Alternative Lineages: The Shisong lü 十誦律 in Japanese Ancient Manuscript Buddhist Canons

Abstract

Traditional studies on Chinese Buddhism have largely relied on printed canons from the Song, Yuan, Ming, and Goryeo dynasties. However, recent discoveries of Dunhuang and Turfan manuscripts, along with growing recognition of Nihon kosha issaikyō (Japanese Ancient Manuscript Canons), have expanded the scope of Buddhist textual research. Despite their significance, Japanese manuscript Buddhist canons remain underexplored, particularly in relation to their textual lineages and connections to Tang-dynasty texts. This study examines Nihon kosha issaikyō through a philological analysis of the Shisong lü (Ten Recitation Vinaya), assessing textual variants, structural patterns, and transmission histories. By situating Nihon kosha issaikyō within the broader East Asian Buddhist tradition, this research clarifies their role in preserving alternative textual lineages beyond standardized printed canons. The findings contribute to a deeper understanding of Buddhist textual transmission, canon formation, and the interplay between manuscript and printed traditions in China, Korea, and Japan. This study highlights the historical processes that shaped East Asian Buddhist canons and offers new insights into their adaptation and preservation across different cultural contexts.

1. Introduction

Traditional studies on Chinese Buddhism have long relied on modern reconstructions of the Chinese Buddhist canon, primarily based on printed editions from the Song 宋, Yuan 元, Ming 明, and Goryeo 高麗 dynasties. However, in recent decades, significant developments have reshaped the field. First, the gradual public release of manuscript discoveries from Dunhuang and Turfan has greatly advanced research in both Buddhist and Chan studies. Second, the exploration and utilization of foreign manuscript resources—commonly referred to as Nihon ko shakyo 日本古寫經 (Japanese ancient manuscript scriptures)—have expanded the diversity and textual stratification of canonical sources. Finally, in addition to facsimile and reprint editions of various printed Buddhist canons, recent technological advances, such as full-text digital image databases, have enabled scholars to engage more directly with these primary sources. These abundant textual materials provide a more robust documentary foundation for Buddhist studies. At the same time, they introduce new philological challenges, particularly in terms of structure, textual content, and transmission history.

Compared to manuscript findings from Dunhuang and Turfan, or the well-studied printed canons of the Song, Yuan, and Goryeo dynasties, Chinese scholarship has only recently begun to explore the concept of Japanese ancient manuscript scriptures1. This study seeks to address this gap by focusing on a specific category within this corpus—the Nihon kosha issaikyō 日本古寫一切經 (Japanese ancient manuscript canons). The term Nihon koshakyō broadly refers to all Chinese-language Buddhist texts copied in Japan prior to the Kamakura 鎌倉 and Muromachi 室町 periods, while the Nihon kosha issaikyō specifically denotes Japanese manuscript Buddhist canons produced during the same period. This study focuses specifically on the Nihon kosha issaikyō, analyzing surviving manuscript Buddhist canons from the Nara 奈良, Heian 平安, and Kamakura periods to assess their textual lineage and historical development.

The history of the Japanese manuscript Buddhist canon dates back to the first half of the eighth century during the Nara period and continued to develop through the Heian and Kamakura periods. In terms of extant textual materials, approximately 2000 scrolls from the Nara-period Buddhist canon survive, whereas manuscript Buddhist canons from the late Heian period onward number in the tens of thousands, spanning over ten major collections. In the 1990s, Buddhist scholars began to recognize that some Heian- and Kamakura-period manuscript Buddhist canons contained ancient and previously lost Buddhist texts (koitsu butten 古逸佛典), prompting comprehensive investigations. As a result, it was hypothesized that manuscript Buddhist canons copied after the twelfth century inherited their textual lineage from the Nara-period manuscript Buddhist canon, which in turn derived from Tang-dynasty Buddhist manuscripts. While scholars have suggested that Heian- and Kamakura-period Buddhist canons inherited their textual lineage from Nara-period manuscripts, direct textual evidence remains limited. This study addresses this gap by providing a concrete philological analysis of the Shisong lü 十誦律.

Many ancient Japanese Buddhist manuscripts, including those of the Shisonglü, exhibit characteristics of what modern codicologists term the “patchwork codex” (baina ben 百衲本). This phenomenon refers to the practical strategy of recombining partial or fragmentary texts over time—often due to missing scrolls or transmission interruptions—by inserting replacement sections or reusing leaves from other manuscripts. As a result, a single scroll may contain material copied at different periods by different hands, leading to textual layering or duplication. This codicological reality reflects the pragmatic ethos of Buddhist textual preservation, in which functionality and ritual usability often took precedence over bibliographic purity. Recognizing this layered structure is essential for interpreting the form and content of the Japanese Shisonglü manuscripts examined in this study, and it helps contextualize their textual “disorder” and redundancy. Through a detailed philological analysis of the Shisong lü, examining its textual variants, structural patterns, and transmission histories, this study assesses whether Japanese manuscript Buddhist canons preserve direct links to Nara-period manuscripts.

To examine this hypothesis, this paper situates the Japanese manuscript Buddhist canon within the broader framework of the East Asian Chinese-language Buddhist canon, revealing the complexities and distinctive features of these transregional textual sources. It further seeks to clarify the position of the Japanese manuscript Buddhist canon in the historical development of East Asian Buddhist textual traditions. By integrating manuscript and printed textual traditions from China, Korea, and Japan, this study aims to offer new insights into the transmission, adaptation, and preservation of Buddhist scriptures in East Asia. In doing so, it contributes to a more nuanced understanding of Buddhist textual history, the evolution of canonical traditions across different cultural contexts, and the complex interactions between regional Buddhist communities—ultimately enriching the field of Buddhist textual studies. By bridging manuscript and printed traditions, this research deepens our understanding of Buddhist textual transmission while contributing to broader discussions on textual authority, canon formation, and the cross-cultural adaptation of religious texts, ultimately shedding new light on the historical processes that shaped East Asian Buddhism and its canonical traditions.

2. The Transmission and Textual Bifurcation of the Shisong lü into Northern and Southern Lineages

As the first Chinese translation of a comprehensive Vinaya 廣律 (guang lü) text and the only complete version of the *Daśa-bhāṇavāra-vinaya (alt. Sarvāstivādavinaya? *Daśādhyāya-vinaya), the Shisong lü 十誦律 was translated into Chinese in the early fifth century CE (see Hirakawa 1960, p. 128). The translation process began on the seventeenth day of the tenth month in the sixth year of the Hongshi 弘始 era of the Yao Qin 姚秦 dynasty (404 CE), with the Central Asian Buddhist monk Punyatāra 弗若多羅 (fl. 404 CE) reciting the Sanskrit text while Kumārajīva 鳩摩羅什 (343–413 CE) rendered it into Chinese. However, the project was abruptly interrupted when Punyatāra passed away, leaving only two-thirds of the text translated2. The following year, Dharmaruci 曇摩流支 (fl. 405 CE) arrived in Chang’an 長安 and, at the request of Huiyuan 慧遠 (334–416 CE) of Mount Lu 廬山, collaborated with Kumārajīva to complete the translation, producing an initial 58-scroll draft. This preliminary version remained complex and unrefined, and before it could be properly revised, Kumārajīva passed away in the eleventh year of the Hongshi era (409 CE)3. Fortunately, his Vinaya teacher, Vimalākṣa 卑摩羅叉 (?–413 CE), was also in Chang’an at the time. After Kumārajīva’s death, Vimalākṣa traveled to the Eastern Jin territory and, at Shijian Monastery 石澗寺 in Shouchun 壽春 (present-day Shouxian, Anhui), undertook a thorough revision and expansion of the initial draft, shaping what would become the 61-scroll expanded edition4.

As a result of this complex translation history, the bifurcation of the Shisong lü into Northern and Southern textual lineages had already taken shape by the early fifth century. The 58-scroll initial draft, completed in Chang’an under the Yao Qin, became the foundation of the Northern lineage, while the revised and expanded 61-scroll edition, finalized in Shouchun under the Eastern Jin, marked the beginning of the Southern lineage. Both lineages have surviving textual witnesses. Manuscripts discovered in Dunhuang and Turfan predominantly belong to the Northern lineage and often contain non-canonical versions of the text (See Hirakawa 1963, pp. 545–51; L. Wang 2021, pp. 84–95; Liu 2021a; 2021b, pp. 7–18; 2022, pp. 27–43). In contrast, the Southern lineage, which later became the canonical version, is preserved in both printed Buddhist canons and Japanese manuscript Buddhist canons. This study focuses on the latter, examining its transmission, textual evolution, and role in shaping the East Asian Buddhist canon.

The Southern lineage of the Shisong lü (canonical texts) comprises two primary components: the Shisong 十誦 (Ten Recitations) and the Lüxu 律序 (Preface to the Vinaya). The Shisong consists of ten recitations (chapters) that address various aspects of monastic discipline. The first through sixth recitations concern the precepts for monks (senglü 僧律), the seventh focuses on regulations for nuns (nilü 尼律), the eighth outlines additional rules (zengyi fa 增一法), the ninth presents dialogues between Upavasatha and the Buddha elaborating on monastic conduct, and the tenth contains miscellaneous regulations (za fa 雜法) addressing disciplinary matters not covered in the preceding sections.

The Lüxu, also known as the “Preface to the Shisong lü” (Shisong lü xu 十誦律序), or simply the “Vinaya Preface” (Pini xu 毗尼序), is typically structured into three scrolls and divided into four sections (pin 品). The first scroll contains Section 1, “The Assembly of Five Hundred Bhikṣus and the Compilation of the Tripiṭaka” (Wubai biqiu ji sanzang fa 五百比丘集三藏法), which describes the earliest efforts to codify the monastic rules, followed by the beginning of Section 2, “The Assembly of Seven Hundred Bhikṣus and the Elimination of Unwholesome Dharma” (Qibai biqiu ji mie’e fa 七百比丘集滅惡法), which discusses subsequent refinements of the Vinaya. The second scroll completes Section 2 and introduces Section 3, “Miscellaneous Vinaya Matters” (Pini za pin 毗尼雜品), which addresses additional monastic regulations. The third scroll consists of Section 4, “Conditions and Circumstances Leading to the Compilation of the Vinaya” (Yinyuan pin 因縁品), which details the historical context, process, and narratives surrounding the establishment of monastic rules. The Lüxu thus provides an essential framework for understanding the formation, codification, and historical development of monastic discipline in the Southern textual lineage of the Shisong lü.

The transmission and evolution of the Shisong lü have long been subjects of scholarly debate, centering on two key issues: the relationship between the Northern and Southern lineages, and the origins of the Lüxu. Early sources from the Southern Dynasties, such as the Gaoseng zhuan 高僧傳 (Biographies of Eminent Monks) and Chu sanzang ji ji 出三藏記集 (Collection of Records about the Translation of the Tripiṭaka), regarded the Southern lineage as a revision of the Northern lineage but made no explicit mention of the Lüxu.5 During the Sui dynasty, Buddhist catalogues such as Fajing’s 法經 (fl. 594 CE) Zhongjing mulu 眾經目錄 (Catalogue of Buddhist Scriptures) and Fei Changfang’s 費長房 Lidai sanbao ji 歷代三寶紀 (Records of the Three Treasures throughout the Successive Dynasties) described the Southern lineage as either a “retranslation” (chongyi 重譯) or “alternative translation” (bieyi 別譯) of the Northern lineage6. By the Tang dynasty, a more balanced perspective emerged in works such as the Datang neidian lu 大唐内典錄 (Great Tang Inner Canon Catalogue)7 and Zhisheng’s 智昇 (fl. 730 CE) Kaiyuan Shijiao lu 開元釋敎錄 (Kaiyuan Catalogue of Buddhist Scriptures). The Kaiyuan Shijiao lu, in particular, refuted earlier claims from the Six Dynasties and Sui sources by conducting a comparative analysis of two Northern lineage versions (one with 58 scrolls and another with 59 scrolls) alongside the 61-scroll Southern lineage version8. It concluded that the 58-scroll Shisong section was translated collaboratively by Punyatāra, Kumārajīva, and others, whereas the three-scroll Lüxu was a later augmentation attributed to Vimalākṣa9. This argument further reinforced the distinction between the two textual lineages and provided a clear historical framework for understanding their divergence.

These historical disagreements have continued to fuel debate among modern scholars. Notably, Japanese scholar Hirakawa Akira 平川彰 examined Northern lineage manuscripts discovered in Dunhuang and preserved in the British Library, identifying significant differences in scroll division, content complexity, classical Chinese phrasing, and translation terminology between these manuscripts and the canonical texts of Southern lineage (See Hirakawa 1963, pp. 545–51). Based on these findings, he supported the view—found in Southern Dynastie sources—that the Southern version was a reconstruction of the Northern version, and argued that the Northern version likely included content corresponding to the Lüxu, albeit with a structure and wording that differed from those of the Southern version, raising doubts on whether Vimalākṣa was truly its translator (See Hirakawa 1960, p. 128).

In contrast, Chinese scholars have focused on two primary differences between the Northern and Southern lineages: variations in the Shisong, particularly in scroll division, content length, and translation terminology (See L. Wang 2021, pp. 84–95; Liu 2022, pp. 27–43). These inconsistencies are not uniform across all chapters, suggesting that the extent of revisions in the Southern lineage depended on the quality of the Northern version’s original translation—sections rendered with greater fidelity underwent fewer modifications (Liu 2022, p. 42). Regarding the absence of the Lüxu in the Northern lineage, Chinese scholars generally maintain that the Lüxu in the Southern lineage was indeed translated by Vimalākṣa (See L. Wang 2021, pp. 84–95). This view aligns with the conclusions of the Kaiyuan Shijiao lu, which asserts that the Southern version inherited and revised the 58-scroll Northern version while incorporating the three-scroll Lüxu. However, despite these revisions, the modifications introduced in the Southern lineage were not substantial enough to constitute a complete “retranslation” (chongyi 重譯) or an entirely “separate translation” (bieyi 別譯).

The evolution of the Lüxu 律序 within the Southern lineage of the Shisong lü underwent significant textual transformations over time. Historical records from the Six Dynasties, Sui, and Tang periods indicate that this newly added preface did not maintain a fixed position relative to the Shisong inherited from the Northern version. Based on its varying placement, Southern lineage texts can be categorized into three distinct types: Lüxu danzhi 律序單置 (Preface Independently Placed), Lüxu houzhi 律序後置 (Preface Post-Placed), and Lüxu qianzhi 律序前置 (Preface Pre-Placed).

These textual variations outline four key phases in the evolution of the Southern lineage. The first phase, during the Six Dynasties, corresponds to the Lüxu danzhi type, where the Lüxu was treated as an independent section, separate from the Shisong10. In this version, the final section of the Shisong was the Tenth Recitation (Dishi song 第十誦), also known as the “Virtuous Recitation” (Shan song 善誦) or the “Vinaya Recitation” (Pini song 毗尼誦).

The second phase, emerging in the Sui and early Tang periods, saw the prevalence of Lüxu houzhi type. In this configuration, the Lüxu was appended to the end of the Tenth Recitation, and the combined section was collectively referred to as the Shan song11. Notably, in this phase, the meanings of Shan song and Pini song diverged from their earlier usage in the Lüxu danzhi configuration. While Shan song continued to designate the final portion of the Shisong lü, it now encompassed both the Tenth Recitation and the Lüxu, whereas Pini song came to refer specifically to the Lüxu.

The third phase, spanning the early Tang period through the Kaiyuan era (713–741 CE), witnessed the emergence of the Lüxu qianzhi type, in which the Lüxu was repositioned between the Ninth and Tenth Recitations.

The fourth phase, spanning from the mid-to-late Tang dynasty through the Five Dynasties period, marked a return to the Lüxu houzhi type, thereby restoring the structural arrangement that had prevailed during the Sui and early Tang periods. This shift was largely catalyzed by the influence of Zhisheng’s Kaiyuan Shijiao lu, which strongly advocated for the Lüxu houzhi configuration12. Zhisheng’s work also introduced terminological reforms that had a lasting impact: he renamed the Lüxu as Pini xu 毗尼序 (Vinaya Preface)13, a designation that has since become standard. By replacing the earlier term Pini song 毗尼誦 (Vinaya Recitation) with Pini xu 毗尼序 (Vinaya Preface), Zhisheng redefined its textual status from a “recitation” (song 誦) to a “preface” (xu 序), thereby consolidating its canonical position in subsequent traditions.

The Kaiyuan Shijiao lu exerted a profound influence on Buddhist catalogues and canonical compilations from the mid-Tang period onward. As a result, manuscript canons produced during the late Tang and Five Dynasties uniformly adopted the Lüxu houzhi configuration as the authoritative structural model (Chi 2021, pp. 31–32; L. Wang 2022, pp. 31–32). This trend continued into the printed canons of the Song 宋, Khitan 契丹, Jurchen 金, Yuan 元, and Goryeo 高麗 dynasties, all of which followed this format, consistently incorporating the Lüxu houzhi version and referring to the Lüxu as the Pini xu.

An examination of extant Buddhist canons confirms that editions of the Shisong lü in the printed canons from the Song, Khitan, Jurchen, Yuan, and Goryeo periods all conform to the Lüxu houzhi structure, consistently comprising a total of 61 scrolls. In contrast, manuscript Buddhist canons from Japan’s Nara, Heian, and Kamakura periods uniquely preserve the Lüxu qianzhi format, thereby distinguishing them from their continental counterparts14. This divergence highlights Japan’s distinctive role in preserving an alternative textual lineage. The following section turns to an analysis of the Shisong lü as transmitted in woodblock-printed Buddhist canons.

3. The Textual Evolution of the Shisong lü: Focusing on the Southern Lineage in the Canonical Tradition

The woodblock-printed editions of the Buddhist canons produced during the Song, Khitan, Jurchen, Yuan, and Goryeo dynasties are generally classified into three major textual lineages. The first is the Central Plains lineage (中原系 Zhongyuan xi), which includes the Kaibao Canon 開寶藏, both editions of the Goryeo Canons 高麗藏, and the Jurchen Canon 金藏. The second is the Northern lineage (北方系 Beifang xi), represented by the Khitan Canon 契丹藏. The third is the Jiangnan lineage (江南系 Jiangnan xi), which comprises various regional editions such as the Fuzhou Canons 福州藏 (including the Dongchan 東禪 and Kaiyuan 開元 editions), the Sixi Canon 思溪藏, the Qisha Canon 磧砂藏, and the Puning Canon 普寧藏 (see Chikusa 2000, pp. 281–89).

While all three lineages of woodblock-printed canons include the Shisong lü in its Lüxu houzhi version, consistently comprising 61 scrolls, their internal structures and textual content exhibit notable variations. The precise origins of these textual sources and their interrelationships remain unclear in many respects. To better understand these differences, this section selects representative editions from each of the three lineages and compares their textual structures (See

Table 1. Comparison of the Lüxu qianzhi and the Lüxu houzhi versions of the Shisong lü.

Lüxu Qianzhi Version		Lüxu Houzhi Versions
Structure	Japanese Manuscript Canons	Structure	Northern Edition	Jiangnan Edition	Central Plains Edition
First Recitation	Scrolls 1–6	First Recitation	Scrolls 1–6	Scrolls 1–6	Scrolls 1–6
Second Recitation	Scrolls 7–13	Second Recitation	Scrolls 7–13	Scrolls 7–13	Scrolls 7–13
Third Recitation	Scrolls 14–18	Third Recitation	Scrolls 14–18	Scrolls 14–18	Scrolls 14–18
	Scroll 19		Scroll 19	Scroll 19	Scroll 19
	Scroll 20		Scroll 20		Scroll 20
				Scroll 20
Fourth Recitation	Scrolls 21–28	Fourth Recitation	Scrolls 21–28	Scrolls 21–28	Scrolls 21–28
Fifth Recitation	Scrolls 29–34	Fifth Recitation	Scrolls 29–34	Scrolls 29–34	Scrolls 29–34
Sixth Recitation	Scrolls 36–41	Sixth Recitation	Scrolls 36–41	Scrolls 36–41	Scrolls 36–41
Seventh Recitation	Scroll 42	Seventh Recitation	Scroll 42	Scroll 42	Scroll 42
					Scroll 43
	Scroll 43		Scroll 43	Scroll 43
					Scroll 44
					Scroll 45
	Scroll 44		Scroll 44	Scroll 44
	Scroll 45		Scroll 45	Scroll 45
					Scroll 46
	Scroll 46		Scroll 46	Scroll 46	Scroll 47
Eighth Recitation	Scroll 47	Eighth Recitation	Scroll 47	Scroll 47	Scroll 48
	Scroll 48		Scroll 48	Scroll 48
					Scroll 49
	Scroll 49		Scroll 49	Scroll 49	Scroll 50
	Scroll 50		Scroll 50	Scroll 50	Scroll 51
Ninth Recitation	Scroll 51	Ninth Recitation	Scroll 51	Scroll 51	Scroll 52
	Scroll 52		Scroll 52	Scroll 52	Scroll 53
	Scroll 53		Scroll 53	Scroll 53
					Scroll 54
				Scroll 54
	Scroll 54		Scroll 54
					Scroll 55
Lüxu	Scroll 55	Tenth Recitation	Scroll 55	Scroll 55	Scroll 56
	Scroll 56		Scroll 56
				Scroll 56	Scroll 57
	Scroll 57		Scroll 57	Scroll 57	Scroll 58
Tenth Recitation	Scroll 58		Scroll 58	Scroll 58
					Scroll 59
	Scroll 59	Lüxu	Scroll 59	Scroll 1	Scroll 60
	Scroll 60		Scroll 60	Scroll 2	Scroll 61
	Scroll 61		Scroll 61	Scroll 3

). Additionally, to highlight the distinctive features of the Lüxu houzhi configuration, the Lüxu qianzhi type preserved in Japanese manuscript Buddhist canons is also incorporated as a point of comparative reference.

The extant Shisong lü in the Central Plains lineage of the Buddhist Canon is preserved in several editions, including the Kaibao Canon (one scroll, Scroll 46)15, the Initial Edition of Goryeo Canon (15 scrolls: Scrolls 12–15, 18–19, 42–45, 48–52)16, the Recarved Edition of Goryeo Canon (61 scrolls)17, and the Jurchen Canon (54 scrolls: Scrolls 1–2, 4–7, 9–20, 22–36, 39–43, 45, 47–61)18. Taking the Recarved Edition of Goryeo Canon (hereafter referred to as the Recarved Goryeo edition) as an example, it comprises 59 scrolls of the Shisong and two scrolls of the Lüxu, totaling 61 scrolls in the Lüxu houzhi format. However, its internal structure does not fully conform to the arrangement described in the Kaiyuan Shijiao lu and appears somewhat irregular and unbalanced. The discrepancy arises from the allocation of 59 scrolls to the Shisong section, which—within the constraints of the 61-scroll framework—necessitated compressing the Lüxu from three scrolls into two. As a result, Scroll 61 (76 folios) is more than twice the length of Scroll 60 (34 folios).

The Northern lineage, represented by the Khitan Canon, has largely been lost. However, its internal structure, including that of the Shisong lü, can be inferred from records in the Xin zan yiqiejing yuan pinci lu 新纘一切經源品次錄 (a treatise from the end of the Tang Dynasty), and the Xin ji zangjing yinyi suihan lu 新集藏經音義隨函錄 (a treatise from the Five Dynasties)19. These sources indicate that the Northern lineage consisted of 58 scrolls of the Shisong and three scrolls of the Lüxu, making it the version most consistent with the structure described in the Kaiyuan Shijiao lu. (See Chi 2021, pp. 25–59). By the late Tang and Five Dynasties periods, this textual form had become the predominant version of the Shisong lü in Northern China, representing its final phase of evolution. In comparison, the Central Plains lineage exhibits several structural differences. The First Recitation diverges in its division point (between Scrolls 3 and 4), and the addition of a scroll in the Seventh Recitation (Scrolls 42–47) shifts all subsequent recitations forward by one scroll. Apart from Scrolls 47, 50–52, and 60, most scroll divisions also differ.

The Jiangnan lineage, comprising multiple regional canons, offers a rich and diverse textual tradition. Taking the Fuzhou Dongchan Edition (Kunai-chō shoryō-bu shūzō kanseki shūran, n.d.)20 as an example, it consists of 58 scrolls of the Shisong and three scrolls of the Lüxu. In this version, the Lüxu is divided into three scrolls, annotated as “Nine”, “Ten”, and “Eleven” (corresponding to scrolls 59, 60, and 61, respectively). Compared to the Northern edition, the Jiangnan editions differ in the segmentation of the Third Recitation (between Scrolls 19 and 20), the Ninth Recitation (between Scrolls 53 and 54), and the Tenth Recitation (between Scrolls 55 and 56)21. Despite these differences, the overall method of segmentation remains highly consistent with that of the Northern edition.

Having examined the structural features of the Central Plains, Northern, and Jiangnan canonical editions, we now turn to the question of whether significant textual differences exist among them. Important clues can be found in the Collation Supplementary Records 校正別錄 (Jiaozheng Bielu), particularly in Scroll 20, which highlights variations in Scroll 5 of the Shisong lü across the Khitan, Initial Goryeo, and Kaibao editions22. These collation records reveal that, compared to the Initial Goryeo and Kaibao editions, the Khitan edition contained three extensive passages, which were later incorporated into the Recarved Goryeo edition. Scroll 5 of the Jurchen edition also retains these three passages, albeit in an abbreviated form, a characteristic of the Central Plains lineage.

In contrast, in the Jiangnan lineage as represented by the Dongchan edition, Scroll 5 not only abbreviates the aforementioned three passages but also drastically shortens two additional sections to just twelve characters each. Notably, these same sections contain 1349 and 1021 characters, respectively, in the Jurchen and Recarved Goryeo editions. Since the Collation Supplementary Records does not provide collation notes for these two instances, it can be inferred that the Kaibao and Initial Goryeo editions did not differ from the Khitan and Recarved Goryeo editions in these instances.

These five cases of textual omission and expansion reveal systematic differences among the three major textual lineages. The Jiangnan lineage consistently presents the most concise versions, with brevity serving as a defining feature. The Central Plains lineage—excluding the Recarved Goryeo edition—adopts an intermediate editorial stance, abbreviating the first three passages while retaining the longer versions of the latter two. In contrast, the Northern lineage preserves the most extensive textual content, with accretive tendencies constituting a hallmark of this tradition.

Given that the Northern lineage represents the final evolutionary stage of the Shisong lü and was dominant in Northern China during the late Tang and Five Dynasties periods, it is possible to infer the relative chronology of these textual traditions based on their textual proximity to the Northern lineage. Unless exceptional circumstances apply, the closer a text aligns with the Northern lineage, the later its likely date of production. By this measure, the Jiangnan lineage—which exhibits the greatest divergence—likely represents the most archaic stratum. The Central Plains lineage occupies an intermediate position, while the Northern lineage, as the most developed, probably did not emerge until the mid-ninth century.

These patterns reveal two major evolutionary trends in the Southern lineage of the Shisong lü during the Tang Dynasty: (1) increasing structural refinement, characterized by greater segmentation and reorganization; and (2) gradual textual accretion, reflected in the incremental inclusion of additional passages. Together, these developments offer a clearer framework for understanding the evolution of the Shisong lü within the broader context of the East Asian Buddhist canonical tradition.

4. The Shisong lü in Japanese Manuscript Buddhist Canons of the Nara, Heian, and Kamakura Periods

During the Kaiyuan era (713–741 CE) of the Tang Dynasty, both the Lüxu houzhi and Lüxu qianzhi versions of the Shisong lü were in circulation in Chang’an. However, after the late tenth century, the Lüxu qianzhi version disappeared from both China and the Korean Peninsula, surviving only in Japanese manuscript Buddhist canons. These Japanese manuscripts provide valuable supplementary sources for the study of ancient Chinese Buddhist literature. This section introduces the manuscript Buddhist canons of the Nara and Heian periods and examines their respective versions of the Shisong lü.

According to historical records, Japan’s tradition of copying manuscript Buddhist canons dates back to the second year of Emperor Tenmu 天武天皇 (673 CE) during the Asuka 飛鳥 period23. By the Nara period (710–794 CE), both governmental and private institutions were actively engaged in the transcription of Buddhist scriptures, resulting in the production of more than twenty distinct official Buddhist canons24. The official Buddhist canon of the Nara period was primarily sponsored by the central government, with transcription activities overseen by two major imperial institutions.

The first of these institutions was the Office of the Empress’s Household 皇后宮職系 (Kōgōgūshiki-kei), which played a pivotal role in organizing large-scale scripture transcription projects. This office was responsible for producing the earliest state-sponsored manuscript Buddhist canon and remained a major center for Buddhist textual production throughout the Nara period. Initially established prior to the eighth year of the Tenpyō 天平 era (736 CE), it gradually evolved into the State Scriptorium 寫經所 (Shakyōsho), which, by the twentieth year of the Tenpyō era (748 CE), came under the administrative authority of the Bureau of the Great Eastern Temple 造東大寺司 (Zō Tōdaiji-shi). This bureau, which oversaw the construction and maintenance of the Great Eastern Temple 東大寺 (Tōdai-ji), subsequently became a central institution for the transmission and preservation of Buddhist texts.

The most significant Buddhist canon produced under the Office of the Empress’s Household was the Scriptures/Canon [copied] at the behest of Empress Kōmy ō’s wish 光明皇后御願經, also known as the Gogatsu-tsuitachi-kyō 五月一日經 (Scriptures/Canon [with the Imperial Wish Text dated] on the first day of the fifth month [of the 12th year of the Tenpyō era (740 CE)]. This monumental manuscript canon was transcribed between the fifth year of the Tenpyō era (733 CE) and the eighth year of the Tenpyō Shōhō 天平勝寶 era (756 CE), under the patronage of Empress Kōmyō (701–760 CE). The Gogatsu-tsuitachi-kyō not only established the standard for subsequent manuscript Buddhist canons but also served as the textual foundation for later transcription projects.

Beyond the Gogatsu-tsuitachi-kyō, the Office of the Empress’s Household also supervised the production of several other manuscript Buddhist canons, including the Great Officer Canon 大官一切經 (Daikan issaikyō), the Early Manuscript Buddhist Canon 先寫一切經 (Sensha issaikyō), the Later Manuscript Buddhist Canon 後寫一切經 (Kōsha issaikyō), the Votive Buddhist Canon of Empress Kōmyō 光明皇太后發願一切經 (Kōmyō Kōtaigō Hotsugan issaikyō), and the Memorial Buddhist Canon for Empress Kōmyō’s Death Anniversary 光明子周忌齋一切經 (Kōmyōshi Shūki-sai issaikyō)25. These manuscript Buddhist canons reflect the imperial court’s sustained commitment to Buddhist textual preservation and ritual practice. They not only played a foundational role in shaping Japan’s early Buddhist scriptural tradition but also contributed to the broader transmission and adaptation of Buddhist texts across East Asia.

The Second institution was the Office of the Imperial Household 內裏系 (Dairi-kei), which as responsible for overseeing the transcription of Buddhist scriptures on behalf of the imperial family. This bureau was originally established to produce the Votive Canon for Emperor Shōmu 聖武天皇御願經 and operated under the Manuscript Buddhist Canon Bureau 寫御書所. Over time, it evolved into the Imperial Scriptorium for Buddhist Texts 奉寫御執經所 (Hōsha Goshitsukyō sho), which was formally established during the Tenpyō Hōji 天平寶字 period (757–765 CE), under the reign of Empress Shōtoku 稱德天皇 (r. 764–770 CE).

One of the most significant transcription projects conducted by this institution was the Scriptures/Canon [copied] at the behest of Empress Shōtoku 稱德天皇御願經, also known as the Jingo-keiun-kyō 神護景雲經. This manuscript Buddhist canon was transcribed between the second year of the Tenpyō Hōji era (758 CE) and the third year of the Jingo-keiun 神護景雲 era (770 CE). In the first year of the Jingo-keiun era (767 CE), a major institutional reform took place: the Tōdai-ji Scriptorium (originally under the Empress’s Bureau of Buddhist Manuscripts) and the Imperial Buddhist Manuscript Bureau were merged into a newly consolidated Imperial Buddhist Manuscript Bureau of the Buddhist Canon 奉寫一切經所 (Hōsha Issaikyōsho). This restructuring streamlined administrative operations and improved coordination in the transcription of Buddhist scriptures. During the Hōki 寶亀 era (770–781 CE), this institution produced the Five-set Buddhist Canons 五部一切經 (Gobu Issaikyō), which included the First Set 先一部 (Saki ichibu), the Second Set 始二部 (Shi nibbu), the Additional Set 更一部 (Kō ichibu), and the Further Additional Set 今更一部 (Imakō ichibu)26. These projects reflect the Office of the Imperial Household’s sustained role in promoting preservation and transmission of Buddhist texts in Japan.

Among the Nara-period Buddhist manuscript canons, the Gogatsu-tsuitachi-kyō and the Jingo-keiun-kyō held a preeminent status as the only two officially designated Buddhist canons (imperially authorized Buddhist canons, 敕定一切經). As imperially sanctioned canons, they carried the highest level of authority among manuscript Buddhist canons in Japan. While both adhered to the textual standards established by the Kaiyuan shijiao lu—the authoritative scripture catalogue compiled in Tang China—they also incorporated a wide range of apocryphal, dubious, and non-canonical scriptures, significantly expanding their content beyond standard texts. As a result, these two imperial canons amassed more than 6500 scrolls in total, making them among the most extensive Buddhist manuscript canons of their time27.

The two institutions responsible for compiling and transcribing these canons—one affiliated with the State Scriptorium at Tōdai-ji and the other associated with the Imperial Palace—played a central role in the production, standardization, and preservation of Buddhist canonical texts during the Nara period. Their transcription efforts were deeply influenced by the textual traditions of Tang-dynasty Buddhist canons, thereby maintaining Japan’s connection to the broader East Asian Buddhist literary landscape. At the same time, these institutions introduced unique adaptations and organizational frameworks, shaping the distinct character of Japan’s early Buddhist manuscript traditions. Through their meticulous efforts, the Gogatsu-tsuitachi-kyō and the Jingo-keiun-kyō not only preserved Buddhist scriptures but also contributed to the localization of Buddhist textual culture in Japan, setting a precedent for the manuscript Buddhist canons of the Heian and Kamakura periods.

The Gogatsu-tsuitachi-kyō corpus comprises approximately 1000 extant scrolls, the majority of which—750 scrolls—are preseved in the Sacred Words Collection of the Shōsōin Repository 正倉院聖語藏 (Shōsōin Shōgozō). The remaining 250 scrolls are dispersed across various private collections (See Horiike 2003, pp. 302–14). Among them, 26 scrolls of the Shisong lü have been identified across multiple repositories. Specifically, 23 scrolls are preserved in the Shōsōin Repository (Scrolls 11–20, 29–37, 39, 40, 51, and 61), while additional individual scrolls are found in the Iwasaki Collection at the Tōyō Bunko Library (Tōyō Bunko Iwasaki Bunko 東洋文庫岩崎文庫, Scroll 22), the Japanese Agency for Cultural Affairs Collection (Nihon Bunka-chō 日本文化廳, Scroll 38), and the Nezu Museum (Nezu Bijutsukan 根津美術館, Scroll 56).

Similarly, the Jingo-keiun-kyō comprises approximately 60 extant scrolls, including six scrolls of the Shisong lü, which are scattered across serval institutions. These include Scroll 17 in the Gotō Art Museum (Gotō Bijutsukan 五島美術館), Scroll 25 in the Kyoto National Museum (Kyōto Kokuritsu Hakubutsukan 京都國立博物館), Scroll 42 in the Shōwa Women’s University Kōyō Museum (Shōwa Joshi Daigaku Kōyō Hakubutsukan 昭和女子大學光葉博物館), Scroll 51 in the Fuse Art Museum (Fuse Bijutsukan 布施美術館), and Scrolls 52 and 55 in the Ishiyama-dera石山寺. Additionally, the Sacred Words Collection of the Shōsōin Repository holds another 740 scrolls attributed to the Jingo-Keiun-kyō28, including 17 scrolls of the Shisong lü. Among these, 16 scrolls have been identified as part of the Further Additional Set Canon, transcribed around the year 775 CE (Iida 2012, pp. 186, 191).

In addition to Nara-period manuscripts, Japan preserves an extensive collection of Buddhist canons transcribed by private temples during the Heian (794–1185 CE) and Kamakura (1185–1333 CE) periods. These manuscript Buddhist canons exhibit considerable variation in content, and the dates of transcription can typically only be inferred from extant colophons. Among them, several representative collections contain relatively complete versions of the Shisong lü, offering valuable evidence for the textural transmission and adaptation of Vinaya texts in Japan.

One of the most significant collections is the Ishiyama-dera Manuscript Buddhist Canon 石山寺一切經, located in Ōtsu City, Shiga Prefecture 滋賀縣大津市. This collection, whose transcription spans from the Nara to the Muromachi period, has a primary corpus dating from 1148 to 1159 and consists of 4485 scrolls, including 57 scrolls of the Shisong lü (with Scrolls 17, 25, and 42 missing; Scroll 51 is currently housed in the Fuse Art Museum) (Ishiyamadera Bunkazai Sōgō Chōsadan 1978, pp. 494–504, 771, 845). Another important collection is the Chūson-ji Manuscript Buddhist Canon 中尊寺一切經, preserved at the Kōyasan Reihōkan Museum in Osaka 大阪高野山靈寶館. Copied between 1117 and 1126, this canon comprises 4296 scrolls, including 60 scrolls of the Shisong lü (missing Scroll 45) (Ueyama 1997, p. 1; Fujisawa 1990, pp. 30–32). The Matsuo Shrine Manuscript Buddhist Canon 松尾社一切經, housed at Myōren-ji Temple in Kyoto 京都市妙蓮寺, was transcribed between 1115 and 1138 and consists of 3545 scrolls, including 51 scrolls of the Shisong lü (missing Scrolls 4–6, 12, 45, 47, 50, 55, and 60) (Nakao 1997, pp. 365, 408–13). Another significant collection is the Kōshō-ji Manuscript Buddhist Canon 興聖寺一切經 in Kyoto 京都市, dating from 1163 to 1169, which includes 5294 scrolls and preserves a complete 61-scroll version of the Shisong lü (Kyōto-fu Kyōiku Iinkai 1998, pp. 196–98, 459). Additional noteworthy collections include the Nanatsu-dera Manuscript Buddhist Canon 七寺一切經, located in Nagoya, Aichi Prefecture 愛知縣名古屋市, which consists of 4954 scrolls, including 56 scrolls of the Shisong lü (missing Scrolls 3, 12, 20, 42, and 49) (Nanatsudera Issaikyō Hozonkai 1968, pp. 82–83, 189). The Kongō-ji Manuscript Buddhist Canon 金剛寺一切經, housed in Kawachinagano, Osaka 大阪府河內長野市, was copied between 1208 and 1268 and comprises 4500 scrolls, including 45 scrolls of the Shisong lü (Scrolls 2, 4–5, 23–24, 31–33, 35–50, 52, 54–59) (Ochiai 2007, pp. 286–89, 502–03).

These Heian- and Kamakura-period manuscript Buddhist canons illustrate the continued transmission and preservation of the Shisong lü in Japan. Their variations in textual structure and content reflect the complex process of textual adaptation, compilation, and preservation across different regions and historical contexts. Moreover, the survival of these manuscripts provides crucial comparative material for understanding the broader transmission history of Vinaya texts in East Asia.

5. The Textual Transmission and Complexity of Japanese Manuscript Versions of the Shisong lü

Among the Shisong lü manuscripts from the Nara, Heian, and Kamakura periods discussed above, all versions—except for the Imakō ichibu canon, whose structural status remains uncertain—place the Lüxu (Scrolls 55–57) before the Shisong proper. This structural arrangement is a defining feature of the Lüxu qianzhi version. However, despite this shared structural characteristic, significant textual variations exist among these Japanese manuscript Buddhist canons, particularly in Scroll 61, which appears in two distinct versions.

Only a limited number of Japanese manuscripts preserve Scroll 61 of the Shisong lü. This analysis focuses on three representative examples: the Matsuo Shrine Manuscript Buddhist Canon, the Nanatsu-dera Manuscript Buddhist Canon, and the Kōshō-ji Manuscript Buddhist Canon. All three confirm their affiliation with the Lüxu qianzhi tradition by positioning the Lüxu (Scrolls 55–57) before the Tenth Recitation. However, they diverge in their treatment of Scroll 61, revealing distinct textual transmission patterns.

A key discrepancy emerges in the Matsuo Shrine Manuscript, where Scroll 61 duplicates the fourth section (i.e., the third scroll) of the Lüxu (Scroll 57), resulting in textual redundancy. In contrast, both the Nanatsu-dera Manuscript and the Kōshō-ji Manuscript preserve the fourth scroll of the Tenth Recitation in Scroll 61, maintaining a more coherent structural organization. This suggests that the Nanatsu-dera and Kōshō-ji manuscripts retain the original configuration of Scroll 61 as found in the Lüxu qianzhi version29, whereas the Matsuo Shrine Manuscript exhibits a degree of textual disarray. Further investigation traces this textual “disorder” in the Matsuo Shrine Manuscript back to the Gogatsu-tsuitachi-kyō of the Nara period. Meanwhile, the Nanatsu-dera and Kōshō-ji manuscripts appear to reflect textual corrections influenced by the Kaibao Canon version of the Shisong lü, which was introduced to Japan at the end of the tenth century.

This section examines the textual discrepancies found in Scroll 61 and explores how they reflect the complexities and unique transmission patterns of Japanese manuscript Buddhist canons. To facilitate this discussion, we first present a comparative analysis of the final portion of the Shisong lü (Scrolls 55–61) as preserved in the Gogatsu-tsuitachi-kyō and the Matsuo Shrine Manuscript Canon (See

Table 2. Comparison of Scrolls 55–61.

Scroll No.	Title	The Shisong lü in Gogatsu-Tsuitachi-Kyō	The Shisong lü in Matsuo Shrine Manuscript Buddhist Canon
Scroll 55	(Missing)	(Not extant)	(Not extant)
Scroll 56	Opening Title	Preface to the Shisong lü, Middle Scroll (56) 十誦律序卷中（五十六）	Preface to the Shisong lü, Second Scroll (56) 十誦律序卷中（五十六）
	Closing Title	(Not known)	Preface to the Shisong lü, Scroll 56 十誦律序卷 卷第五十六
Scroll 57	Opening Title	(Not extant)	The Shisong lü, Third Scroll (Yinyuan Chapter) (57) 十誦律卷下 (因缘品 五十七)
	Closing Title	(Not extant)	The Shisong lü, Scroll 57 十誦律卷第五十七
Scroll 58	Opening Title	(Not extant)	The Shisong lü, First Scroll of Virtuous Recitation (58) 十誦律善誦卷第一 (五十八)
	Closing Title	(Not extant)	The Shisong lü, Scroll 58 十誦律卷第五十八
Scroll 59	Opening Title	(Not extant)	The Shisong lü, Second Scroll of Virtuous Recitation (59) 十誦律菩誦卷第二 (五十九)
	Closing Title	(Not extant)	The Shisong lü, Second Scroll of Virtuous Recitation 十誦律菩誦卷二
Scroll 60	Opening Title	(Not extant)	The Shisong [lü], Third Scoll of [Virtuous] Recitation 十誦□□誦卷第三
	Closing Title	(Not extant)	(Unable to verify)
Scroll 61	Opening Title	Preface to the Shisong lü, Fourth Section of Causal Chapter (Scroll 61) 十誦律序因缘品第四（卷六十一）	Preface to the Shisong lü, Fourth Section of the Yinyuan Chapter, (Scroll 61) 十誦律序因缘品第四（卷第六十一）
	Closing Title	The Shisong lü, Scroll 61 十誦律卷第六十一	The Shisong lü, Scroll 61 十誦律卷第六十一

).

The comparative analysis of Scrolls 55–61 in the Gogatsu-tsuitachi-kyō and the Matsuo Shrine Manuscript Buddhist Canon highlights key differences in scroll structure, titles, and textual placement, revealing critical inconsistencies that shaped later manuscript traditions. One of the most striking anomalies in the Gogatsu-tsuitachi-kyō appears in the final scrolls of the Shisong lü. Notably, Scrolls 55 and 57–60 are missing, with only Scrolls 56 and 61 extant. Since Scroll 56 contains the second scroll of the Lüxu, it is reasonable to infer that Scroll 55 originally contained the first scroll and Scroll 57 the third. This configuration confirms that the Gogatsu-tsuitachi-kyō follows the Lüxu qianzhi version, placing it in alignment with the Matsuo Shrine Manuscript.

More significantly, Scroll 61 in the Gogatsu-tsuitachi-kyō erroneously duplicates the content of Scroll 57 (the third scroll of the Lüxu)—a structural anomaly that is closely mirrored in the Matsuo Shrine Manuscript. This correspondence suggests that the root of the Scroll 61 inconsistency observed in later Japanese manuscript Buddhist canons can be traced back to the Gogatsu-tsuitachi-kyō. A key question thus arises: Why did the Gogatsu-tsuitachi-kyō include this erroneous duplication, and what became of the fourth scroll of the Tenth Recitation (Scroll 61), which should have contained the final section?

The answer cannot be found within the Gogatsu-tsuitachi-kyō itself. However, later Japanese manuscript Buddhist canons—particularly the Matsuo Shrine Manuscript—provide valuable clues. Copied in the late Heian period, the Matsuo Shrine Manuscript closely replicates the structural features of the Gogatsu-tsuitachi-kyō. It places the Lüxu in Scrolls 55–57, contains only three scrolls of the Tenth Recitation (Scrolls 58–60), and lacks the expected fourth scroll. Instead, it erroneously duplicates the third scroll of the Lüxu in Scroll 61. This strongly suggests that the Matsuo Shrine Manuscript inherited the textual lineage of the Gogatsu-tsuitachi-kyō, along with its structural inconsistencies. The persistence of this anomaly into the twelfth century highlights the durability of transmission errors within Japanese manuscript Buddhist canons.

By contrast, other late Heian manuscripts—such as the Nanatsu-dera Manuscript and the Kōshō-ji Manuscript—successfully corrected this structural anomaly. This correction did not originate from within Japan, but was facilitated by the introduction of the Kaibao Canon, which was transmitted to Japan in the late tenth century. Prior to the arrival of the Kaibao Canon, all extant Japanese manuscript versions of the Shisong lü appear to have lacked the fourth scroll of the Tenth Recitation.

To understand why the Gogatsu-tsuitachi-kyō erroneously duplicated the third scroll of the Lüxu in Scroll 61, it is necessary to examine its compilation process. The imperial manuscript canon project for the Gogatsu-tsuitachi-kyō began in the fifth year of the Tenpyō era (733). By the seventh month of the eighth year (736 CE), the project had adopted a compilation policy based on the Kaiyuan shijiao lu, using Buddhist scriptures imported by Genbō 玄昉 (d. 746 CE) as the primary textual source. By the fourth month of the twelfth year (740 CE), the canon had reached a total of 3531 scrolls, at which point the Vow Text of Empress Kōmyō was appended to all scrolls.

In the thirteenth year of the Tenpyō era (741 CE), the project was transferred to the Fukuju-ji Scriptorium 福壽寺寫經所, which was later renamed the Konkōmyō-ji Scriptorium 金光明寺寫經所. By the fourteenth year of the Tenpyō era (742 CE), the canon had expanded to 4561 scrolls—approaching but still falling short of the Kaiyuan shijiao lu’s target of 5048 scrolls. At this point, acquiring additional source texts became increasingly difficult. Consequently, the scriptorium shifted its focus to compiling apocryphal and dubious scriptures, as well as Buddhist commentaries not listed in the Kaiyuan shijiao lu. This strategic shift ensured that the Gogatsu-tsuitachi-kyō would not only become Japan’s largest Buddhist canon, but also serve as a comprehensive repository of all Buddhist texts available in Japan at the time. By the eighth year of Tenpyō Shōhō era (756 CE), following the death of Emperor Shōmu, the manuscript canon was officially finalized30.

The Shisong lü within the Gogatsu-tsuitachi-kyō consists of 61 scrolls, but its compilation occurred in two distinct phases, drawing upon source texts from different textual lineages. The first phase (Tenpyō 9–12, 737–740 CE) involved the imperial scriptorium transcribing the initial 60 scrolls based on manuscripts imported by Genbō31. This manuscript set appears to have belonged to the Lüxu qianzhi type, but it notably lacked the final scroll of the Tenth Recitation (Scroll 61). As the project adhered to standards set by the Kaiyuan shijiao lu, the absence of this final scroll was noted, though no solution was available at the time. The second phase (Tenpyō 14–19, 742–747 CE) saw the Konkōmyō-ji Scriptorium discover an alternative version of Scroll 61 in the Inner Hall 内堂 (Naidō). This manuscript was transcribed and appended to the Gogatsu-tsuitachi-kyō in the fourteenth year of the Tenpyō era (742 CE). However, it was not fully incorporated into the canon until 746–747 CE (Tenpyō 18–19)32, indicating a period of uncertainty and deliberation regarding its inclusion. This delayed integration highlights the textual complexities surrounding the Shisong lü in Japanese manuscript Buddhist canons and reflects broader anxieties regarding textual authority and consistency in canonical transmission.

The newly discovered Scroll 61 contained two distinct texts: the “Yinyuan pin, the Fourth Section of [the Preface to the Shisong lü] 因緣品第四”—a duplication of Scroll 57 from the original 60-scroll set—and the Buddha’s Discourse on the Gravity of Violating Precepts 佛說犯戒罪報輕重經, an entirely separate Vinaya text33. This presented a dilemma for the scriptorium. Although the manuscript was labeled as Scroll 61, its contents did not correspond to the anticipated fourth scroll of the Tenth Recitation. The scriptorium faced two options: either to catalog the manuscript separately as an apocryphal Vinaya scripture or to append it to the existing 60-scroll set to maintain the intended numerical total. Initially, they chose the former, treating the scroll as an independent text. However, by the nineteenth year of the Tenpyō era (747 CE), as the Gogatsu-tsuitachi-kyō was being finalized as the official Buddhist canon of Tōdai-ji, they adopted the latter strategy—incorporating the duplicate Scroll 61 into the canon to ensure numerical completeness34.

This case highlights a fundamental tension in the compilation of the Gogatsu-tsuitachi-kyō. The scriptorium aspired to construct an ideal Buddhist canon in accordance with the authoritative Kaiyuan shijiao lu, yet practical constraints made this goal unattainable. Many texts—including the Shisong lü—were incomplete, and the extant Buddhist materials in Japan did not fully conform to Chinese canonical models. As a result, compromises were inevitable, leading to textual inconsistencies that would persist in Japanese manuscript canons for centuries. This analysis underscores the complexities of textual transmission in East Asia and the enduring challenges of preserving Buddhist scriptures with absolute fidelity.

6. Concluding Remarks

By tracing the textual evolution of the Shisong lü, this study has examined the origins and transmission of Japanese manuscript Buddhist canons from the Nara, Heian, and Kamakura periods within the broader context of East Asian Buddhist textual traditions. Since its initial Chinese translation in the fifth century, the Shisong lü underwent an early bifurcation into two lineages: the Northern lineage, preserved only in unearthed manuscripts from Dunhuang and Turfan, and the Southern lineage, which became the dominant version incorporated into Buddhist canons. Within the Southern lineage, two distinct textual traditions emerged: the Lüxu houzhi (Post-Appended Preface) version, found in printed Buddhist canons from the Song, Yuan, and Goryeo dynasties, and the Lüxu qianzhi (Pre-Appended Preface) version, preserved in Japanese manuscript Buddhist canons. While the former reflects the standardized textual tradition of mid-to-late Tang Buddhism, the latter retains pre-Kaiyuan textual features. Japanese manuscript Buddhist canons thus represent an alternative textual lineage, serving as a complementary counterpart to printed Buddhist canons and offering valuable insights into early Buddhist textual formations and the transmission of Vinaya texts in East Asia.

This study has demonstrated that Japanese manuscript Buddhist canons evolved through processes of adaptation, hybridization, and selective revision. The Nara-period Gogatsu-tsuitachi-kyō exemplifies this dynamic, having been reconstructed from a 60-scroll Lüxu qianzhi version combined with a single scroll from a Lüxu houzhi version, resulting in a structurally unique and internally redundant text. This hybrid form circulated in Japan for over two centuries, with its inconsistencies remaining uncorrected until the introduction of the Kaibao Canon in the late tenth century. Only thereafter did some Japanese manuscript Buddhist canons begin to align with printed canons, addressing textual discrepancies and moving toward greater standardization. These findings confirm that while Heian- and Kamakura-period manuscript Buddhist canons inherited textual elements from Nara-period sources, their transmission was neither direct nor static; rather, it involved continuous modification in response to new textual imports and evolving monastic needs.

As the final recipient of Buddhist textual culture in East Asia, Japan relied on external sources while actively adapting them to its own manuscript tradition. During the Nara period, Japan depended heavily on imported Buddhist texts, necessitating the integration and repurposing of extant materials. The subsequent arrival of printed Buddhist canons, particularly the Kaibao Canon, further reshaped textual practices, promoting increased standardization and contributing to a broader shift in textual authority. This transition from manuscript-based to print-based canonical authority marks a pivotal transformation in the history of East Asian Buddhist textuality and warrants further scholarly attention.

On a broader scale, this study highlights the distinct yet complementary roles of three major categories of Buddhist textual resources. First, unearthed Buddhist manuscripts from Dunhuang and Turfan preserve the Northern lineage of the Shisong lü, offering critical insights into its earliest textual forms and regional variations. Second, printed Buddhist canons from the Song, Yuan, and Goryeo dynasties reflect the standardized Southern lineage, demonstrating how editorial intervention and institutional oversight shaped the evolution of the Vinaya tradition. Third, Japanese manuscript Buddhist canons preserve pre-Kaiyuan features of the Southern lineage, providing crucial evidence for reconstructing earlier, pre-codified textual traditions that were later refined in printed canons.

A systematic integration of these three textual traditions—Dunhuang and Turfan manuscripts, Song-Yuan-Goryeo printed canons, and Japanese manuscript Buddhist canons—will allow for a more precise reconstruction of the historical evolution and transmission of the Shisong lü. Future research should continue to explore these interconnections to deepen our understanding of the formation, transmission, and institutionalization of Buddhist canons across East Asia. Ultimately, this integrative approach reinforces the philological foundations of Buddhist studies and sheds new light on the historical processes that shaped the transmission, adaptation, and preservation of Buddhist texts.