---
title: The Three-Body Architecture of Bodhisattva Sīla
subtitle: Asaṅga's Three-Fold Structural Invention and Its Chinese Reception
volume: 戒門 SĪLA
paper_id: "SILA-P12"
paper_number: P12
language: en
date: 2026-05-21
author: "釋慧鏡 (Shi Huijing)"
license: CC BY-NC-SA 4.0
series: Six Practice Gates
abstract_en: "The 'three clusters of pure precepts' (sānjù jìngjiè 三聚淨戒) — Mahāyāna Buddhism's most influential structural reframing of monastic ethics — has received remarkably little focused analysis in Western Buddhist studies. This paper argues that the three clusters are not a list of three rules but a three-dimensional architectural design for bodhisattva ethics: the restraint cluster (saṃvara-śīla) inherits the seven-fold prātimokṣa as its structural foundation; the accumulation-of-virtue cluster (kuśala-dharma-saṃgrāhaka-śīla) brings the totality of bodhisattva practice within the precept's purview; and the sattva-welfare cluster (sattvārtha-kriyā-śīla) names the benefit of all sentient beings as the defining dimension of the bodhisattva precepts. Two philological findings are advanced. First, the four-character compound sānjù jìngjiè itself appears in none of the three primary transmission texts (T1581, T1582, T1485); it stabilizes only in the Chinese commentarial layer — Zhiyi (538–597) and Daoxuan (596–667) organize the framework as the "three clusters," and the fixed four-character compound is attested in Mingkuang's Tang-era *Tiantai pusa jie shu* (T1812). Second, T1581 (Dharmakṣema) and T1582 (Guṇavarman), translated from the same Sanskrit Bodhisattvabhūmi, employ almost entirely non-overlapping nomenclatures; T1582 supplies not the names but the strongest doctrinal argument — the Upāli dialogue — for the third cluster's defining status. A three-way dialogue with Funayama Toru, Michael Zimmermann, and Charles Goodman situates the contribution. Mingkuang's Tiantai pusa jie shu (T1812) supplies the paper's eponymous title: an explicit mapping of the three clusters onto the three bodies (trikāya) of Buddhahood, recasting precept-classification as a structural map of the entire bodhisattva path."
---

# The Three-Body Architecture of Bodhisattva Sīla

**Asaṅga's Three-Fold Structural Invention and Its Chinese Reception**

*中文主題：三聚淨戒之三身架構——大乘的結構發明與漢傳承接*

---

**Author:** 釋慧鏡 Shi Huijing
**Date:** May 2026
**License:** CC BY-NC-SA 4.0
**Version:** 1.0

---

> 「律儀戒，能令心住；攝善法戒，自成佛法；攝眾生戒，成就眾生。是名菩薩一切事。」
> *(Translator's rendering: "The restraint precept stabilizes the mind; the accumulation-of-virtue precept perfects one's own Buddha-dharma; the sattva-welfare precept accomplishes the liberation of sentient beings. This is called the totality of the bodhisattva's work.")*
> — *Pusa dichi jing* (T1581), fascicle 5, translated by Dharmakṣema (Northern Liang, 5th century)[^1]

---

## Abstract

The "three clusters of pure precepts" (*sānjù jìngjiè* 三聚淨戒) — Mahāyāna Buddhism's most influential structural reframing of monastic ethics — has received remarkably little focused analysis in Western Buddhist studies. This paper argues that the three clusters are not a list of three rules but a three-dimensional architectural design for bodhisattva ethics: the restraint cluster (*saṃvara-śīla*) inherits the seven-fold *prātimokṣa* as its structural foundation; the accumulation-of-virtue cluster (*kuśala-dharma-saṃgrāhaka-śīla*) brings the totality of bodhisattva practice within the precept's purview; and the sattva-welfare cluster (*sattvārtha-kriyā-śīla*) names the benefit of all sentient beings as the *defining* dimension of the bodhisattva precepts. Two philological findings are advanced. First, the four-character compound *sānjù jìngjiè* itself appears in none of the three primary transmission texts (T1581, T1582, T1485); it stabilizes only in the Chinese commentarial layer — Zhiyi (538–597) and Daoxuan (596–667) organize the framework as the "three clusters," and the fixed four-character compound is attested in Mingkuang's Tang-era *Tiantai pusa jie shu* (T1812). Second, T1581 (Dharmakṣema) and T1582 (Guṇavarman), translated from the same Sanskrit *Bodhisattvabhūmi*, employ almost entirely non-overlapping nomenclatures; T1582 supplies not the names but the strongest doctrinal argument — the Upāli dialogue — for the third cluster's defining status. A three-way dialogue with Funayama Toru, Michael Zimmermann, and Charles Goodman situates the contribution. Mingkuang's *Tiantai pusa jie shu* (T1812) supplies the paper's eponymous title: an explicit mapping of the three clusters onto the three bodies (*trikāya*) of Buddhahood, recasting precept-classification as a structural map of the entire bodhisattva path.

**Keywords:** three clusters of pure precepts; *saṃvara-śīla*; *kuśala-dharma-saṃgrāhaka-śīla*; *sattvārtha-kriyā-śīla*; bodhisattva precepts; *Bodhisattvabhūmi*; *Pusa dichi jing*; *Pusa shan jie jing*; *Yogācārabhūmi*; *Tiantai pusa jie shu*; Dharmakṣema; Guṇavarman; Zhiyi; Mingkuang; *trikāya*; Asaṅga; structural invention; Mahāyāna ethics; Chinese reception

## 摘要

三聚淨戒不是三條戒律的清單，而是菩薩戒的三維結構性設計：律儀戒（*saṃvara-śīla*）承接聲聞七眾別解脫律儀為結構基礎，攝善法戒（*kuśala-dharma-saṃgrāhaka-śīla*）將個人修行的全部善業納入戒的對境，攝眾生戒（*sattvārtha-kriyā-śīla*）以利益眾生為菩薩戒的定義性向度。本篇追蹤兩個漢譯路徑（曇無讖 T1581／求那跋摩 T1582）在術語體系上的分歧，揭示「三聚淨戒」這一四字固定術語並不見於任何傳譯底本，而是漢傳詮釋層的概括標籤。本篇與船山徹（傳譯史軸）、Michael Zimmermann（梵文文本軸）、Charles Goodman（倫理哲學軸）三家研究形成承讓與推進的對話。最後通過明曠《天台菩薩戒疏》（T1812）的三聚↔三身明文映射，本篇展示三聚不僅是戒律分類，更是菩薩道止／智／悲三維的整體結構圖譜。

**關鍵詞：** 三聚淨戒、攝律儀戒、攝善法戒、攝眾生戒、饒益有情戒、菩薩戒、《菩薩地持經》、《菩薩善戒經》、《瑜伽師地論》、《菩薩戒義疏》、曇無讖、求那跋摩、智顗、三身架構、結構性發明

---

## I. Redrawing the Boundary: From Śrāvaka Precepts to Bodhisattva Precepts

The previous paper in this volume (P11) closed with two diachronic crises *within* the early *Vinaya* literature — the Council of Vesālī's ruling on the ten contested practices, and Devadatta's five-point demand — both of which the *saṅgha* eventually disposed of through procedural boundary-renegotiation. Yet the very fact that such procedures could be invoked already presupposed a particular framework: the *śrāvaka* framework, whose central concern is individual liberation under conditions of communal purity within the *saṅgha* boundary. The procedures presupposed the frame; they did not interrogate it.

A different question, however, presses from outside the *śrāvaka* horizon. If the field of ethical concern in Mahāyāna practice extends beyond the seven-fold *prātimokṣa* community to the welfare of all sentient beings, then where exactly does the ethical boundary lie? Can the purity-standard of the *śrāvaka* precepts carry the weight of a path whose ultimate concern is the liberation of every sentient being? The question is not whether the *prātimokṣa* is defective — it is whether the *prātimokṣa*, as conceived, can accommodate the soteriological scope that the Mahāyāna proposes to bring within ethics.

The fourth-century Yogācāra master Asaṅga, in the *Śīlapaṭala* (戒品 *Jiepin*, "Chapter on Right Conduct") of his *Bodhisattvabhūmi*, answers this question by redrawing the boundary of ethics itself along three functional dimensions. The first dimension (the restraint cluster, *saṃvara-śīla*) inherits the *prātimokṣa* of the seven-fold community and retains it as a structural foundation. The second dimension (the accumulation-of-virtue cluster, *kuśala-dharma-saṃgrāhaka-śīla*) expands the field of ethical concern outward to encompass the entirety of meritorious activity proper to the bodhisattva path. The third dimension (the sattva-welfare cluster, *sattvārtha-kriyā-śīla*) installs the welfare of all sentient beings as the *defining* dimension of the bodhisattva precepts — not as an additional rule but as the criterion by which the precepts count as bodhisattva precepts at all. The structural principle is *not the abandonment of the śrāvaka precepts but their resituation within a higher-order frame*: the Mahāyāna does not discard the *prātimokṣa*; it preserves it and reframes it from above.

One comparative honesty-discipline must be observed at the threshold. The Pāli tradition is not entirely without the *concept* of the precept-keeper bestowing benefit on sentient beings. The *Aṅguttara Nikāya* 8.39 already records, in functional terms, that the one who keeps the precepts "gives the gift of freedom from fear, freedom from enmity, and freedom from oppression" to immeasurable beings, and partakes thereby in immeasurable merit.[^2] This is a functional antecedent — but it does not, in Pāli, become a *structural classification*. The triadic taxonomy of ethical concern (restraint / accumulation-of-virtue / sattva-welfare) as a comprehensive design for the bodhisattva's ethical field is Asaṅga's structural integration, not an inheritance from the Pāli canon.

The thesis of this paper proceeds from that observation. The three clusters are not a list of three rules; they are a three-dimensional architectural design for what bodhisattva ethics is. The structural argument unfolds across the sections that follow: §II tracks the two distinct Chinese-translation paths (T1581 and T1582) and a critical terminological-history finding — that the four-character compound *sānjù jìngjiè* 三聚淨戒, the very name by which the framework is known in East Asian Buddhism, appears in none of the three primary transmission texts; §III analyzes the internal logic of the three dimensions, here labeled *stillness*, *wisdom*, *compassion* (or *止 / 智 / 悲*); §IV stages a three-way dialogue with the principal modern scholars who have engaged this material — Funayama Toru, Michael Zimmermann, and Charles Goodman; §V tracks the Chinese reception from Asaṅga's Sanskrit text through two further Chinese translations to the moment in the Tiantai precept-commentary tradition — the *Tiantai pusa jie shu* (T1812, Mingkuang's Tang-era redaction) — when the three clusters become explicitly mapped onto the three bodies of Buddhahood; §VI synthesizes. The paper's own title, *The Three-Body Architecture*, is named for that mapping. It is not the present paper's interpretive projection: the commentary states the three-cluster ↔ three-body correspondence explicitly, and the title acknowledges that explicit claim as its source.

---

## II. Two Translations, Two Terminological Paths

The strongest textual anchor for the three clusters is a single concise passage near the end of fascicle 5 of T1581 (*Pusa dichi jing* 《菩薩地持經》, translated by Dharmakṣema during the Northern Liang in the early fifth century):

> 「自性等九種戒，當知三戒所攝，所謂律儀戒、攝善法戒、攝眾生戒。又復三種，略說能為菩薩三事：一者律儀戒，能令心住；二者攝善法戒，自成佛法；三者攝眾生戒，成就眾生。是名菩薩一切事。」
> *(Translator's rendering: "The nine precept-types beginning with intrinsic-nature precepts — know that they are subsumed under three precept-types, namely the restraint precept (*lǜyí jiè* 律儀戒), the accumulation-of-virtue precept (*shèshànfǎ jiè* 攝善法戒), and the sattva-welfare precept (*shèzhòngshēng jiè* 攝眾生戒). These three, stated briefly, accomplish three things for the bodhisattva: (1) the restraint precept stabilizes the mind; (2) the accumulation-of-virtue precept perfects one's own Buddha-dharma; (3) the sattva-welfare precept accomplishes the liberation of sentient beings. This is called the totality of the bodhisattva's work.")*[^1]

This single passage performs two operations at once: it establishes the *names* of the three precept-types (restraint / accumulation-of-virtue / sattva-welfare) and assigns each a *function* (stabilizing the mind / perfecting one's own Buddha-dharma / accomplishing sentient beings). The terminology that subsequent East Asian Buddhism uses for the three clusters of bodhisattva precepts — including the very nomenclature this paper is reading — descends almost entirely from Dharmakṣema's translation-bureau diction. T1581 fixed the East Asian vocabulary for more than fifteen hundred years.

### 2.1 A Frequently Overlooked Fact: The *Pusa shan jie jing* Does Not Use This Vocabulary at All

T1582 (*Pusa shan jie jing* 《菩薩善戒經》), translated by Guṇavarman in the Liu Song dynasty in the mid-fifth century, derives from the *same* Sanskrit textual base as T1581 — a circulating recension of the *Bodhisattvabhūmi* (the *Śīlapaṭala* of the *Yogācārabhūmi*'s bodhisattva-section). Yet a comprehensive search through the roughly 1,500 lines of T1582 confirms that the three characteristic terms — *shèlǜyí* 攝律儀, *shèshànfǎ* 攝善法, *shèzhòngshēng* 攝眾生 — *do not appear in T1582 at all*.[^3] T1582 uses a completely different nomenclature path: the first cluster is called simply *jiè* 戒 ("precepts," referring to the seven-fold *prātimokṣa*); the third cluster is called *lìtā jiè* 利他戒 ("altruistic precepts").

The significance of this fact is structural. The standard East Asian three-cluster nomenclature is not the result of "two Chinese translations speaking with one voice." It is the result of T1581's translation-bureau monopolizing the eventual terminological field. T1582's contribution to the fixed vocabulary that subsequent commentators would inherit is essentially nil.

⚠ A philological discipline follows. The dominance of T1581's terminology is a historical outcome of translation-bureau dynamics and reception history, *not* of T1581 being doctrinally more accurate than T1582. This paper's comparative posture juxtaposes both translation paths as parallel transmission events, without treating T1581 as a "standard version."

### 2.2 The Distinctive Contribution of the *Pusa shan jie jing*: The Upāli Dialogue

Although T1582 contributes almost nothing to the *naming* of the three clusters, it preserves a passage that T1581 lacks: a dialogue between the Buddha and Upāli in the opening chapter of the bodhisattva-section, in which the Buddha articulates the fundamental difference between the *śrāvaka* precepts and the bodhisattva precepts.

> 「以是義故，優波離！汝應宣說，聲聞戒急、菩薩戒緩；聲聞戒塞、菩薩戒開；聲聞戒中應說因緣，菩薩戒中則不應說。」
> *(Translator's rendering: "For this reason, Upāli, you should proclaim: the śrāvaka precepts are strict, the bodhisattva precepts are flexible; the śrāvaka precepts are closed, the bodhisattva precepts are open; in the śrāvaka precepts one should state circumstantial conditions, but in the bodhisattva precepts one should not.")*[^4]

The Buddha then illustrates the contrast through the case of greed and aversion:

> 「優波離！若諸菩薩犯如恒河沙等貪，如是菩薩不名毀戒。若犯一瞋因緣毀戒，是名破戒。何以故？優波離！瞋恚之心能捨眾生，貪愛之心能護眾生。若愛眾生，不名煩惱；瞋捨眾生，名重煩惱。」
> *(Translator's rendering: "Upāli, if a bodhisattva commits transgressions motivated by greed in number equal to the sands of the Ganges, such a bodhisattva is not called one who has broken the precepts. If [a bodhisattva] commits even one transgression due to a condition of aversion, this is called breaking the precepts. Why? Because, Upāli, a mind of aversion abandons sentient beings, while a mind of affection-and-care protects sentient beings. To care for sentient beings is not called an affliction; to abandon sentient beings through aversion is called a severe affliction.")*[^5]

This is, by *śrāvaka*-precept standards, an extraordinary inversion. In the *śrāvaka* analysis of the afflictions, greed is the primary defilement and aversion the secondary; here the entire opposition is reversed. The reason is structurally precise. The criterion by which the bodhisattva precepts judge a precept-act is no longer *individual purity* (under which greed would count as a major affliction); it is *whether sentient beings are benefited* (under which aversion, because it "abandons sentient beings," counts as the more severe transgression, and greed, because it "cares for sentient beings," can be tolerated even at enormous quantitative scale).

This passage supplies the strongest primary-text argument in the entire corpus for the *defining status* of the third cluster (the sattva-welfare cluster). The third cluster is not the third item in a list; it is the dimension that determines what the bodhisattva precepts are *as bodhisattva precepts*. The functional division of labor between the two Chinese translations is now visible: T1581 supplies the three-fold naming and the function-triad; T1582 supplies the doctrinal argument for why the third name names a *defining* dimension rather than merely an additional one. The two translations are not redundant parallels; they are complementary anchors.

### 2.3 A Second, Independent Three-Precept Framework in the *Pusa shan jie jing*

It is worth noting that T1582 contains, in fascicle 9, an entirely independent three-precept classification that has nothing to do with the *saṃvara* / *kuśala-dharma-saṃgrāhaka* / *sattvārtha-kriyā* triad:

> 「四波羅蜜攝六波羅蜜有三戒：一者隨戒戒；二者隨心戒；三者隨智戒……離是三戒無菩薩戒，菩薩三戒攝一切戒。」
> *(Translator's rendering: "The four perfections, in subsuming the six perfections, contain three precepts: first, the precept of following discipline; second, the precept of following the mind; third, the precept of following wisdom… Apart from these three precepts there are no bodhisattva precepts; the bodhisattva's three precepts subsume all precepts.")*[^6]

This second framework — *suí jiè* 隨戒 / *suí xīn* 隨心 / *suí zhì* 隨智 (precept-as-discipline / precept-as-meditative-stage / precept-as-wisdom) — is organized along the *threefold training* axis (*śīla* / *samādhi* / *prajñā*), not along the bodhisattva-functional-object axis of the three clusters. ⚠ The two frameworks must not be conflated: they sort the field by different principles. The presence of *two* coexisting three-precept frameworks within T1582 itself is further evidence that the four-character compound *sānjù jìngjiè* had not yet achieved a single canonical form in the transmission-period texts.

### 2.4 The Core Terminological-History Finding: *Sānjù jìngjiè* Is Not a Canonical Compound

This brings the most important philological finding of the present paper into view. The four-character compound *sānjù jìngjiè* 三聚淨戒 — the *name* by which the entire framework is now known throughout East Asian Buddhism — does not appear in any of the three primary transmission texts.[^7]

The closest approximations are:
- **T1581** uses *sān zhǒng jiè jù* 三種戒聚 ("three precept-clusters") just once, in fascicle 5;
- **T1582** uses *jiè* 戒 / *shànfǎ jiè* 善法戒 / *lìtā jiè* 利他戒 (no compound name for the triad as a whole);
- **T1485** (the *Pusa yingluo benyejing* 《菩薩瓔珞本業經》, on which more below) uses *sān shòu mén* 三受門 ("three-gateway reception"), not *sānjù*.

The four-character compound *sānjù jìngjiè* stabilizes only in the Chinese *commentarial* layer — Zhiyi's *Pusa jie yishu* 《菩薩戒義疏》 (T1811, Sui) and Daoxuan's *Sifenlü shanfan buque xingshi chao* 《四分律刪繁補闕行事鈔》 (T1804, Tang) organize the framework as the "three clusters" (*sānjù* 三聚), and the fixed four-character compound is attested in Mingkuang's *Tiantai pusa jie shu* 《天台菩薩戒疏》 (T1812, Tang).[^8]

This finding is *not* a challenge to the authority of the three-cluster framework. Asaṅga's three-fold structure is stably present across all the translations of the *Bodhisattvabhūmi*; the framework itself is fully canonical. The point is more delicate: the *fixed four-character name* by which the framework is known is a downstream consolidation produced by Chinese commentators reading and systematizing the transmission texts. ⚠ The corollary discipline is therefore: one cannot accurately say "the Buddha taught the three clusters of pure precepts in the *Bodhisattvabhūmi*" or "T1581 presents the *sānjù jìngjiè*." The accurate formulation is: "Asaṅga's *Bodhisattvabhūmi* establishes a three-fold structure (*saṃvara-śīla* / *kuśala-dharma-saṃgrāhaka-śīla* / *sattvārtha-kriyā-śīla*), which Chinese commentarial tradition consolidated under the fixed compound *sānjù jìngjiè*."

This is the first philological contribution of the present paper: clarifying the layer to which the compound *sānjù jìngjiè* properly belongs.

---

## III. The Inner Logic of the Three-Dimensional Structure: Stillness, Wisdom, Compassion

Returning to the function-triad of T1581 — *能令心住* / *自成佛法* / *成就眾生* — we may now examine each of the three dimensions in turn.

### 3.1 The Restraint Cluster: The Structural Base of Stillness

The restraint cluster (*saṃvara-śīla* 攝律儀戒) takes "stabilizing the mind" as its functional definition. Its object is *the cessation of harmful conduct*, concretely instantiated by the prohibitions enumerated in the seven-fold *prātimokṣa* (the offenses of killing, stealing, sexual misconduct, false speech, the trading of alcohol, the disparagement of the four-fold community, the self-praise and other-disparagement, and so on — including, in the Mahāyāna codification, the ten major and forty-eight minor bodhisattva precepts).

The crucial structural point here is *continuity rather than rupture*: the restraint cluster does not constitute a new Mahāyāna rule-set replacing the *prātimokṣa*; it *inherits* the *prātimokṣa* and incorporates it as the foundational layer of bodhisattva ethics. T1581 makes this explicit in fascicle 1, where the bodhisattva's four *pārājika*s are said to be transmitted by the Buddhas and bodhisattvas "through immeasurable wisdom and skillful means."[^9] The bodhisattva's four major offenses are *not* a duplicate of the *śrāvaka prātimokṣa*'s four (sexual misconduct, theft, killing, gross false speech); they are layered above it, structurally inheriting it while extending it under the bodhisattva-mind framework. The restraint cluster thus simultaneously preserves the *śrāvaka*-purity dimension and subsumes it under the bodhisattva orientation. This is the doctrinal warrant for the principle that "the Mahāyāna does not abolish the *śrāvaka* precepts but preserves them in a higher structure."

### 3.2 The Accumulation-of-Virtue Cluster: The Maturation of Wisdom

The accumulation-of-virtue cluster (*kuśala-dharma-saṃgrāhaka-śīla* 攝善法戒) takes "perfecting one's own Buddha-dharma" as its functional definition. The phrase *shèshànfǎ* 攝善法 means literally "embracing all virtuous *dharma*s" — the totality of meritorious activity that the bodhisattva path comprises, including the three trainings (*śruta* / *cintā* / *bhāvanā* — hearing, reflection, meditative cultivation), the six perfections, and every form of positive cultivation that constitutes the bodhisattva's path.

The structural significance of this cluster is that the field of ethical concern is no longer restricted to *what one must not do* (the *saṃvara* dimension); it now extends to *what one must do* — every form of beneficial activity capable of accomplishing the path of awakening is brought within the field of the precepts. At the *abhidharma* layer, Sthiramati's *Mahāyānābhidharmasamuccayavyākhyā* (T1606, translated by Xuanzang), commenting on the *Bodhisattvabhūmi*, makes this connection explicit:

> 「如〈菩薩地〉說。攝善法戒者，謂六波羅蜜多若依純雜相資助說……」
> *(Translator's rendering: "As the Bodhisattvabhūmi states: the accumulation-of-virtue precept refers to the six pāramitās, whether taken in their pure form or in their mutual combination…")*[^10]

The Tang-era Chinese *abhidharma* tradition thus reads the second cluster as structurally coextensive with the entire architecture of the six perfections. The implication is far-reaching: the totality of bodhisattva practice falls within the precepts' field of concern.

### 3.3 The Sattva-Welfare Cluster: Compassion as Defining Dimension

The sattva-welfare cluster (*sattvārtha-kriyā-śīla*) takes "accomplishing sentient beings" as its functional definition. The Chinese terminology here requires a strict disambiguation: T1581 (Dharmakṣema) renders this term *shèzhòngshēng jiè* 攝眾生戒; T1579 (Xuanzang's translation of the *Yogācārabhūmi*) renders it *ráoyì yǒuqíng jiè* 饒益有情戒. ⚠ The two Chinese terms correspond to the *same* Sanskrit concept (*sattvārtha-kriyā-śīla*) but belong to different translation-bureau vocabularies and must not be used interchangeably within a single citation. When citing T1581, the correct rendering is *shèzhòngshēng*; when citing T1579, the correct rendering is *ráoyì yǒuqíng*.[^11]

Returning to the doctrinal substance: the *defining* status of this third cluster has already been established in §2.2 via the Upāli dialogue. The third cluster is not an additional category alongside the first two; it is the dimension that *defines* the bodhisattva precepts as bodhisattva precepts. The judgment-standard has shifted: from individual purity to actual benefit conferred upon sentient beings.

The internal logic of the three dimensions is now visible:

- **Restraint cluster** = *stillness* (the *protective* dimension) — taking the *śrāvaka prātimokṣa* as structural foundation;
- **Accumulation-of-virtue cluster** = *wisdom* (the *maturational* dimension) — encompassing the entirety of bodhisattva practice;
- **Sattva-welfare cluster** = *compassion* (the *defining* dimension) — installing the welfare of beings as the ultimate criterion.

The three clusters are not three parallel rules; they are three *functional dimensions* of a single ethical field. Stillness, wisdom, and compassion interpenetrate; the absence of any one would prevent the structure from being recognizably bodhisattva ethics at all.

### 3.4 Ordering Variation: Another Interpretive Position

The standard ordering in T1581 and T1579 — restraint → accumulation-of-virtue → sattva-welfare — is preserved consistently in those two translation paths. Yet a third translation, T1485 (the *Pusa yingluo benyejing* 《菩薩瓔珞本業經》, translated by Zhu Fonian during the Yao Qin in the late fourth century), preserves a *different* ordering:

> 「今為諸菩薩結一切戒根本，所謂三受門：攝善法戒，所謂八萬四千法門；攝眾生戒，所謂慈悲喜捨化及一切眾生皆得安樂；攝律儀戒，所謂十波羅夷。」
> *(Translator's rendering: "For all bodhisattvas, I now establish the root of all precepts, namely the three-gateway reception: the accumulation-of-virtue precept, that is, the eighty-four thousand dharma-gateways; the sattva-welfare precept, that is, transforming all sentient beings through the four immeasurables so that they all attain happiness; the restraint precept, that is, the ten pārājikas.")*[^12]

The comparison may be summarized:

| Translation | First | Second | Third |
|---|---|---|---|
| T1581 (Dharmakṣema, Northern Liang, 5c) | restraint | accumulation-of-virtue | sattva-welfare |
| T1579 (Xuanzang, Tang, 7c) | restraint | accumulation-of-virtue | sattva-benefit (*ráoyì yǒuqíng*) |
| T1485 (Zhu Fonian, Yao Qin, late 4c) | accumulation-of-virtue | sattva-welfare | restraint |

T1485 thus places the restraint cluster *last* and the accumulation-of-virtue cluster *first*. During the transmission window of the fourth to seventh centuries, when the framework had not yet stabilized into a single canonical form on Chinese soil, T1485 represents a distinct interpretive emphasis — one in which the eighty-four thousand dharma-gateways subsumed under the accumulation-of-virtue cluster constitutes the primary weight, rather than the linear "restraint as base" ordering of T1581 and T1579.

The discipline here must be precise. T1485's ordering is *not* to be characterized as "an immature version" or "an error in sequence." Such an evaluation would presuppose that the T1581/T1579 ordering is uniquely correct. The descriptive characterization that this paper adopts is that T1485 represents a *distinct interpretive position* — one that foregrounds the accumulation-of-virtue cluster as primary. The "restraint as base" reading (T1581/T1579 lineage) is one interpretive option among several, not the only legitimate one.

This ordering variation matters for the paper's argument. Although Asaṅga's three-fold structure is stable across all Chinese transmissions of the *Bodhisattvabhūmi* — three dimensions, distinguished by functional object — *which* dimension is read as the primary weight admits genuine interpretive variation. This paper adopts a descriptive stance throughout, refraining from any judgment of relative ordering merit.

---

## IV. Three Scholarly Perspectives: Funayama, Zimmermann, Goodman

The three-cluster framework has been engaged in modern scholarship from three distinct methodological positions: a transmission-history axis (Funayama Toru), a Sanskrit-textual axis (Michael Zimmermann), and an ethical-philosophical axis (Charles Goodman). The present paper situates itself in concede-then-extend dialogue with each.

### 4.1 Funayama Toru: The Transmission-History Axis

Funayama Toru of Kyoto University has produced the most systematic body of research on the reception of T1581 and T1582 in Chinese Buddhism.[^13] His work tracks the circulation of the two translations through the fifth to seventh centuries, demonstrating that the *Pusa shan jie jing* (T1582) and the *Pusa dichi jing* (T1581) circulated *in parallel* until Xuanzang's seventh-century translation of the full *Yogācārabhūmi* (T1579) gradually displaced T1582 from its parallel position.

*Concession.* Funayama's transmission-history research is the most detailed historical tracking available, in any language, of how T1581 and T1582 fared on Chinese soil. This paper benefits directly from his philological frame: the historical dominance of T1581's terminology is to be explained by translation-bureau dynamics and reception patterns, not by a thesis of doctrinal superiority.

*Extension.* Transmission-history research, by its nature, presupposes that *what is being transmitted* has already been identified at the structural level. Funayama tracks how the three clusters traveled through Chinese Buddhism; the present paper asks what the structural design *is* — what makes the three clusters the kind of object whose transmission history is worth tracking in the first place. The structural analysis offered here is what Funayama's transmission-history research takes as its object.

### 4.2 Michael Zimmermann: The Sanskrit-Textual Axis

Michael Zimmermann of the University of Hamburg has produced the most extensive English-language Sanskrit-textual study of the *Śīlapaṭala* (the bodhisattva-precept chapter of the *Bodhisattvabhūmi*) in his contribution to volume 75 of the Harvard Oriental Series.[^14] His work confirms the three-fold structure as an internal integration within the Yogācāra system and clarifies the Sanskrit terminology.

*Concession.* Zimmermann's work is the closest the existing Western scholarship comes to direct analysis of the three-cluster structure. The Sanskrit-Chinese correspondences this paper relies upon — *saṃvara-śīla* for *shèlǜyí jiè*, *kuśala-dharma-saṃgrāhaka-śīla* for *shèshànfǎ jiè*, *sattvārtha-kriyā-śīla* for both *shèzhòngshēng jiè* (T1581) and *ráoyì yǒuqíng jiè* (T1579) — depend on Zimmermann's textual reconstruction.

*Extension.* Zimmermann's analysis remains at the layer of the Sanskrit *Bodhisattvabhūmi* itself. The present paper extends the analysis along two further axes that Zimmermann's textual work does not pursue: (i) the comparative analysis of how the *two* Chinese translations (T1581, T1582) render the same Sanskrit source through non-overlapping nomenclatures, and (ii) the reception trajectory through which the Tiantai precept-commentary tradition (Mingkuang's *Tiantai pusa jie shu*, T1812) transforms the framework by mapping the three clusters onto the three bodies of Buddhahood (§V below). The structural argument is thereby moved from textual reconstruction outward into transmission and reception.

### 4.3 Charles Goodman: The Ethical-Philosophical Axis

Charles Goodman, in *Consequences of Compassion*, reads Asaṅga's bodhisattva-precept framework through the lens of Western consequentialist ethics, with particular attention to the "openness" of the bodhisattva precepts relative to the *śrāvaka* precepts (which he sees as a key marker of consequentialist structure).[^15]

*Concession.* Goodman's reading does correctly identify a real structural feature: the bodhisattva precepts *are* open relative to the *śrāvaka* precepts in ways the Upāli dialogue makes vivid (§2.2 above). T1582's reversal of the standard greed/aversion hierarchy is genuinely what consequentialist accounts of ethics would predict — a benefit-oriented criterion replacing a purity-oriented criterion.

*Extension.* Yet Goodman's framework captures the *phenomenon* at the expense of the *structure*. T1582's "benefit-rather-than-purity as the criterion of the bodhisattva precepts" is not the same as a consequentialist claim that "the rightness of an act depends on its consequences." The bodhisattva-precept criterion is the *sattva-welfare cluster operating as the defining dimension that constitutes the other two clusters as bodhisattva precepts*. Sattva-welfare is not the third item in a list of three items; it is the dimension that runs through restraint and accumulation-of-virtue and makes them bodhisattva precepts rather than *śrāvaka* precepts. The consequentialist reading captures the openness; it does not capture the *integration*. The paper's argument is that the three clusters constitute a structural design in which the third dimension defines the field as such — a structural claim that consequentialist categories cannot capture without remainder.

### 4.4 Honest Disclosure: Sparse Western Coverage

A frank disclosure is necessary at this point. *No English-language monograph exists with the three clusters of pure precepts as its primary structural topic.* Studies of the bodhisattva precepts in Western scholarship have approached the material primarily through three angles: ordination ritual (Paul Groner and others on the East Asian ordination-platform tradition), the *Fanwang jing* lineage (a partial body of work, mostly in article form), and ethical philosophy (Goodman, Mark Siderits, and others). The structural-invention question — what *is* the three-cluster framework as a structural design — has received remarkably little dedicated attention.[^16]

This is the corrective position of the present paper: to fill a genuine structural-analytic gap in Western-language scholarship on Mahāyāna ethics, and to lay groundwork for further work along this axis.

---

## V. Chinese Reception: From Three Precept-Types to a Three-Body Architecture

Asaṅga's three-fold structure does not stabilize on Chinese soil in a single moment. It moves through several stages: the initial fifth-century translation by Dharmakṣema (T1581), the parallel translation by Guṇavarman (T1582), the seventh-century retranslation under Xuanzang (T1579), the *abhidharma*-layer endorsement in Sthiramati's commentary (T1606), and — its decisive moment — the Tang-era mapping in the *Tiantai pusa jie shu* (T1812, Mingkuang's redaction) onto the three bodies of Buddhahood.

### 5.1 Xuanzang's Terminological Revision

When Xuanzang retranslated the full *Yogācārabhūmi* in the seventh century (T1579), the relevant passage on the three clusters of bodhisattva precepts (fascicles 40–41) became:

> 「謂律儀戒能安住其心，攝善法戒能成熟自佛法，饒益有情戒能成熟有情。」
> *(Translator's rendering: "The restraint precept stabilizes the mind; the accumulation-of-virtue precept ripens one's own Buddha-dharma; the sattva-benefit precept (ráoyì yǒuqíng jiè) ripens sentient beings.")*[^17]

The substantive structure is identical to T1581's formulation. The lexical change occurs only in the third term: Dharmakṣema's *shèzhòngshēng jiè* 攝眾生戒 becomes Xuanzang's *ráoyì yǒuqíng jiè* 饒益有情戒. Both render the same Sanskrit *sattvārtha-kriyā-śīla*, but the diction belongs to different translation-bureau conventions. *Ráoyì yǒuqíng* tracks the literal Sanskrit *sattvārtha* ("welfare of sentient beings") more closely; *shèzhòngshēng* uses the more idiomatic Chinese sense of "embracing sentient beings." ⚠ When citing T1579, the appropriate rendering is *ráoyì yǒuqíng jiè*; when citing T1581, *shèzhòngshēng jiè*; the two are not interchangeable within a single passage.

A philological note worth recording: the *Mahāprajñāpāramitā-śāstra* (T1509, attributed to Nāgārjuna and translated by Kumārajīva in the early fifth century) — by far the largest Mādhyamika-tradition treatise translated into Chinese — *does not contain the three-cluster framework at all*. A full-text search of T25 No. 1509 returns *zero occurrences* of the *saṃvara-śīla* / *kuśala-dharma-saṃgrāhaka-śīla* / *sattvārtha-kriyā-śīla* triad. T1509 employs a different three-precept taxonomy entirely (*saṃvara-śīla* / dhyāna-generated *śīla* / undefiled *śīla*), organized by *stage of the path* rather than by *functional object*.[^18] This is an independent line within the Chinese treatise corpus and should not be invoked as supporting evidence for the three-cluster framework. T1509's *śīla-pāramitā* discussion is the topic of the next paper in this volume (P13).

Sthiramati's *Mahāyānābhidharmasamuccayavyākhyā* (T1606, translated by Xuanzang), by contrast, explicitly invokes the *Bodhisattvabhūmi* in confirming the three-cluster authority, as cited in §3.2 above. This is the clearest *abhidharma*-layer endorsement of the three-cluster framework within the Chinese treatise corpus.

### 5.2 The *Tiantai pusa jie shu*: The Three-Cluster ↔ Three-Body Mapping

The decisive moment in the Chinese reception of the three clusters is the *Tiantai pusa jie shu* 《天台菩薩戒疏》 (T1812), the Tang-era redaction of Tiantai bodhisattva-precept commentary by Mingkuang 明曠. Using the single precept of "non-killing" as a worked example, the commentary maps the three clusters onto the three bodies of Buddhahood (*trikāya*) as a causal correspondence:

- The **restraint cluster** (cessation of harmful conduct, the body fully separated from defilement) → cause of the *dharmakāya* (法身因)
- The **accumulation-of-virtue cluster** (the disciplined practice of virtue, the realization of dharma-knowledge) → cause of the *saṃbhogakāya* (報身因)
- The **sattva-welfare cluster** (the combined cessation of harm, practice of virtue, and compassionate benefit toward beings) → cause of the *nirmāṇakāya* (應身因)

The commentary then states explicitly:

> 「故知戒戒三聚互融，三觀三身相即。三聚三身，既無優劣，四十八輕十重，等持心性，寧有淺深？」
> *(Translator's rendering: "Therefore know: in every precept the three clusters interpenetrate; the three contemplations and the three bodies are mutually identical. Between the three clusters and the three bodies there is neither superiority nor inferiority. Among the forty-eight minor and ten major precepts, all are equally held within mind-nature — how could there be shallow and deep?")*[^19]

This is the most consequential structural pronouncement in the entire Chinese reception of the three clusters. Several points must be emphasized. The three-cluster ↔ three-body mapping is *the commentary's explicit claim*, articulated in the *Tiantai pusa jie shu* itself; *it is not this paper's interpretive projection*. The paper's title, *The Three-Body Architecture*, is named *for that mapping* — not invented by the present author. (This sets P12's authorial posture apart from that of P11, where the Vesālī ↔ *karma-seng-bheda* mapping was P11's own interpretive completion of a typology Daoxuan left open. P12's three-body mapping requires no such interpretive completion: the commentary states the correspondence explicitly.)

The significance of this mapping is that it transforms the three clusters from a *taxonomic classification of precepts* into a *structural map of the bodhisattva path*. The three clusters are no longer merely the three directions one orients toward when receiving precepts; they are the causal grounds, in the precept-practice phase, of the three bodies of Buddhahood. Every act of keeping a bodhisattva precept is thereby an act of three-directional cultivation toward the three bodies. This is the Tiantai tradition's distinctive structural contribution to the precept domain.

### 5.3 The Commentary's Second Mapping: The Four Vast Vows ↔ The Three Clusters

In the same *Tiantai pusa jie shu*, the commentary performs an additional mapping that further integrates the three-cluster framework into the bodhisattva path's larger architecture: the correspondence between the *four vast vows* and the three clusters.

> 「不出四弘三聚。成道知法即攝善法，誓斷煩惱即攝律儀，願度眾生即攝眾生。」
> *(Translator's rendering: "They do not go beyond the four vast vows and the three clusters. Perfecting the path and knowing the dharma is the accumulation-of-virtue cluster; vowing to sever afflictions is the restraint cluster; vowing to liberate sentient beings is the sattva-welfare cluster.")*[^20]

The correspondence may be tabulated:

| Four Vast Vows | Three Clusters |
|---|---|
| "Sentient beings without bounds, I vow to liberate" | sattva-welfare cluster |
| "Afflictions without end, I vow to sever" | restraint cluster |
| "Dharma-gateways without measure, I vow to learn" | accumulation-of-virtue cluster |

The four vast vows — "Sentient beings without bounds, I vow to liberate; afflictions without end, I vow to sever; dharma-gateways without measure, I vow to learn; the unsurpassed Buddha-path, I vow to accomplish" — are the standard formula of bodhisattva aspiration in East Asian Buddhism, still recited in daily liturgy today. The commentary's mapping reveals that the four vast vows and the three clusters of pure precepts are not two parallel devotional formulas; they are the *same bodhisattva aspiration* articulated at two different layers — vow-power (the four vows) and precept-substance (the three clusters) — that mutually pervade.

With this, the historical sequence is complete: from Asaṅga's structural invention in fourth-century India, through Dharmakṣema's translation-bureau in the fifth-century Northern Liang that fixed the terminology, through Xuanzang's seventh-century retranslation that revised the third cluster's name, to the Tang-era *trikāya* mapping of the *Tiantai pusa jie shu* that transformed precept-classification into a structural map of the entire bodhisattva path.

---

## VI. Conclusion: The Three-Dimensional Framework as a Map of the Bodhisattva Path

### 6.1 Synthesis

The three clusters of pure precepts are not a list of three rules. They are a three-dimensional architectural design for bodhisattva ethics. The restraint cluster inherits the *prātimokṣa* of the seven-fold community as the structural foundation for the cessation of harmful conduct. The accumulation-of-virtue cluster extends the field of ethical concern to the entirety of bodhisattva practice as the dimension of wisdom's maturation. The sattva-welfare cluster installs the benefit of all sentient beings as the *defining* dimension by which the bodhisattva precepts count as bodhisattva precepts at all. The three dimensions interpenetrate; the absence of any one would render the structure no longer bodhisattva ethics.

Two philological clarifications have been advanced. First, the four-character compound *sānjù jìngjiè* 三聚淨戒 itself appears in none of the three primary transmission texts (T1581, T1582, T1485). It is a downstream consolidation produced by the Chinese commentarial layer. The framework's authority is in Asaṅga; the framework's *name* is in the Chinese commentaries. Second, T1581 and T1582 — though translated from the same Sanskrit *Bodhisattvabhūmi* — employ almost entirely non-overlapping nomenclatures. The terminological dominance of T1581 in subsequent East Asian Buddhism is a historical outcome of translation-bureau dynamics and reception patterns, not a measure of doctrinal accuracy. T1582's contribution is not the names but the strongest primary-text argument — the Upāli dialogue — for the defining status of the sattva-welfare cluster.

Beyond these philological clarifications, the paper's structural argument is completed by the explicit mapping in the *Tiantai pusa jie shu* (T1812) of the three clusters onto the three bodies of Buddhahood. With that mapping, the three clusters become not merely a precept-classification but a structural map of the bodhisattva path as a whole — a path in which stillness, wisdom, and compassion are the three dimensions whose cultivation grounds, respectively, the *dharmakāya*, the *saṃbhogakāya*, and the *nirmāṇakāya*.

### 6.2 Open Questions

The present paper addresses the structural question — *what are the three clusters of pure precepts?* — but leaves two further directions for subsequent inquiry. First, the three clusters establish the structural design of the bodhisattva precepts; but how do bodhisattva precepts become a *pāramitā* (a perfection — that is, a completed virtue of the bodhisattva path)? This is the central question of the *śīla-pāramitā* discussion in T1509 (*Mahāprajñāpāramitā-śāstra*), which the next paper in this volume (P13) takes up. Second, the three clusters serve as the framing structure within which the more detailed nine-fold taxonomy of precepts (the *navavidha-śīla* of the *Bodhisattvabhūmi*) operates; how the nine subdivisions unfold within the three-fold frame is the topic of the paper following (P14, focused on T1579). The present paper has supplied the master architecture; P13 and P14 will treat, respectively, the *pāramitā*-completion question and the nine-fold subdivision question.

P11 closed Part III of this volume by reading the *śrāvaka*-internal historical crises of boundary-renegotiation. The present paper, P12, opens Part IV by showing how, in the Mahāyāna turn, the *concept of boundary itself* is redrawn — from the procedural boundary of an in-*saṅgha* deliberation to the three-dimensional architectural design of bodhisattva ethics. The three clusters are the structural pivot of that redrawing. They permit the bodhisattva precepts to retain the *śrāvaka prātimokṣa* as their structural base while installing the welfare of all sentient beings as the defining dimension. The entirety of the bodhisattva path is thereby brought, through the second cluster, within the field of ethical concern; the entirety of sentient existence is brought, through the third cluster, within the field of ethical concern's *ultimate criterion*.

---

## Footnotes

[^1]: *Pusa dichi jing* 《菩薩地持經》, fascicle 5, translated by Dharmakṣema (Northern Liang, fifth century). *Taishō* vol. 30, T1581, p. 0918b01–06.

[^2]: *Aṅguttara Nikāya* 8.39 (*Abhisanda Sutta*, "The Eight Streams of Merit"). PTS edition; SuttaCentral translation by Bhikkhu Sujato (CC0). The functional formulation here — that the precept-keeper bestows immeasurable freedom from fear, enmity, and oppression upon innumerable beings, partaking thereby in immeasurable merit — supplies the Pāli-canonical functional antecedent for what becomes, in Asaṅga's *Bodhisattvabhūmi*, the sattva-welfare cluster. *Philological note*: the Pāli canon does not, however, integrate this functional content into a *taxonomic structural classification* of the precepts. The three-fold structural taxonomy (restraint / accumulation-of-virtue / sattva-welfare) as a comprehensive design is Asaṅga's structural integration, not a Pāli inheritance.

[^3]: *Philological confirmation*: a full-text search of T1582 *Pusa shan jie jing* 《菩薩善戒經》 (Guṇavarman, Liu Song, mid-fifth century, *Taishō* vol. 30) returns *zero occurrences* of the three Chinese terms *shèlǜyí* 攝律儀, *shèshànfǎ* 攝善法, and *shèzhòngshēng* 攝眾生 used as the names of precept-clusters. T1582 employs the alternative nomenclature: the first cluster as simply *jiè* 戒 (referring to the seven-fold *prātimokṣa*); the third cluster as *lìtā jiè* 利他戒 ("altruistic precepts"). The naming path is wholly distinct from T1581.

[^4]: *Pusa shan jie jing* 《菩薩善戒經》, fascicle 1, "Preface to the Bodhisattva-Ground," translated by Guṇavarman (Liu Song, fifth century). *Taishō* vol. 30, T1582, p. 0961c23–26.

[^5]: *Pusa shan jie jing* 《菩薩善戒經》， fascicle 1, T30n1582_p0962a14–20, translated by Guṇavarman.

[^6]: *Pusa shan jie jing* 《菩薩善戒經》, fascicle 9, "Stage of Final Conduct," T30n1582_p1009b07–11, translated by Guṇavarman. This second three-precept framework (*suí jiè* / *suí xīn* / *suí zhì* — precept-following / mind-following / wisdom-following) is unique to T1582 and is organized along the *threefold-training* axis (*śīla* / *samādhi* / *prajñā*), not along the bodhisattva-functional-object axis of the three clusters. The two frameworks are not interchangeable.

[^7]: *Terminological-history note (D-P12-A1)*: The four-character compound *sānjù jìngjiè* 三聚淨戒 does not appear in any of the three primary transmission texts (T1581 *Pusa dichi jing*, T1582 *Pusa shan jie jing*, T1485 *Pusa yingluo benyejing*). T1581 fascicle 5 uses *sān zhǒng jiè jù* 三種戒聚 ("three precept-clusters") just once (T30n1581_p0918b01); the consolidated label *sānjù jìngjiè* solidifies only in the Chinese commentarial layer: Zhiyi's *Pusa jie yishu* (T1811) and Daoxuan's *Sifenlü shanfan buque xingshi chao* (T1804) organize the framework as the "three clusters" (三聚, three occurrences each) without the full four-character compound, which is attested in Mingkuang's *Tiantai pusa jie shu* (T1812, Tang; two occurrences, CBETA-verified). The corollary discipline is that the framework's authority resides in the Sanskrit *Bodhisattvabhūmi* and its three Chinese translations; the four-character *name* is a Chinese-commentarial consolidation.

[^8]: Zhiyi 智顗 (538–597), *Pusa jie yishu* 《菩薩戒義疏》, *Taishō* vol. 40, T1811; Daoxuan 道宣 (596–667), *Sifenlü shanfan buque xingshi chao* 《四分律刪繁補闕行事鈔》, *Taishō* vol. 40, T1804. These two commentaries systematize the framework under the "three clusters" heading; the fixed four-character compound *sānjù jìngjiè* is attested in Mingkuang's *Tiantai pusa jie shu* 《天台菩薩戒疏》 (T1812, Tang).

[^9]: *Pusa dichi jing*, fascicle 1, T30n1581_p0910c, translated by Dharmakṣema. The "four *pārājika*s" mentioned in this passage are the bodhisattva four major offenses, *not* the *śrāvaka prātimokṣa*'s four (sexual misconduct, theft, killing, gross false speech). The bodhisattva-precept ten-major / forty-eight-minor schema layered on top of, rather than replacing, the *śrāvaka prātimokṣa* base.

[^10]: *Mahāyānābhidharmasamuccayavyākhyā* 《大乘阿毘達磨雜集論》, fascicle 12, composed by Sthiramati (sixth century), translated by Xuanzang (Tang dynasty). *Taishō* vol. 31, T1606, p. 0750c09.

[^11]: *Terminology non-interchangeability discipline (D-P12-A3)*: T1581 (Dharmakṣema) third cluster = *shèzhòngshēng jiè* 攝眾生戒 (= *sattvārtha-kriyā-śīla*); T1579 (Xuanzang) third cluster = *ráoyì yǒuqíng jiè* 饒益有情戒 (= same Sanskrit). The two Chinese terms render the *same* Sanskrit concept but belong to different translation-bureau vocabularies. When citing T1581, *shèzhòngshēng* is correct; when citing T1579, *ráoyì yǒuqíng* is correct. The two must not be substituted for one another within a single passage.

[^12]: *Pusa yingluo benyejing* 《菩薩瓔珞本業經》, fascicle 2, translated by Zhu Fonian 竺佛念 (Yao Qin, late fourth century). *Taishō* vol. 24, T1485, p. 1020b29–c03.

[^13]: Funayama Toru 船山徹, principal contributions on the reception of T1581 and T1582 in fifth- to seventh-century Chinese Buddhism, in *Journal of Asian History* vol. 38, no. 2 (2004), and in related articles in *Tōhō gakuhō* (《東方學報》, the Kyoto University Humanities Research Bulletin). *Paraphrase summary; no verbatim accessed* (no open-access English-language full text confirmed).

[^14]: Michael Zimmermann, "The Chapter on Right Conduct in the *Bodhisattvabhūmi*," in *Harvard Oriental Series* vol. 75 (Cambridge, MA: Harvard University Press, 2013), pp. 872–883. *Paraphrase summary; no verbatim accessed* (paywalled).

[^15]: Charles Goodman, *Consequences of Compassion: An Interpretation and Defense of Buddhist Ethics* (New York: Oxford University Press, 2009), chapter 4. *Paraphrase summary; no verbatim accessed* (paywalled).

[^16]: *Honest disclosure (D-P12 sparse-coverage discipline)*: No English-language monograph exists with the three clusters of pure precepts as its primary structural topic. Western scholarship on the bodhisattva precepts has proceeded along three principal axes — ordination ritual and lineage (Paul Groner and others on the East Asian ordination-platform tradition), the *Fanwang jing* tradition (a partial body of article-form work), and ethical philosophy (Goodman, Mark Siderits, and others). The structural-design question — what is the three-cluster framework *as* a structural design — has received remarkably little dedicated treatment. The present paper situates itself as addressing this gap.

[^17]: *Yogācārabhūmi* 《瑜伽師地論》， fascicle 40, translated by Xuanzang (Tang, seventh century). *Taishō* vol. 30, T1579, p. 0523a02–03.

[^18]: *Philological note (T1509 absence record)*: T1509 *Mahāprajñāpāramitā-śāstra* 《大智度論》 (attributed to Nāgārjuna; translated by Kumārajīva in the Later Qin, early fifth century) contains *zero occurrences* of the *saṃvara-śīla* / *kuśala-dharma-saṃgrāhaka-śīla* / *sattvārtha-kriyā-śīla* framework (confirmed by full-text search of T25 No. 1509). T1509 employs a distinct three-precept taxonomy organized by stage of the path — *saṃvara-śīla*, dhyāna-generated *śīla*, and undefiled *śīla* — rather than by functional object. T1509 should *not* be invoked as a supporting witness for the three-cluster framework. The *śīla-pāramitā* dimension of T1509 is the topic of paper P13 of this series.

[^19]: Mingkuang 明曠, *Tiantai pusa jie shu* 《天台菩薩戒疏》, *Taishō* vol. 40, T1812, T40n1812_p0584b07–10. The three-cluster ↔ three-body mapping articulated in this passage — *dharmakāya* cause via the restraint cluster, *saṃbhogakāya* cause via the accumulation-of-virtue cluster, *nirmāṇakāya* cause via the sattva-welfare cluster — is the commentary's explicit doctrinal claim, not the present paper's interpretive projection.

[^20]: Mingkuang 明曠, *Tiantai pusa jie shu* 《天台菩薩戒疏》, *Taishō* vol. 40, T1812, T40n1812_p0580c22–25. The four-vast-vows ↔ three-clusters mapping articulated here, followed immediately by the three-body assignment of the three clusters (如持不殺…法身因也…).

---

## References

### Primary canonical sources (CBETA Electronic Tripiṭaka 2024)

- *Pusa dichi jing* 《菩薩地持經》 ("Sūtra of the Bodhisattva's Stages of Sustaining"). Translated by Dharmakṣema (Northern Liang, 414–426). *Taishō* vol. 30, T1581. Principal anchor for the three-cluster naming and the function-triad (fascicle 5, T30n1581_p0918b01–06).
- *Pusa shan jie jing* 《菩薩善戒經》 ("Sūtra of the Bodhisattva's Good Precepts"). Translated by Guṇavarman (Liu Song, 431). *Taishō* vol. 30, T1582. Principal anchor for the Upāli dialogue (fascicle 1, T30n1582_p0961c23–p0962a20) and the alternative three-precept framework (fascicle 9, T30n1582_p1009b07–11).
- *Pusa yingluo benyejing* 《菩薩瓔珞本業經》 ("Sūtra of the Original Acts that Adorn the Bodhisattva"). Translated by Zhu Fonian (Yao Qin, late fourth century). *Taishō* vol. 24, T1485. Source of the alternative ordering of the three precept-categories (fascicle 2, T24n1485_p1020b29–c03).
- *Yogācārabhūmi* 《瑜伽師地論》 ("Treatise on the Stages of Yoga Practice"). Attributed to Maitreya; translated by Xuanzang (Tang, 646–648). *Taishō* vol. 30, T1579. Source of the Xuanzang terminology *ráoyì yǒuqíng jiè* 饒益有情戒 (fascicle 40, T30n1579_p0523a02–03).
- *Mahāyānābhidharmasamuccayavyākhyā* 《大乘阿毘達磨雜集論》 ("Mahāyāna Compendium of Abhidharma with Commentary"). Composed by Sthiramati (sixth century); translated by Xuanzang (Tang). *Taishō* vol. 31, T1606. Source of the *abhidharma*-layer endorsement of the three-cluster framework (fascicle 12, T31n1606_p0750c09).
- *Tiantai pusa jie shu* 《天台菩薩戒疏》 ("Tiantai Commentary on the Bodhisattva Precepts"). Redacted (*shanbu* 刪補) by Mingkuang (Tang). *Taishō* vol. 40, T1812. Source of the three-cluster ↔ three-body explicit mapping (T40n1812_p0584b07–10) and the four-vast-vows ↔ three-clusters mapping (T40n1812_p0580c22–25).
- *Pusa jie yishu* 《菩薩戒義疏》 ("Commentary on the Meaning of the Bodhisattva Precepts"). Attributed to Zhiyi (538–597), recorded by Guanding. *Taishō* vol. 40, T1811.

### Secondary canonical sources

- *Sifenlü shanfan buque xingshi chao* 《四分律刪繁補闕行事鈔》 ("Compendium on Procedures and Activities Drawn from the *Sifenlü*"). Composed by Daoxuan (596–667). *Taishō* vol. 40, T1804. Source for the terminological consolidation of *sānjù jìngjiè* in the Chinese commentarial layer.
- *Mahāprajñāpāramitā-śāstra* 《大智度論》 ("Treatise on the Great Perfection of Wisdom"). Attributed to Nāgārjuna; translated by Kumārajīva (Later Qin, 402–406). *Taishō* vol. 25, T1509. *Explicitly recorded as absent from the three-cluster framework*; cited here for the philological-absence record only. Treated as primary source for P13 (*śīla-pāramitā*) of this volume.
- *Aṅguttara Nikāya* 8.39 (*Abhisanda Sutta*). PTS edition; SuttaCentral ID an8.39; translation by Bhikkhu Sujato (CC0). Cited as Pāli-canonical functional antecedent.

### Modern scholarship

- Funayama, Toru 船山徹. "The Acceptance of the *Bodhisattvabhūmi* in Six Dynasties to Tang Chinese Buddhism" and related articles. *Journal of Asian History* 38, no. 2 (2004); related work in *Tōhō gakuhō* 《東方學報》, Kyoto University Humanities Research Bulletin. *Paraphrase summary; no verbatim accessed* (no open-access English-language full text confirmed).
- Goodman, Charles. *Consequences of Compassion: An Interpretation and Defense of Buddhist Ethics*. New York: Oxford University Press, 2009. Chapter 4 ("The Ethics of the Mahāyāna Sūtras") on the bodhisattva precepts' relative openness. *Paraphrase summary; no verbatim accessed* (paywalled).
- Zimmermann, Michael. "The Chapter on Right Conduct in the *Bodhisattvabhūmi*." In *Harvard Oriental Series* vol. 75. Cambridge, MA: Harvard University Press, 2013, pp. 872–883. Sanskrit-textual reconstruction of the *Śīlapaṭala*. *Paraphrase summary; no verbatim accessed* (paywalled).

### Sanskrit-Chinese terminology table

| Chinese | Sanskrit | Source translation |
|---|---|---|
| 律儀戒 *lǜyí jiè* | *saṃvara-śīla* | common to T1581, T1579 |
| 攝善法戒 *shèshànfǎ jiè* | *kuśala-dharma-saṃgrāhaka-śīla* | common to T1581, T1579 |
| 攝眾生戒 *shèzhòngshēng jiè* | *sattvārtha-kriyā-śīla* | T1581 (Dharmakṣema diction) |
| 饒益有情戒 *ráoyì yǒuqíng jiè* | *sattvārtha-kriyā-śīla* | T1579 (Xuanzang diction) |
| 利他戒 *lìtā jiè* | (functional equivalent) | T1582 (Guṇavarman diction) |

---

## Cross-References

### Part IV internal (this paper's volume position)

- **P11** (*Two Sangha Crises and the Mechanism of Boundary-Renegotiation*) — Part III's concluding paper, treating the diachronic *śrāvaka*-internal crises of Vesālī and Devadatta as paradigms of procedural boundary-renegotiation and foundational-posture conflict. The present paper's §I opens Part IV by reframing the boundary question itself: from the procedural-boundary question that P11 closes upon to the three-dimensional architectural design that P12 advances.
- **P13** (*Śīla-Pāramitā in the Mahāprajñāpāramitā-śāstra*) — the present paper's structural argument establishes the three-cluster framework; P13 takes up the question of how the bodhisattva precepts become a *pāramitā* (a perfection of the bodhisattva path), the central concern of T1509 *Mahāprajñāpāramitā-śāstra*. The present paper's §V records T1509's structural absence from the three-cluster framework precisely as preparation for the P13 treatment of T1509's distinct *śīla-pāramitā* discussion.
- **P14** (*The Nine Precept-Categories of the Bodhisattvabhūmi Śīlapaṭala*) — the present paper's three-cluster framework serves as the framing structure within which the nine-fold *navavidha-śīla* taxonomy operates. P14 will develop the nine subdivisions within the three-fold frame established here.
- **P15** (*The Pusa Yingluo Benyejing and the East Asian Ordination Architecture*) — the present paper's record of T1485's distinct ordering (§3.4) anticipates the more extended treatment of T1485 in P15.

### Volume-internal (Volume IV SĪLA)

- **SĪLA Part I** (P1 *What* Sīla *Actually Means*) — the volume's foundational reading of the term *śīla*. The present paper's three-fold structural design of bodhisattva ethics builds upon the volume's foundational understanding.
- **SĪLA Part II** (P7 *Uposatha and the Genius of Intermittent Monasticism*; P8 *Sīlagandha — the Fragrance That Travels Against the Wind*) — the *śīla* axis of the Pāli-canonical foundation; the present paper presupposes the Pāli-canonical grounding established in Part II.
- **SĪLA Part III** (P9 *The Architecture of the* Pātimokkha; P10 *Three Vinaya Lineages in Conversation*; P11 *Two Sangha Crises and the Mechanism of Boundary-Renegotiation*) — the *śrāvaka*-internal *Vinaya* literature treated in Part III is what the present paper, opening Part IV, brings into the Mahāyāna structural reframing.

### Cross-volume

- **NIAN Part I** — the Pāli-root foundational reading is methodologically continuous with the present paper's commitment to reading each tradition on its own terms before structural integration.
- **DĀNA Part II** (the Mahāyāna turn within the *dāna* volume) — structurally parallel to the present paper: where DĀNA Part II reframes the gift-category as *dāna-pāramitā*, P12 reframes the precept-category as the three-cluster architecture. The parallel is structural, not thesis-borrowing.
- **SAMĀDHI volume (forward)** — the *śīla*-to-*dhyāna* transition treated in SĪLA P24 prepares the entry into the SAMĀDHI volume; the present paper's *trikāya* mapping in §V anticipates the Tiantai integration that SAMĀDHI will develop further.

---

*CBETA collation:* The Chinese-translation primary anchors of the present paper — the *Pusa dichi jing* (T1581, *Bodhisattvabhūmi* translation, Dharmakṣema, Northern Liang fifth century, fascicle 5 three-cluster naming-and-function-triad passage at T30n1581_p0918b01–06, principal terminological anchor); the *Pusa shan jie jing* (T1582, parallel translation of the same Sanskrit base, Guṇavarman, Liu Song fifth century, fascicle 1 Upāli dialogue at T30n1582_p0961c23–p0962a20 with the greed/aversion-inversion at p0962a14–20 supplying the strongest doctrinal argument for the sattva-welfare cluster's defining status, and the fascicle 9 alternative three-precept framework at T30n1582_p1009b07–11); the *Pusa yingluo benyejing* (T1485, Zhu Fonian, Yao Qin late fourth century, fascicle 2 alternative ordering passage at T24n1485_p1020b29–c03 with the accumulation-of-virtue / sattva-welfare / restraint sequence); the *Yogācārabhūmi* (T1579, Xuanzang, Tang seventh century, fascicle 40 three-cluster passage at T30n1579_p0523a02–03 with the *ráoyì yǒuqíng jiè* 饒益有情戒 lexical revision); the *Mahāyānābhidharmasamuccayavyākhyā* (T1606, Sthiramati composed and Xuanzang translated, fascicle 12 *abhidharma*-layer endorsement at T31n1606_p0750c09); and the *Tiantai pusa jie shu* (T1812, Mingkuang's Tang-era redaction, three-cluster ↔ three-body explicit mapping at T40n1812_p0584b07–10 and four-vast-vows ↔ three-clusters mapping at T40n1812_p0580c22–25 — principal anchor for the paper's structural argument) — all have been collated against the CBETA Electronic Tripiṭaka 2024. The *Mahāprajñāpāramitā-śāstra* (T1509, Nāgārjuna / Kumārajīva) is *explicitly recorded as absent* from the three-cluster framework (full-text search of T25 No. 1509 confirms zero occurrences of the *saṃvara-śīla* / *kuśala-dharma-saṃgrāhaka-śīla* / *sattvārtha-kriyā-śīla* triad); T1509 employs a distinct three-precept taxonomy organized by stage of the path and is reserved for P13's *śīla-pāramitā* treatment. The Pāli reference *Aṅguttara Nikāya* 8.39 (*Abhisanda Sutta*, the eight streams of merit) is cited via SuttaCentral with Bhikkhu Sujato's translation under CC0, as the functional-antecedent reference. Modern scholarship: Funayama Toru (*Journal of Asian History* 38.2, 2004 and *Tōhō gakuhō* related articles), Michael Zimmermann (*Harvard Oriental Series* vol. 75, Harvard University Press 2013, pp. 872–883), and Charles Goodman (*Consequences of Compassion*, Oxford University Press 2009, chapter 4) are paraphrase engagements with explicit *paraphrase summary; no verbatim accessed* honest-disclosure markers as required by the project's scholarship conventions for paywalled and limited-access sources.

*CBETA collation:* all *Taishō* T-number citations above verified against CBETA Electronic Tripiṭaka 2024; Pāli citations via SuttaCentral with Bhikkhu Sujato's translation under 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 Twelve · May 2026*