 The Mūlasarvāstivāda Saṅghabhedavastu outlier — at the narrative-framing-posture layer, not the item-content layer

A single recension, the Mūlasarvāstivāda Saṅghabhedavastu (T1450) fascicle 10, presents a narrative-framing posture that diverges from the other four. The relevant passage at T24 p. 0149b records Devadatta addressing the community as follows:

"If a bhikṣu dwells within a village-house, then he is properly a bhikṣu; for those who dwell in unoccupied places, in araṇya-regions, this is not the case."¹²

Here village-residence is framed as the proper register of bhikṣu-status, and forest-dwelling — the very mode commended elsewhere — is framed as deviation. This is an anti-eremitic narrative-framing posture, an outlier against the pro-eremitic consensus of the other four enumerating recensions. The 4-versus-1 divergence sits not at the level of which five items are listed — the items themselves do not differ in any structurally remarkable way — but at the level of how the narrative frames those items.

Two qualifications discipline the citation of this finding. The first is that the Mūlasarvāstivāda tradition is not monolithic on the question. The anti-eremitic narrative-framing posture is limited to this fascicle of T1450. It does not extend across the wider Mūlasarvāstivāda cluster: the Bhikṣu-vinaya (T1442), the Mūlasarvāstivāda-bhikṣu-prātimokṣa gathā collection (T1443), and the Kṣudrakavastu (T1451) do not share this redactional posture. The Mūlasarvāstivāda cluster's internal non-monolithic structure, in other words, is itself a part of the finding. This is continuous with the double-register structure of the Mūlasarvāstivāda cluster identified at P10 §II.3 in a different sub-domain (the placement of the garudhammas). The second qualification is that the inversion is at the redactional-posture layer, not at the item-content layer. If one reads only the listed five items, T1450 fascicle 10 looks much like the other listing recensions. The inversion is in how those items are narratively framed, not in the items themselves. Calling this an "item-five inversion" or anything of that shape would be textually wrong; the finding lives at the level of narrative-framing posture.

3.4 What 4-versus-1 divergence implies

Why does T1450 fascicle 10 alone present this anti-eremitic posture? The full answer depends on the redactional history of the Mūlasarvāstivāda Vinaya corpus, which the present paper does not undertake to exhaust. The present claim is more limited: that the divergence exists, at which layer (narrative-framing posture, not item enumeration), and with what discipline of citation (limited to T1450 fascicle 10, not monolithic across the cluster). What the finding makes visible, for comparative purposes, is that cross-recensional reading cannot stop at the level of enumerated items. A narrative-framing-posture divergence may, in some cases, signal a deeper redactional position than an item-enumeration divergence would.

3.5 The Mahāsāṅghika reframing of the schism-narrative

The Mahāsāṅghika-vinaya's decision not to enumerate the five points is consistent with its decision not to enumerate the dasa-vatthūni. In fascicles 7 and 26 of T1425, the Devadatta narrative centres on "what was not laid down he laid down; what was laid down he opened up" and on "the Prātimokṣa — whether to study or not, just as my own intent desires", that is, Devadatta's claim of a right to opt out of the communal Prātimokṣa-recitation. The claim is read in Mahāsāṅghika as the threshold point: Prātimokṣa-recital secession is what schism is.¹³

The implication of this reframing is sharp. For Mahāsāṅghika, what matters in a schism is not whether five specific ascetic items are lawful, but whether the seceding party still falls within the same uposatha-recitation community. If not, the saṅgha-boundary is broken, and the item-by-item content of the disputed practices is, structurally, secondary.

One historiographic point deserves explicit treatment here. Earlier modern scholarship (including André Bareau's foundational Les Sectes bouddhiques du Petit Véhicule of 1955 and a number of subsequent currents) sometimes coupled the Mahāsāṅghika divergence with a popular suggestion that the Mahāsāṅghika tradition was "sympathetic to the Vajjiputtakas". Bhikkhu Sujato, in Sects and Sectarianism (Santipada / SuttaCentral, 2007), explicitly retracted this conflation. The Mahāsāṅghika-vinaya contains no text-attested support for any Mahāsāṅghika sympathy with the Vajjiputtakas; the suggestion travelled through a chain of associations of the form "Mahāsāṅghika ≈ great-assembly ≈ majority faction ≈ deposed majority opposed to the Vajjiputtakas", none of which is grounded in the text itself.¹⁴ The episode is itself a good example of text-internal reconstruction discipline: when the text fails to support a long-circulating attribution, the responsibility is to retract the attribution explicitly. The point is taken up again in §IV.

3.6 The two crisis-natures distinguished

The two episodes thus name two structurally distinct kinds of saṅgha-boundary crisis. The Vesālī episode is a procedural-rule conflict: the disputed practices can be measured against the existing rule-corpus, and the recognitive verdict restores consensus. The Devadatta episode is a foundational-posture conflict: the disputed posture cannot be measured against the existing rule-corpus because it concerns the Buddha's posture on what the rule-corpus is, and procedure has no purchase on the dispute. §V will return to this distinction in the vocabulary of Daoxuan's two-fold typology.

IV. Dialogue with Modern Scholarship — Sujato, Prebish, Schopen

The comparative method of the present paper stands in conversation with three figures in modern Vinaya studies whose work is differently situated in relation to the questions just raised. The conversation is staged not as triangulated balance but as a natural pivot: from Bhikkhu Sujato's text-internal reconstruction to Gregory Schopen's archaeological-historical caveat, with Charles S. Prebish positioned as the methodological precedent that grounds both.

4.1 Sujato — cross-recension stability as evidence of an early common layer

Bhikkhu Sujato's Sects and Sectarianism (2007) argues — within the limits of text-internal reconstruction, without reliance on archaeological or epigraphic data — that the cross-recension stability of the Sattasati procedure and the three-fold negative verdict is itself evidence that this episode belongs to a pre-sectarian common layer of the Vinaya. The reasoning: had each tradition fabricated the episode independently in the post-sectarian period, one would expect significantly less procedural stability and significantly less convergence on the verdict-language. The actual cross-recension structure is more parsimoniously accounted for by supposing a common stratum shared before the recensions diverged.¹⁵ (Sujato's volume is published under CC0 by Santipada / SuttaCentral and is freely available; the citation here paraphrases his methodological position rather than reproducing extended verbatim.)

The position is in harmony with the observation of §II of the present paper: the three-fold negative verdict (adhammika, avinayika, apagatasatthusāsana) is stable across the four enumerating recensions, and salt-1 / gold-silver-10 bracketing is stable across the four. These structural stabilities are not the product of late synthesis. On the Sattasati episode, then, the present paper aligns with Sujato.

A second feature of Sujato's argument is methodologically generalisable. The retraction of the "Mahāsāṅghika sympathised with the Vajjiputtakas" conflation, noted in §III.5 above, is exemplary of the self-discipline of text-internal reconstruction: where the text does not support a long-standing scholarly attribution, the responsibility is to retract the attribution, regardless of how rooted in the field's history the attribution has become.

4.2 Prebish — comparative Vinaya as methodological precedent

Charles S. Prebish's Buddhist Monastic Discipline: The Sanskrit Prātimokṣa Sūtras of the Mahāsāṃghikas and Mūlasarvāstivādins (1975) is the foundational work of modern Anglophone comparative-Vinaya studies. Its core methodological commitment is the refusal to treat any single recension as the standard against which the others are to be measured as deviations.¹⁶ Its working method is item-by-item cross-mapping across the Prātimokṣa tables of the various traditions, undertaken not to determine which recension is "original" but to bring into view what the recensions share and where they differ. The volume is not in open-access status; we paraphrase its methodological position here rather than reproducing verbatim text (paraphrase summary; no verbatim accessed).

Two features of Prebish's method bear on the present paper. The first is the refusal of any single recension as standard — a commitment that the present paper inherits and applies. The second is the move from item-by-item cross-mapping to recognising that, beyond the level of items, there can be genre-level divergence that no item-by-item table can fully accommodate. The Mahāsāṅghika-vinaya's non-enumeration of both the five points and the dasa-vatthūni is precisely such a genre-level divergence. The present paper, then, both inherits Prebish's methodological refusal and extends his approach: where item-by-item mapping is appropriate (the four enumerating recensions on the dasa-vatthūni), we deploy it; where genre-level divergence intervenes (T1425), we read the divergence as a fifth theoretical pole rather than a fifth data point in the same table. Prebish is the precedent; this is the extension.

4.3 Schopen — the archaeological-historical caveat

Gregory Schopen's Bones, Stones, and Buddhist Monks (1997), and his subsequent essays, advance a methodological corrective with respect to the historical reading of Vinaya narratives. Vinaya texts are not transparent reports of the events they narrate; they are rhetorical-political documents that reflect the institutional concerns of the redactors at the time of their redaction. The redaction of the various Vinaya corpora is itself a long, post-schism process. What is recoverable from the texts is principally the redactors' posture at the time of redaction; what is not directly recoverable, without strong corroborating evidence from material culture and epigraphy, is the event-history of the episodes the texts narrate (paraphrase summary; no verbatim accessed).¹⁷

Schopen's caveat and Sujato's text-internal reconstruction are not opposed, but they generate a natural pivot for the present paper. Sujato's claim is that cross-recension stability is internal evidence of an early common layer. Schopen's claim is that even cross-recension stability does not, on its own, license inference to historical actuality. The two claims are compatible: the Vinaya literature's redactional stability may attest a textual stratum shared before sectarian divergence, while the historical event the textual stratum describes may remain only indirectly accessible. The pivot is the distinction between textual coherence and historical reconstructibility.

4.4 The present paper's position between the three figures, and the cameo register

The present paper situates itself, with regard to these three figures, as follows. With Sujato: in harmony. The five architectural moves of §II take cross-recension stability as their organising structure, and the present paper treats the three-fold negative verdict and the salt-1 / gold-silver-10 bracketing as evidence of an early common stratum at the textual level. With Prebish: in inheritance. The refusal of any single recension as standard is the methodological floor of the present paper. With Schopen: caveat accepted without abandoning text-internal work. The present paper does not claim direct historical access either to the actuality of the Devadatta crisis or to the actuality of the Vesālī council; its claims are limited to the structural patterns that the Vinaya literature itself preserves — patterns that are present in the redacted texts and that the redactors (whenever, wherever) jointly produced.

One brief cameo register acknowledges related work. André Bareau's Les Sectes bouddhiques du Petit Véhicule (1955) is a foundational comparative-Vinaya contribution and a precursor to Prebish's project (paraphrase summary; no verbatim accessed). Recent text-internal cross-tradition work by Bhikkhu Anālayo on the bhikkhunī canon, Ann Heirman on the Dharmaguptaka recension, and Shayne Clarke on the Mūlasarvāstivāda cluster has refined recension-specific scholarly depth in ways that the comparative comparative method employed here partly draws on.

A further extension is, however, available. All three of the central figures, and the cameo cluster too, work within a methodological tradition rooted in nineteenth- and twentieth-century Western comparative religious studies. The history of comparative Vinaya scholarship in East Asia, by contrast, is significantly older — and significantly continuous with the indigenous Vinaya-master tradition. The next section turns to that tradition.

V. East Asian Indigenous Reception — Daoxuan and Yijing as Twin Anchors

The work of East Asian Vinaya masters is sometimes classified as "intra-traditional commentary" rather than as "comparative Vinaya scholarship", with the implication that it stands outside the comparative enterprise. The classification is convenient but obscures one important feature of the late seventh- and early eighth-century moment: Daoxuan (596–667) and Yijing (635–713) were performing comparative Vinaya work, with cross-recensional awareness, in two distinct but complementary registers. This section reads them as twin indigenous-reception anchors.

5.1 Daoxuan's two-fold typology — mechanism-classification

Daoxuan classifies saṅgha-breaking into two kinds in his own commentarial works — his Jiemo shu (Sifenlü shanbu suiji jiemo shu / 業疏, X0728) states that "according to the Vinaya, the breaking of the saṅgha is also of two kinds . . . karma-breaking is a lighter offence, dharma-wheel-breaking constitutes an ānantarya sin", and his Sifenlü xingshichao (T1804), in the thirteen ordination-obstacles, distinguishes "schism — that is, the dharma-wheel saṅgha; the breaking of the karma-saṅgha is not [counted among the obstacles]". The crispest formulation of the two-fold scheme is transmitted within the Nanshan line by Yuanzhao's Sifenlü xingshichao zichi ji (T1805, Song), glossing the "schism-by-admonition precept" (破僧違諫戒) of the Xingshichao's 隨戒釋相篇:

"The breaking of the saṅgha is twofold. The first, called fa-lun-seng-bheda ['breaking of the dharma-wheel saṅgha'], is when one establishes five doctrines to convert the world and breaks the four refuges and the eightfold path. The second, called karma-seng-bheda ['breaking of the procedural saṅgha'], is when within the same boundary parallel saṅgha-acts are performed. . . . As for those who do not complete the fruit: in this Jambudvīpa-age, only Devadatta has accomplished the breaking of the saṅgha definitively. The rest are merely preliminaries; this is limited to the period when the Buddha was in the world, because after the parinirvāṇa there is no rival teaching with which to contend. As for karma-seng-bheda, it may attain its fruit, and it remains possible after the parinirvāṇa."³

The classification is structurally precise. Fa-lun-seng-bheda is a breaking at the doctrinal-foundation layer: it requires the establishment of a competing doctrine and the breaking of the Buddha-dharma's constitutive structures (four refuges, eightfold path). It is possible only during the Buddha's lifetime, because after the parinirvāṇa there is no living rival teaching to compete with. Karma-seng-bheda is a breaking at the procedural-act layer: it consists of parallel saṅgha-acts being performed within a single territorial boundary (the uposatha, pavāraṇā, and other community-acts being conducted independently by two parties within the same sīmā). It can occur after the parinirvāṇa, because it does not require a living rival.

The transmitted formulation names Devadatta as the paradigm case of fa-lun-seng-bheda: "in this Jambudvīpa-age, only Devadatta has accomplished the breaking of the saṅgha definitively". The attribution is explicit within Daoxuan's framework. This corresponds directly to the present paper's reading of Devadatta's five-point demand in §III as foundational-posture conflict.

A discipline of honesty is required with respect to the Vesālī side of the mapping. Daoxuan does not, in this same passage, name the Council of Vesālī as the paradigm case of karma-seng-bheda. His description of karma-seng-bheda is general typology — same-boundary parallel saṅgha-acts, possibly bearing fruit, possible after the parinirvāṇa — without designating Vesālī as its paradigm. The application of Daoxuan's framework to Vesālī as the paradigm case of karma-seng-bheda is the present paper's interpretive completion of Daoxuan's framework, not Daoxuan's own attribution. The mapping is supported by the internal logic of Daoxuan's two-fold classification — the Vesālī procedure was indeed conducted within the Vinaya's procedural register, and the underlying dispute concerned procedural acts within the same regional saṅgha — but the attribution to Daoxuan himself stops at the general typology. We propose the mapping; we do not assign it to Daoxuan.

Daoxuan's Sifenlü xingshichao (T1804), fascicle 1, does contain a specific Vesālī judgment, on the matter of the uncut sitting-cloth (one of the ten vatthūni): "the additions made in some Vinayas are made outside the legitimate occasions; in the Sifenlü's 'Seven Hundred Council' the keeping of an uncut sitting-cloth . . . such practice should long since have been abandoned, yet now it is being revived — this is to act on the Vajji's deviant practice".¹⁸ Daoxuan here issues a specific verdict on a particular vatthu, but does not extend the verdict into a general karma-seng-bheda classification of the Vesālī episode. The present paper's mapping is therefore a combination: it joins Daoxuan's two-fold framework (in his Jiemo shu / 業疏 and Xingshichao) with his specific Vesālī judgment in the Xingshichao (T1804) to read Vesālī as the paradigm case of karma-seng-bheda. The combination is interpretive, and so identified.

A further note on the scope of Daoxuan's use in this paper. The Nanshan line (Yuanzhao's Zichi ji, T1805) also preserves an eighteen-point taxonomy of saṅgha-breaking (the nine-deviations / nine-correctitudes structure). That taxonomy is part of the wider Nanshan Vinaya-architectural project and is treated in subsequent papers of this volume; the present paper invokes it only by taxonomic name and does not expand the eighteen points. Here Daoxuan is engaged in the register of mechanism-classifier, not in any further doctrinal register; the further doctrinal registers of his work belong to later papers in this volume.

5.2 Yijing's cross-recension methodology — methodological anchor

Where Daoxuan is the mechanism-classifier of East Asian Vinaya, Yijing is the methodologist. His Nanhai jigui neifa zhuan, fascicle 1, made the cross-recensional Indian situation visible to the Chinese reading public from direct observation: "in the western country [India] the two practices coexist and their principles do not contend", and to the methodological critic: "on Chinese soil the Vinaya-holding schools pull on one another's recensions, and the commentaries and excerpts have become entangled and disorderly", with the methodological prescription: "if one verifies by the Vinaya text itself, it is not so; for judgments of light and heavy, a few lines of text suffice".¹

Three methodological claims are folded into this dense passage. First, the plural coexistence of Vinaya traditions in India is a fact, observed from twenty-five years' direct experience at Nālandā, not a doctrinal speculation: the "western country's two practices" do not contend in their principles. Second, plural coexistence does not require harmonising synthesis. The Chinese practice of "pulling on one another's recensions" produces tangled commentary, not clarity. Third, the Vinaya text itself is the criterion of judgment: "verifying by the Vinaya text" — zhun-yan lü-wen (准驗律文) — is the discipline by which judgments of relative weight are to be made, and "a few lines of text suffice" when the discipline is followed.

These three claims taken together are a methodological position in the cross-recensional comparative Vinaya register. Yijing is not engaging the Devadatta or Vesālī narratives at the level of historical event; he is engaging the outcome of the Indian situation — that multiple recensions had reached India, and continued to operate in India, without principled contradiction — and is recommending a corresponding methodology to Chinese Vinaya scholarship.

A discipline of citation is warranted here. Yijing functions as a methodological anchor, not a narrative anchor. His Nanhai jigui neifa zhuan does not narrate the Devadatta crisis or the Council of Vesālī. To cite him for "what he records about Vesālī" would mis-attribute; the contribution of his text in the present argument is the methodological frame.

5.3 The twin-anchor correspondence

Daoxuan and Yijing together constitute the twin indigenous-reception anchor of the present paper:

Anchor	Master	Source	Layer of contribution
Mechanism-classification	Daoxuan	Jiemo shu (業疏, X0728) + Sifenlü xingshichao (T1804); crisp formulation transmitted in Yuanzhao's Zichi ji (T1805)	Two-fold typology of saṅgha-breaking (foundational-doctrine vs procedural-act)
Cross-recensional methodology	Yijing	Nanhai jigui neifa zhuan (T2125) fascicle 1	Explicit methodological statement of "verifying by the Vinaya text" across recensions

The two are not isolated. Daoxuan's two-fold typology develops in close awareness of the Vinaya corpus's multi-recensional character; Yijing's methodological statement develops in close awareness of the actual multi-recensional Indian situation observed directly. Together, in two distinct registers — one a Chinese architectural-classification register, one a transregional methodological register — they evidence cross-recensional Vinaya awareness as indigenous to East Asian Vinaya scholarship in the late seventh and early eighth centuries.

This indigenous awareness precedes the founding work of modern Western comparative-Vinaya studies (Prebish 1975) by approximately thirteen hundred years. The historical priority is not the point. The point is that the methodological resources for the comparative reading that this paper attempts were already organised on East Asian soil, in the writings of working Vinaya masters who were in direct conversation with the same multi-recensional reality that contemporary scholarship engages. The reception axis traced in this section is one in which the comparative work of §§ II–IV stands continuous with — not in competition with — the indigenous Vinaya tradition's own self-conducting comparative method.

VI. Conclusion — Closing of Part III and Anticipation of Part IV

This paper is the third in a three-paper sequence that constitutes Part III of the present volume on sīla. The three papers approach the Vinaya through methodologically distinct but architecturally cumulative lenses. P9 read the single-tradition self-legislating logic of the Pātimokkha — the casuistic case-law of a single recension's internal categorial structure. P10 read the synchronic structure of three Vinaya lineages — the cross-five-recension comparison of the eight garudhammas' placement and the eight-category prātimokṣa sequence. The present paper read the diachronic structure of two saṅgha-boundary crises through the comparative lens, with the additional axis of indigenous reception by East Asian Vinaya masters.

The shared method across the three papers is comparative reading without preferred recension. The layer-by-layer progression: P9 single-tradition-internal structural disclosure; P10 cross-recension synchronic comparison; P11 cross-recension diachronic comparison joined with indigenous-reception axis. The three together compose the volume's three-layer presentation of Vinaya method — recension-internal, cross-recensional synchronic, cross-recensional diachronic with reception.

The structural distinction between the two crisis-natures — Devadatta's foundational-posture conflict (fa-lun-seng-bheda) and Vesālī's procedural-act conflict (karma-seng-bheda) — has a consequence beyond the boundaries of the Vinaya literature itself. It marks the limit of the saṅgha's internal procedural mechanism. Procedure can renegotiate boundaries where the underlying foundation holds; where the foundation is itself contested, procedure has no purchase. The next section of the volume turns to the Mahāyāna reframing of that limit. Paper twelve (P12) opens Part IV on the trīṇi-śīla-skandhāḥ (三聚淨戒) — the three-fold ethical scheme of the bodhisattva path — and the bodhisattva's reformulation of the saṅgha-boundary question as the cross-boundary ethical question, articulated through the three dimensions of restraint, beneficial conduct, and the ingathering of beings (攝律儀 / 攝善法 / 攝眾生). The Vinaya's procedural boundary, in the Mahāyāna turn, is reframed as the bodhisattva's ethical boundary; the seven hundred monks' deliberative consensus is reframed as the bodhisattva's three-fold ethical dimension. The present paper does not undertake that reframing; its claims terminate at the boundary of cross-recensional Vinaya comparison.

A final note. Yijing's seventh-century recommendation — to verify by the Vinaya text itself — is, in the end, a recommendation of respect. To read each recension on its own terms, recognising what each recension has chosen to enumerate and what each recension has chosen to reframe, is to respect both the recensions and the Vinaya masters across centuries who already laboured to read them comparatively. The cross-recensional reading that the present paper has attempted is one form, among many possible, of that respect.

References

Primary Pāli sources (SuttaCentral, CC0)

Cullavagga VII, Saṅghabheda-kkhandhaka (Pāli Vinaya II 180–203). SuttaCentral ID pli-tv-kd17. Mahāsaṅgīti edition; Brahmali / Sujato English translation.
Cullavagga XI, Pañcasatika-kkhandhaka (Pāli Vinaya II 284–294). SuttaCentral ID pli-tv-kd21. First Council narrative; apaññattaṁ nappaññapeyya placement.
Cullavagga XII, Sattasati-kkhandhaka (Pāli Vinaya II 294–308). SuttaCentral ID pli-tv-kd22. Second Council narrative; dasa-vatthūni enumeration; ubbāhikā committee; three-fold verdict.

Primary Chinese-translation sources (CBETA Electronic Tripiṭaka 2024)

Sifenlü (《四分律》, Dharmaguptaka). Translated by Buddhayaśas and Zhu Fonian, Late Qin (410–413). Taishō vol. 22, T1428. Fascicle 46 (Sanghabheda-vastu) and fascicle 54 (Qibai jifa pini, 七百集法毘尼). Verdict at p. 0968c18.
Wufenlü （《彌沙塞部和醯五分律》， Mahīśāsaka). Translated by Buddhajīva and Daosheng, Liu Song (423–424). Taishō vol. 22, T1421.
Shisonglü （《十誦律》， Sarvāstivāda). Translated by Puṇyatara, Kumārajīva et al., Late Qin (404–409). Taishō vol. 23, T1435.
Mohesengqilü (《摩訶僧祇律》, Mahāsāṅghika). Translated by Buddhabhadra and Faxian, Eastern Jin (416–418). Taishō vol. 22, T1425. Fascicles 7, 26 (Devadatta narrative without enumerated five points); fascicle 33 (Second Council narrative without enumerated ten points; five purities; nine-fold ordinance; tenth vagga).
Mūlasarvāstivāda Saṅghabhedavastu (《根本說一切有部毘奈耶破僧事》). Translated by Yijing, Tang (700–712). Taishō vol. 24, T1450. Fascicle 10 anti-eremitic narrative-framing posture passage at p. 0149b19.
Mūlasarvāstivāda Kṣudrakavastu (《根本說一切有部毘奈耶雜事》). Translated by Yijing, Tang. Taishō vol. 24, T1451. Fascicle 40 Second Council narrative; ten-point enumeration at p. 0411c04; verdict at p. 0414b.
Yuanzhao, Sifenlü xingshichao zichi ji (《四分律行事鈔資持記》). Song. Taishō vol. 40, T1805. Gloss on the Xingshichao's "schism-by-admonition precept": crisp formulation of the two-fold typology of saṅgha-breaking and the Devadatta-paradigm attribution (transmitting Daoxuan's framework).
Daoxuan, Sifenlü shanbu suiji jiemo shu (《四分律刪補隨機羯磨疏》, Jiemo shu / 業疏). Tang. Manji Zokuzōkyō X0728. Autograph source of the two-fold saṅgha-breaking classification ("according to the Vinaya, the breaking of the saṅgha is also of two kinds").
Daoxuan, Sifenlü xingshichao (《四分律行事鈔》). Tang. Taishō vol. 40, T1804. Fascicle 1, two-fold saṅgha-breaking in the thirteen ordination-obstacles + specific verdict on the uncut sitting-cloth in the Sifenlü "Seven Hundred Council" narrative.
Yijing, Nanhai jigui neifa zhuan (《南海寄歸內法傳》). Tang (691–712). Taishō vol. 54, T2125. Fascicle 1, "verifying by the Vinaya text" methodological statement at T54 pp. 0205c10–0206a04.

Modern scholarship

Bareau, André. Les Sectes bouddhiques du Petit Véhicule. Saigon: École française d'Extrême-Orient, 1955. Foundational comparative-Vinaya contribution. Paraphrase summary; no verbatim accessed (paywalled).
Prebish, Charles S. Buddhist Monastic Discipline: The Sanskrit Prātimokṣa Sūtras of the Mahāsāṃghikas and Mūlasarvāstivādins. University Park: Pennsylvania State University Press, 1975. Methodological foundation for modern Anglophone comparative-Vinaya studies; refusal of any single recension as standard. Paraphrase summary; no verbatim accessed (paywalled).
Schopen, Gregory. Bones, Stones, and Buddhist Monks: Collected Papers on the Archaeology, Epigraphy, and Texts of Monastic Buddhism in India. Honolulu: University of Hawai'i Press, 1997. Methodological corrective on the historical reading of Vinaya narratives. Paraphrase summary; no verbatim accessed (paywalled).
Sujato, Bhikkhu. Sects and Sectarianism: The Origins of Buddhist Schools. Taipei: Santipada / SuttaCentral, 2007. Text-internal reconstruction of the Second Council; explicit retraction of the Mahāsāṅghika–Vajjiputtaka conflation. CC0; verbatim citation permitted.

Cross-References

Part III internal

P9 (The Architecture of the Pātimokkha) — single-tradition self-legislating casuistic case-law; Part III's first paper. The present paper's §VI retrospect of Part III references P9 as the recension-internal layer of the volume's Vinaya method.
P10 (Three Vinaya Lineages in Conversation) — synchronic five-recension comparison of the eight garudhammas' placement; Part III's second paper and the present paper's direct predecessor. P10's stable-core / variable-periphery finding parallels the present paper's judgment-core / ordering-variance finding on the dasa-vatthūni; P10's Mūlasarvāstivāda double-register finding is a structural cousin to the present paper's Mūlasarvāstivāda cluster non-monolithic finding on T1450 fascicle 10. P10 §VI's forward-link to P11 is honoured here in §I as the synchronic-to-diachronic axis-pivot.

Forward-link within the volume

P12 (The Three-fold Ethical Scheme of the Bodhisattva — Triśīla and the Three-Body Architecture) — opens Part IV. The present paper's §VI forward-link prepares the Mahāyāna reframing of the saṅgha-boundary question as the bodhisattva's three-fold ethical question. Triśīla / 三聚淨戒 territory is reserved to P12; the present paper invokes it only as forward-link, not as in-scope material.
P16–P18 (Daoxuan jietigui architecture) — the Vinaya-substance ontology is reserved to these papers. The present paper's §V uses Daoxuan in the register of mechanism-classifier, not as ontologist; the eighteen-point taxonomy of nine deviations / nine correctitudes is invoked only by taxonomic name, without expansion.
P24 (Śūraṅgama Four-fold Pure Instruction) — the sīla-to-dhyāna transition territory. The present paper's §VI extension of "boundary from intra-saṅgha procedural to bodhisattva cross-realm ethical" is one step preparatory to the further extension treated in P24.

Volume-internal

SĪLA Part II (P7 Uposatha and the Genius of Intermittent Monasticism; P8 Sīlagandha — the Fragrance That Travels Against the Wind) — AN 11.1's kimatthiya eleven-link chain placing sīla at the head of the path to wisdom is canonical ground shared across the Vinaya-lineages; the present paper's Vinaya-diachronic reading rests on the sikkhā axis presented in Part II.
SĪLA Part I (P1 What Sīla Actually Means) — the volume's three-axis foundational reading.

Cross-volume

NIAN Part I — the Pāli-root foundation reading is methodologically continuous with the present paper's discipline of reading each tradition on its own terms.
DĀNA Part II (Mahāyāna turn) — structurally parallel: where DĀNA Part II reframes the gift category as the bodhisattva's dāna-pāramitā, the present paper's §VI forward-link anticipates a parallel reframing of the Vinaya-boundary category as the bodhisattva's triśīla-boundary. The parallel is structural, not thesis-borrowing.
SAMĀDHI Part forward — the sīla-to-dhyāna transition treated in SĪLA P24 prepares the entry into the SAMĀDHI volume.

CBETA collation: The Chinese-translation primary anchors of the present paper — the Dharmaguptaka-vinaya (T1428, Dharmaguptaka, translated by Buddhayaśas and Zhu Fonian, Late Qin 410–413, fascicle 46 Sanghabheda-vastu + fascicle 54 Qibai jifa pini, with the three-fold negative verdict "非法非毘尼非佛所教" at T22 p. 0968c18); the Wufenlü (T1421, Mahīśāsaka, translated by Buddhajīva and Daosheng, Liu Song 423–424, Second Council and Sanghabheda fascicles); the Shisonglü (T1435, Sarvāstivāda, translated by Puṇyatara, Kumārajīva et al., Late Qin 404–409, including the fascicle-37 Buddha's "praise-and-permission" dual teaching language and the Second Council fascicle); the Mahāsāṅghika-vinaya (T1425, Mahāsāṅghika, translated by Buddhabhadra and Faxian, Eastern Jin 416–418, fascicles 7 + 26 Sanghabheda without enumerated five points + fascicle 33 Second Council narrative organised around the five purities and nine-fold ordinance sequence, with the uniquely Mahāsāṅghika ninth rubric "Dharma-conformity" 法隨順法, and the silver-and-gold-prohibition theme in the tenth vagga — theoretical-fifth-pole principal anchor); the Mūlasarvāstivāda Saṅghabhedavastu (T1450, translated by Yijing, Tang 700–712, fascicle 10 "若有比丘住村舍內" at T24 p. 0149b19 — anti-eremitic narrative-framing posture principal anchor); the Mūlasarvāstivāda Kṣudrakavastu (T1451, translated by Yijing, fascicle 40 Second Council narrative + ten-point enumeration at T24 p. 0407c23–p. 0414b19, with the ten-point list at p. 0411c04 — principal anchor) — all citations have been collated against the CBETA Electronic Tripiṭaka 2024. The mechanism-classification anchors — Yuanzhao's Sifenlü xingshichao zichi ji (T1805, Song, gloss on the Xingshichao's "schism-by-admonition precept" at T40 p. 0290b, "破僧有二" two-fold fa-lun-seng-bheda / karma-seng-bheda typology with the explicit attribution "閻浮一化、唯提婆一人破僧究竟" of Devadatta as the paradigm case of fa-lun-seng-bheda, transmitting Daoxuan's framework — principal anchor; note that Taishō 40 No. 1805 is Yuanzhao's Zichi ji, not Daoxuan's Hanzhu jieben shu [the latter is X0714], and the autograph source of the two-fold classification is Daoxuan's Jiemo shu / 業疏 X0728 "據《律》破僧，亦有二種……破羯磨犯輕，破法輪成逆" together with his Xingshichao T1804 thirteen-obstacles passage) and Sifenlü xingshichao (T1804, Daoxuan, fascicle 1 section 5 line 352 specific Vesālī judgment on the uncut sitting-cloth "諸律增者，於緣外增之。《四分》〈七百結集〉中，得畜不截坐具……此應久廢，今往往重興，則用跋闍妄法也", limited to this specific vatthu and not extended into a general karma-seng-bheda classification of the Vesālī episode as a whole) — all citations collated against CBETA. The Yijing anchor — Nanhai jigui neifa zhuan (T2125, Yijing, Tang 691–712, fascicle 1 "神州持律、諸部互牽、章鈔繁雜……西國雙行、理無乖競……准驗律文則不如此" at T54 pp. 0205c10–0206a04, source of the present paper's epigraph) — collated against the CBETA XML edition (T54n2125). The Pāli primary anchors — Cullavagga VII Saṅghabheda-kkhandhaka (Vinaya II 180–203, SuttaCentral ID pli-tv-kd17, CC0, with Devadatta's five-point demand at 3.14.9 and the Buddha's refusal at 3.15.11), Cullavagga XI Pañcasatika-kkhandhaka (Vinaya II 284–294, pli-tv-kd21, with Mahākassapa's apaññattaṁ nappaññapeyya, paññattaṁ na samucchindeyya principle), and Cullavagga XII Sattasati-kkhandhaka (Vinaya II 294–308, pli-tv-kd22, with the eight-member ubbāhikā committee, the dasa-vatthūni, the seven-hundred quorum satta bhikkhusatāni anūnāni anadhikāni, and the three-fold verdict adhammika avinayika apagatasatthusāsana) — cited by PTS numeration with the Mahāsaṅgīti edition and Brahmali / Sujato translation on SuttaCentral (CC0) as cross-reference. Modern scholarship: Sujato, Sects and Sectarianism (Santipada / SuttaCentral 2007, CC0) is the principal text-internal-reconstruction anchor for §IV.1, with verbatim citation permitted under CC0; Prebish, Buddhist Monastic Discipline (Penn State University Press 1975), Schopen, Bones, Stones, and Buddhist Monks (University of Hawai'i Press 1997), and Bareau, Les Sectes bouddhiques du Petit Véhicule (École française d'Extrême-Orient 1955) are paraphrase engagements with explicit paraphrase summary; no verbatim accessed honest-disclosure markers as required by the project's scholarship conventions for paywalled sources.

CBETA collation: all Taishō T-number citations above verified against CBETA Electronic Tripiṭaka 2024; Pāli citations against the Mahāsaṅgīti edition and Brahmali / Sujato translation on SuttaCentral (CC0).

Author: 釋慧鏡 Shi Huijing Project: Point-to-the-moon · Six Practice Gates License: CC BY-NC-SA 4.0 Repository: github.com/pointing-at-the-moon/six-practice-gates

Pointing at the Moon · Six Practice Gates · Volume IV SĪLA · Paper Eleven · May 2026

Footnotes

Yijing, Nanhai jigui neifa zhuan (《南海寄歸內法傳》), fascicle 1, in Taishō shinshū daizōkyō vol. 54, T2125, pp. 0205c20–0206a04 (continuous block). The phrases "various schools pull on one another" (諸部互牽), "the western country's two practices coexist without principles contending" (西國雙行、理無乖競), and "verifying by the Vinaya text" (准驗律文) are clustered around T54 p. 0205c20. Yijing also makes explicit reference, in this same fascicle, to the four-recensions-in-coexistence Indian context. ↩ ↩²
Cullavagga XII, Sattasati-kkhandhaka (七百集篇), Vinaya II 294–308; SuttaCentral ID pli-tv-kd22. The three-fold verdict sabbāneva imāni bhikkhave vatthūni adhammikāni avinayikā apagatasatthusāsanāni marks the threefold negative judgment on the dasa-vatthūni. ↩
The two-fold classification of saṅgha-breaking (fa-lun-seng-bheda / karma-seng-bheda) is Daoxuan's own framework, found in his autograph works: his Sifenlü shanbu suiji jiemo shu (《四分律刪補隨機羯磨疏》, Jiemo shu / 業疏) states "according to the Vinaya, the breaking of the saṅgha is also of two kinds . . . karma-breaking is a lighter offence, dharma-wheel-breaking constitutes an ānantarya sin" (Manji Zokuzōkyō X0728), and his Sifenlü xingshichao (T1804), in the thirteen ordination-obstacles, distinguishes "schism — that is, the dharma-wheel saṅgha; the karma-saṅgha-breaking is not [an obstacle]". The crisp formulation quoted here — "the breaking of the saṅgha is twofold . . . in this Jambudvīpa-age only Devadatta has accomplished the breaking of the saṅgha definitively" (破僧有二……閻浮一化、唯提婆一人破僧究竟) — is from Yuanzhao's Sifenlü xingshichao zichi ji (《四分律行事鈔資持記》), glossing the "schism-by-admonition precept" of the Xingshichao's 隨戒釋相篇, Taishō vol. 40, T1805, p. 0290b. (Note: Taishō 40 No. 1805 is "Yuanzhao, Sifenlü xingshichao zichi ji" (Song), not Daoxuan's Sifenlü hanzhu jieben shu; the latter is Manji Zokuzōkyō X0714.) The application of Daoxuan's framework to the Vesālī council as the paradigm case of karma-seng-bheda is the present paper's interpretive completion; the passage describes karma-seng-bheda in general-typology terms ("within the same boundary parallel saṅgha-acts are performed... it may attain its fruit, and is possible after the parinirvāṇa") without designating Vesālī as its paradigm. ↩ ↩²
On the dating of the Council of Vesālī, the Pāli tradition places it one hundred years after the parinibbāna; some Chinese-translated traditions transmit one hundred and ten years. The present paper does not take a position on the dating question; it engages the textual structure as preserved across the recensions, not the chronology. ↩
Cullavagga XII, SuttaCentral pli-tv-kd22. The original fragment: sabbāneva imāni bhikkhave vatthūni adhammikāni avinayikā apagatasatthusāsanāni... satta bhikkhusatāni anūnāni anadhikāni dutiyaṁ saṅgītiṁ akaṁsu ("monks, all these practices are not Dhamma, not Vinaya, departed from the Teacher's instruction... seven hundred monks, neither more nor less, conducted the Second Recitation"). ↩
Dharmaguptaka-vinaya (《四分律》), fascicle 54, Qibai jifa pini (〈七百集法毘尼〉), in Taishō vol. 22, T1428, p. 0968c18. The three-fold negation "not Dhamma, not Vinaya, not what the Buddha taught" (非法非毘尼非佛所教) corresponds structurally to the Pāli's adhammika, avinayika, apagatasatthusāsana. ↩
Cullavagga XI, Pañcasatika-kkhandhaka (五百集篇), SuttaCentral pli-tv-kd21. Apaññattaṁ nappaññapeyya, paññattaṁ na samucchindeyya, yathāpaññattesu sikkhāpadesu samādāya vatteyya — the principle's explicit textual location is Cullavagga XI (the First Council narrative), not Cullavagga XII (the Second Council narrative). The Sattasati's recognitive verdict operationalises the principle without re-citing its letter. ↩
Mūlasarvāstivāda-vinaya kṣudraka-vastu (《根本說一切有部毘奈耶雜事》), fascicle 40, in Taishō vol. 24, T1451, pp. 0407c23 (opening of the "remainder of the five hundred and seven hundred council" narrative) through 0414b (verdict section); the ten-item enumeration appears at p. 0411c04. This fascicle is the Mūlasarvāstivāda anchor for the Second Council. Its enumeration follows the broader pattern of the four enumerating recensions, with minor middle-segment reordering consistent with the periphery variation noted in the main text. ↩
Mahāsāṅghika-vinaya (《摩訶僧祇律》), fascicle 33, in Taishō vol. 22, T1425. The silver-and-gold prohibition theme appears in the tenth vagga (a procedural-miscellany section). The recension does not produce a complete ten-point enumeration; its narrative architecture is organised around the five purities (pañca-vidha-pariśuddhi, 五淨法) and a nine-fold ordinance sequence (nava-aṅga, 九法序). The ninth rubric, "Dharma-conformity" (dharma-anudharma, 法隨順法), is uniquely Mahāsāṅghika and is structurally continuous with the ninth rubric finding of P10. ↩
Cullavagga VII, Saṅghabheda-kkhandhaka (破僧揵度), Vinaya II 180–203; SuttaCentral pli-tv-kd17. The Devadatta five-point demand is at pli-tv-kd17 3.14.9. ↩
Cullavagga VII 3.15.11. Na kho ahaṁ devadatta evaṁ vadāmi. yo icchati āraññiko hotu, yo icchati gāmante viharatu — "I do not so speak, Devadatta. Let him who wishes be a forest-dweller, let him who wishes dwell in the village outskirts" — manifests the constitutive difference between the Buddha's permission-architecture and Devadatta's prescription-architecture. ↩
Mūlasarvāstivāda Saṅghabhedavastu (《根本說一切有部毘奈耶破僧事》), fascicle 10, in Taishō vol. 24, T1450, p. 0149b19. The passage "if a bhikṣu dwells within a village-house, then he is properly a bhikṣu; for those who dwell in unoccupied places, in araṇya-regions, this is not the case" (若有比丘住村舍內，乃名比丘；若處空閑、阿蘭若處者，無有此事) presents the anti-eremitic narrative-framing posture. Two qualifications attach to citation of this passage. First, the anti-eremitic narrative-framing posture is limited to this fascicle; the wider Mūlasarvāstivāda cluster — including T1442, T1443, and T1451 — does not share this redactional posture, which extends the double-register finding of P10 on the same cluster. Second, the inversion sits at the narrative-framing-posture layer, not at the item-content layer: the five items themselves do not significantly differ from those of the other enumerating recensions. The inversion is in how those items are narratively framed, not in the items themselves. ↩
Mahāsāṅghika-vinaya, fascicles 7 and 26, T1425. The Devadatta narrative centres on "what was not laid down he laid down, what was laid down he opened up" (於十二事不制者制、制者便開, fascicle 7) and on "the Prātimokṣa — whether to study or not, just as my own intent desires" (波羅提木叉欲學不學自從我意, fascicle 26), the latter being Devadatta's claim to a self-determined right of opting out of the communal Prātimokṣa-recitation, that is, Prātimokṣa-recital secession. The recension does not produce the five-point enumeration; the typical items (forest-dwelling, ascetic-cloister, abstention from fish and meat, etc.) do not appear as a list in the recension's Devadatta material. ↩
Bhikkhu Sujato, Sects and Sectarianism: The Origins of Buddhist Schools (Taipei: Santipada / SuttaCentral, 2007), CC0. Sujato explicitly retracts the long-circulating suggestion that the Mahāsāṅghika tradition was "sympathetic to the Vajjiputtakas", a suggestion that runs through André Bareau's Les Sectes bouddhiques du Petit Véhicule (1955) and subsequent currents. Sujato argues that the Mahāsāṅghika-vinaya contains no text-attested support for any such sympathy, and that the conflation arose through a chain of associations ("Mahāsāṅghika ≈ great-assembly ≈ majority faction ≈ deposed majority opposed to the Vajjiputtakas") not grounded in the text. The retraction is itself exemplary of text-internal reconstruction discipline. ↩
Sujato, Sects and Sectarianism (Santipada / SuttaCentral, 2007), CC0. Sujato's thesis: the cross-recension stability of the Council of Vesālī's procedure and verdict-language constitutes internal evidence that the episode belongs to a pre-sectarian common stratum of the Vinaya. The present paper inherits this text-internal method (without claiming, as §IV.4 makes explicit, direct historical access to the actuality of the events described). ↩
Charles S. Prebish, Buddhist Monastic Discipline: The Sanskrit Prātimokṣa Sūtras of the Mahāsāṃghikas and Mūlasarvāstivādins (University Park, PA: Pennsylvania State University Press, 1975). The methodological commitment to refusing any single recension as standard against which the others are to be measured is the foundational gesture of modern Anglophone comparative-Vinaya studies. Source is paywalled; the position cited here is a paraphrase (paraphrase summary; no verbatim accessed). ↩
Gregory Schopen, Bones, Stones, and Buddhist Monks: Collected Papers on the Archaeology, Epigraphy, and Texts of Monastic Buddhism in India (Honolulu: University of Hawai'i Press, 1997), and subsequent essays. Schopen's methodological corrective: Vinaya narratives are rhetorical-political documents reflecting the redactors' institutional concerns at the time of redaction, not transparent reports of the events they narrate. The present paper accepts Schopen's caveat and limits its claims accordingly. Source is paywalled; the position cited here is a paraphrase (paraphrase summary; no verbatim accessed). ↩
Daoxuan, Sifenlü xingshichao (《四分律行事鈔》), fascicle 1, in Taishō vol. 40, T1804. The specific Vesālī judgment on the uncut sitting-cloth — "the additions made in some Vinayas are made outside the legitimate occasions; in the Sifenlü's 'Seven Hundred Council' the keeping of an uncut sitting-cloth . . . such practice should long since have been abandoned, yet now it is being revived — this is to act on the Vajji's deviant practice" (諸律增者，於緣外增之。《四分》〈七百結集〉中，得畜不截坐具……此應久廢，今往往重興，則用跋闍妄法也) — is Daoxuan's verdict on this specific vatthu. Daoxuan does not extend this verdict into a general karma-seng-bheda classification of the Vesālī episode as a whole. ↩

Source on GitHub ↗