Visible Layouts, Hidden Dynamics: Reading, Reproducing, and Reframing Chinese Buddhist Glossaries

Abstract

This paper investigates how the layout strategies of Xuanying’s Yiqiejing yinyi (mid-7th c.), the earliest surviving Chinese Buddhist glossary, evolved across manuscripts, Buddhist Canon editions, and Qing-era scholarly reprints from the 7th to 19th centuries. While Xuanying’s work serves as the central case due to its breadth of preservation and representativeness, this study also references Huiyuan’s glossary (early-8th c.) to highlight broader patterns of reception and adaptation, particularly in late imperial China. Through a usability–production efficiency framework, the study identifies a continuum from the flexible manuscript layouts to the standardized double-line format used in Buddhist woodblock printing, and later to Qing-era adaptations that integrated Buddhist glossaries into evidential studies. It argues that layout decisions were influenced not merely by practical considerations of use and production but also by changing conceptions of textual function and authority. It also highlights the unintended effects of layout standardization, which at times introduced new interpretive complexities. By demonstrating how layout actively influenced the reproduction and reception of Buddhist glossaries, this study offers a new perspective on the intersection of materiality, textual transmission, and reading practices in pre-modern China.

1. Introduction

Yiqiejing yinyi 一切經音義 (Sound and Meaning of All the Scriptures, hereafter abridged as Xuanying’s Yinyi),1 composed by Xuanying 玄應 (d. ca. 663) in the early Tang dynasty (618–907),2 is the oldest surviving and relatively complete glossary of the Chinese Buddhist Canon. Drafted while Xuanying served as a proofreader at Xuanzang’s 玄奘 (602–664) translation bureau in Chang’an 長安,3 this twenty-five-juan glossary extracts and explains puzzling characters and phrases from Buddhist scriptures. Arranged in concordance with the sequence of interpreted texts, it endeavors to elucidate their pronunciation, meaning, and other relevant details to facilitate the comprehension of Buddhist texts.

Among Chinese Buddhist glossaries, Xuanying’s Yinyi is especially noteworthy—not only for its early date and paradigmatic status as a model for later works but also for the exceptional breadth of its textual transmission. It survives in more manuscript and block-printed versions than any other glossary of its genre, providing a uniquely rich corpus for examining long-term changes in layout across different periods and media. This article also discusses Huiyuan’s 慧苑 (b. ca. 674–678) Xinyi Dafangguang Fohuayan jing yinyi 新譯大方廣佛華嚴經音義 (Sound and Meaning of the Newly Translated Avatamsaka Sutra, hereafter abridged as Huiyuan’s Yinyi), which gained renewed prominence in the Qing period (1644–1911) alongside Xuanying’s Yinyi, especially among literati engaged in evidential studies (kaoju xue 考據學).4 Other glossaries, though of historical importance, are not discussed in detail here because they failed to achieve wide dissemination or gradually fell into obscurity.5

The surviving copies of Xuanying’s Yinyi, while transmitting the same work, exhibit remarkably diverse characteristics in both content and presentation. This variability stems from the text’s unique nature as a supplementary reference work: despite its inclusion in the Buddhist Canon, the glossary maintained greater flexibility than canonical scriptures, allowing for adjustments by copyists, readers, and collators. Its practical function and text-based organizational structure attracted an audience beyond the Buddhist community, necessitating various layout adaptations to accommodate different users’ needs. The extensive corpus of surviving documents thus offers valuable insights into how a single text could be presented and adapted in response to diverse user requirements and scribal practices.

Previous scholarship has classified extant copies of Xuanying’s Yinyi according to features such as script size, annotation structure, and the visual coherence between entry heads and annotations. While this typological approach has clarified the chronology and textual relationships among surviving copies, it has largely treated layout as a descriptive category rather than an interpretive one. As Nugent (2023, p. 238) reminds us, a text encompasses more than just the literal message conveyed by its words; it also includes how the words are arranged, styled, and presented. Thus, layout plays a crucial role in shaping how intended readers might engage with the text. Drawing on Gibson’s (2014, pp. 139–61) affordance theory, this study views manuscripts and printed books as interactive artifacts that structure engagement and guide interpretation. From this perspective, the layout of Buddhist glossaries offers a lens through which to examine editorial intention, reader interaction, and the shifting dynamics of textual transmission.

This paper approaches layout as a strategy of information management, investigating how it shaped textual function, authority, and editorial intervention. Drawing on a comprehensive range of currently accessible pre-modern versions, it emphasizes the continuity between manuscript and printed traditions—from fluid manuscript practices to standardized Buddhist woodblock-printed editions, and later to its repurposing in Qing evidential scholarship. Moving beyond conventional philological concerns with textual lineage, this study treats surviving copies as material traces of textual engagement, shaped by the hands and decisions of copyists and compilers.

The analysis begins by reassessing layout variation through a newly proposed usability–production efficiency framework, derived from close observation of surviving examples. It then turns to manuscript traditions, highlighting the diverse scribal strategies used in presenting Buddhist glossaries. The discussion next traces the emergence and consolidation of the double-line annotation format in Buddhist woodblock printing and examines the unintended consequences that accompanied its standardization. Finally, it considers the transformation of layout in Qing editions, where Buddhist glossaries were reformatted to cater to the needs of evidential studies. By integrating layout into the history of textual transmission and reception, this paper argues that layout is not a passive feature of textual reproduction, but an active force that shaped how texts were read, interpreted, and reproduced.

2. Reassessing Layout Variations: A Usability–Production Efficiency Framework

2.1. Mapping Existing Scholarship

Scholars have long noted the differences among the surviving copies of Xuanying’s Yinyi. Ishizuka and Ikeda (1991, p. 4) categorize manuscripts found in Western China and Japan into seven groups (see Figure 1), using three primary criteria: (1) whether entry heads and annotations share the same font size, (2) whether entry heads and annotations are written continuously—both within the same entry and across different entries—and (3) whether annotations appear in single or double lines. They link these layout styles to distinct historical periods, arguing that the earliest format (A) featured a uniform text size with a space left after the second line of annotations. Later formats (B, C) emphasized prominent entry heads, while other versions (D, E) adopted small double-line annotations, and others (F, G) used small single-line annotations.

Li (2022, 2023) later extends this classification system to early printed Buddhist Canons, including the Fuzhou Canon (Fuzhou zang 福州藏), Sixi Canon (Sixi zang 思溪藏), and Qisha Canon (Qisha zang 磧砂藏) editions produced between the 11th and 13th centuries. She notes that layout styles often vary not only across different juans within the same edition but also sometimes even within a single entry, where single- and double-line annotation formats may coexist. These observations underscore the fluidity and variability of layout, suggesting that it was not a fixed visual convention, but an adaptive element of textual transmission.

While Ishizuka and Ikeda’s framework provides a detailed typological breakdown, it also introduces an overly granular classification, focusing on minor spacing and typographic variations. This approach, while valuable for historical tracing, risks fragmenting the analysis and diverting attention from the more fundamental differences that shaped both textual production and reader engagement.

To address these limitations, layout should be conceptualized not merely as a visual feature but as a functional intersection between production constraints and usability needs. Copyists and compilers did not choose formats arbitrarily; instead, their decisions were shaped by practical considerations, such as reducing paper usage and ensuring the ease of production, as well as functional concerns, such as how effectively the text could be retrieved, read, and interpreted. At the same time, different layouts had direct consequences for readers, influencing retrievability, readability, and interpretive strategies.

2.2. A Usability–Production Model: Four Approaches to Layout

In light of these considerations, I propose simplifying the classification system into two primary criteria. (1) Entry Arrangement: Whether each new entry begins on a new line or continues on the current line. Writing continuously until the margin conserves space, whereas starting each new entry on a fresh line improves retrievability by making entries easier to locate and distinguish. (2) Annotation Structure: Whether annotations appear in a single line or are divided into two lines. Single-line annotations are simpler to produce and create a uniform layout, whereas double-line annotations, though more labor-intensive to arrange, highlight entry heads more distinctly and optimize page usage.

From these two criteria, four broad layout types emerge: (1) single-line annotation with each entry on a new line; (2) single-line annotation with continuous lines; (3) double-line annotation with each entry on a new line; (4) double-line annotation with continuous lines. Each layout carries distinct advantages and drawbacks when evaluated through the dual lens of usability (retrievability and reading flow) and production efficiency (efficient paper use and labor efficiency).

To illustrate these trade-offs, the radar chart (Figure 2) maps each layout type across four dimensions: retrievability, reading flow, material efficiency, and labor efficiency. The relative placement of each layout on these axes is not derived from quantitative measurement but from comparative analysis—based on recurring features observed across a broad range of manuscript and printed copies, as well as on how different configurations facilitate or hinder textual use and production. This chart does not aim to produce objective metrics, but to offer a heuristic model that visualizes the trade-offs inherent in different layout strategies (see note 6 for a detailed explanation of the evaluation criteria).7

The chart immediately reveals how each layout balances different priorities. For example, the single-line annotation with continuous lines (red shape) offers a better reading flow by reducing the need for frequent line breaks and maximizing material efficiency by conserving paper space. However, this format complicates retrievability, making it harder to locate specific entries unless additional visual markers—such as red dots or enlarged fonts—are introduced, increasing production complexity.

Conversely, double-line annotation with each entry on a separate line (blue shape) enhances retrievability by clearly differentiating entry heads, making them easier to locate. However, this format requires a precise alignment and often increases paper usage, demanding more labor-intensive production efforts.

Two particularly flattened diamond shapes stand out in the chart: the orange shape, representing single-line annotation with new lines per entry, and the green shape, representing double-line annotation with continuous lines. The trade-offs between these formats are striking. Single-line annotation with new lines excels in retrievability and labor efficiency, as it requires minimal formatting effort, but it ranks lower in material efficiency and reading flow due to the frequent breaks in text. In contrast, double-line annotation with continuous lines improves material efficiency and reading flow, yet sacrifices retrievability and labor efficiency by making entries harder to distinguish.

These stark differences between layout types provide a practical framework for understanding how copyists navigated competing demands—balancing textual organization, production feasibility, and the reading experience. More broadly, they reveal how layout decisions were not merely aesthetic choices but integral to the strategies of textual transmission and use.

Having established these four primary layout categories and their trade-offs, the next step is to examine how these formats are reflected in surviving manuscripts. Rather than being purely stylistic variations, different layouts represent deliberate information management strategies, shaped by the practical demands of canonical preservation, textual navigation, and scholarly study. The following sections will explore this continuum, tracing how layout standardization evolved in response to both historical circumstances and the shifting needs of Buddhist textual culture.

3. A Spectrum of Layout Strategies in Manuscripts

3.1. Canonical Conformity: Manuscripts Aligned with Scriptural Norms

Manuscripts of Buddhist glossaries featuring single-line annotation with line breaks frequently appear in formal, canonical-style copies, suggesting that they were intended for monastic collections, either as part of the Buddhist Canon or as institutional reference works. A representative example is Dunhuang manuscript S.3469 (Figure 3), which adopts visual conventions typical of canonical Buddhist scriptures, such as a uniform script size, consistent 17-character lines, and single-line entries that begin on separate lines. Notably, the manuscript adapts these conventions to highlight headwords more explicitly: when annotations extend beyond one line, subsequent lines are slightly indented—typically by the space of a single character—thereby helping to visually distinguish the start of new entries.

This practice produces visual regularity while reducing production complexity: each entry could be copied continuously with minimal formatting effort, unlike double-line annotations that require careful alignment. However, this same layout occasionally may also pose challenges for readers. Since headwords and annotations appear in the same script size and alignment, variation in headword length—even though most consist of two characters—can blur the internal boundary between a headword and its gloss. Meanwhile, when scribes fail to consistently begin new entries on separate lines, adjacent headwords may be conflated, as seen in manuscript P.3734, where entries “ji li” 亟立 and “shan shuo” 覢鑠 were mistakenly merged.9

When examined in the broader context of philological manuscripts excavated from Dunhuang—including rhyme books, graph dictionaries, exegetical texts, and glossaries across multiple traditions—the use of a layout in which headwords and annotations appear in the same script size without differentiation stands out in Xuanying’s Yinyi manuscripts. Surveying the philological texts collected in the Dunhuang jingbu wenxian heji 敦煌經部文獻合集10—spanning Buddhist, Confucian, and Daoist traditions—suggests a consistent tendency across genres to visually distinguish headwords from annotations. Most of these manuscripts employ a smaller script for annotations, whether they follow a single-line or double-line format.

In contrast, the majority of extant Xuanying’s Yinyi manuscripts from Dunhuang adopt a layout in which headwords and annotations appear in the same script size and alignment, using line breaks rather than changes in script size or alignment to indicate structural boundaries. This formatting deviates notably from common conventions observed in Dunhuang philological manuscripts. Ishizuka and Ikeda (1991, p. 3) have argued that this layout reflects the original form of Xuanying’s Yinyi, and that the practice of using a smaller script for annotations developed later. However, this explanation does not fully account for why Xuanying’s Yinyi emerged as such an outlier within the broader landscape of philological manuscript practices. Rather than attributing the anomaly solely to chronological evolution or manuscript survival bias, it may be more productive to consider how its formatting was shaped by the textual function, institutional role, or regional scribal practices. These factors offer a more comprehensive basis for understanding the rationale behind its canonical-leaning layout.

This distinctive layout can be better understood in light of Xuanying’s Yinyi’s unfinished status and its evolving role within the Buddhist Canon. According to Daoxuan’s 道宣 (596–667) preface to the work, Xuanying passed away before finalizing the work, leaving it in a provisional state subject to variation. Moreover, the glossary’s structure—covering approximately 460 Buddhist texts in sequence—enabled readers to consult it alongside a wide range of scriptures, embedding it within the logic of canonical study.

This close association is further reflected in Buddhist bibliographic traditions. Xuanying’s Yinyi is the only glossary listed in Daoxuan’s Datang neidian lu 大唐內典錄 (664), and it continues to appear in later catalogs such as the Kaiyuan shijiao lu 開元釋教錄 and the Zhenyuan xinding shijiao mulu 貞元新定釋教目錄. As an influential bibliographic model, the Datang neidian lu persisted in Dunhuang even through the 10th century, shaping how monastic libraries organized and classified their holdings (Fang 1997, pp. 26–28; Zhu 2022, p. 97).

Dunhuang sources further demonstrate the glossary’s integration into monastic collections. Catalogs from Longxing Monastery 龍興寺 (P.4039, S.2079) list Xuanying’s Yinyi as comprising 24 juans in three sections, while a lending record (P.3654) confirms its active use by a monk named Shenwei 神威. Catalogs compiled by Daozhen 道真, a monk at the Sanjie Monastery 三界寺 in the early 10th century, record an increase in extant juans—from eleven (BD14129) to thirteen (Dunyan 345)—suggesting continued efforts to complete the set through transcription.11 One manuscript (S.3538) even tracks this process, marking which juans were present, newly copied, or still missing, and recording scribal assignments and paper usage. These materials provide a rare insight into the management of monastic libraries, illustrating the meticulous efforts involved in preserving and replenishing Buddhist textual collections.

These examples demonstrate that Xuanying’s Yinyi was not merely a supplementary aid but an integrated component of Buddhist monastic collections. Its repeated appearance in monastic inventories, borrowing records, and copying plans reflects not just passive storage but a stable institutional presence. This structural association with the Buddhist Canon and its supporting apparatus may help explain why many of its manuscripts adopt a layout that visually echoes canonical sutra transcription, albeit with practical modifications such as the slight indentation of continuation lines.

Yet, this layout style recurs in manuscripts that differ widely in scribal execution and material context. For instance, Dunhuang manuscript P.3095 presents a roughly written version of Xuanying’s Yinyi in red ink on reused paper, with extended line lengths that suggest an attempt to maximize space.12 Such examples highlight how canonical formatting conventions were applied across diverse production settings. Rather than serving as a fixed marker of textual authority, layout was shaped by scribal habits, institutional norms, and pragmatic constraints. Its recurrence across both well-executed and hastily copied manuscripts reflects a negotiation between idealized visual norms and the practical realities of textual reproduction.

3.2. Pragmatic Adaptations: Manuscripts Oriented Toward Economy and Accessibility

Beyond manuscripts that closely resemble canonical formats, a group of Xuanying’s Yinyi manuscripts exhibits more pragmatic tendencies, emphasizing usability, economy, and customized adaptation to specific functional needs. These manuscripts often diverge from standardized patterns in both content and layout, reflecting how glossarial texts were reshaped to serve practical ends.

Some manuscripts, for instance, employ a double-line annotation format in which each gloss spans two lines in smaller script, and entries proceed continuously without line breaks between them. Unlike line-break formats—where each entry begins on a new line—this layout results in a denser, more compact structure, visually distinguished by the script size. The result is a highly compact layout that enhances information density and conserves paper—an especially valuable consideration in resource-constrained environments like Dunhuang. Compared to the more segmented line-break format, this design also offers a smoother reading experience for extended glosses by reducing the need for frequent visual resets.

However, this format also presents challenges for scribes. Because annotations had to be split evenly between two lines, scribes needed to estimate the length of glosses in advance to ensure a spatial balance. Errors in estimation led to inconsistencies: when too many characters were written in the first line, the second line appeared sparse; conversely, if the first line was too short, the second line became densely packed. Moreover, the absence of line breaks made retrieval less efficient, especially for readers scanning the page for specific headwords. In contrast to line-break formats—where entries begin on new lines and headwords align horizontally near the top of each line—this continuous layout distributes headwords unevenly throughout the page, making them harder to locate at a glance.

Dunhuang manuscript P.2901 (Figure 4) exemplifies a highly pragmatic adaptation of Xuanying’s Yinyi, characterized by a compressed layout and selective content retention.13 It applies a double-line continuous layout, economizing on paper while preserving a clear visual distinction between headwords and glosses. Yet analysis of the portion shown in Figure 4 indicates that over 50 percent of entries failed to achieve balanced spacing. In some cases, the first line was overcrowded, leaving the second line underused; in others, the distribution was reversed.

One example illustrates this challenge in practice. The entries for “bing yan” 并𩞚 (highlighted in yellow in Figure 4) and “tan zha” 嘆咤 (highlighted in blue) were mistakenly reversed in order. While transcribing “bing yan,” the scribe nearly reached the margin before switching to the second line, leaving insufficient space for the annotation of “tan zha”. To compensate, the scribe converted “tan zha” into a single-line annotation and even extracted the phonetic explanation of “zha” into a separate entry (highlighted in green) to fit within the remaining space. This was not an intentional restructuring of the text but rather an adaptation made out of necessity, illustrating the challenges of maintaining uniformity in a double-line continuous format.

Other manuscripts exhibit alternative strategies for organizing the layout to facilitate access or readability. For example, manuscript Φ.367 maintains a uniform script size throughout but introduces spacing and red dots before each new entry, offering a simple yet effective way to visually separate individual entries. These interventions reflect a user-oriented logic, prioritizing quick access and retrievability over rigid adherence to inherited layout conventions.

Together, these manuscripts suggest that scribes and readers treated glossarial texts not as fixed repositories, but as adaptable tools responsive to diverse copying conditions and user needs. Whether through compact double-line formats, punctuation-like devices, or selective gloss extraction, the layout and organization of Buddhist glossaries were continually negotiated and reconfigured. These interventions demonstrate a flexible approach to textual production—one that prioritized functionality and responded directly to the practical demands of manuscript culture.

3.3. Layout Variation and Functional Adaptation

The preceding sections have illustrated how different layout choices in Xuanying’s Yinyi manuscripts were shaped by overlapping factors—scribal practices, institutional expectations, material constraints, and functional goals. Rather than forming fixed categories, these layout types occupy a broad functional continuum. At one end lie standardized formats that closely align with canonical scripture presentation; at the other, more compact and adaptive formats reflect heightened attention to usability and economy.

The case of Xuanying’s Yinyi offers a particularly revealing lens through which to observe these variations. Its surviving manuscripts demonstrate a range of visual strategies that depart from conventional glossarial norms. Some closely follow canonical models, using a uniform script size and clear line breaks. Others reveal considerable adjustment in both structure and visual presentation—compressing or abbreviating content, omitting nonessential citations, or modifying the layout for greater efficiency.

These variations unfolded along two primary axes, reflecting how scribes visually structured information both across and within entries. One involves the visual structuring between entries: whether each new headword begins on a fresh line, whether red dots or spacing are used to mark new terms, and how compact or dispersed the entries appear on the page. The other concerns visual differentiation within each entry: how the headword is distinguished from the gloss, whether through font size, spacing, indentation, or alignment.

Manuscripts such as Ch/U.6788 and Ch/U.6782d, for example, maintain a single-line format but use script size variation to distinguish glosses from headwords. The sixth juan of Xuanying’s Yinyi manuscript preserved in the Shōgōzō 聖語藏 collection, meanwhile, adopts extra spacing between headwords and annotations to clarify the internal structure without interrupting the overall layout flow. In addition to spacing and script size adjustments, color and visual markers also played a role in enhancing text organization and retrievability. Manuscript Φ.367 employs red dots before each entry to help separate headwords, despite using a uniform character size throughout.

In some manuscripts, these strategies are applied consistently; in others, multiple layout types appear within a single text. In manuscripts from Kongō-ji 金剛寺 in Japan, both single- and double-line annotation formats are used within the same juan: shorter glosses occupy a single line, while longer ones are divided across two.15 These hybrid approaches suggest that layout decisions were sometimes modular, combining standard features with context-specific modifications.

Rather than treating layout as a passive result of scribal routine, these examples reveal that visual organization was itself a field of decision-making—shaped by purpose, audience, and physical conditions. Xuanying’s Yinyi was not simply transmitted in fixed form; its layout was continually reconfigured through the flexible and often improvisational handling of visual structure.

While these layout strategies are well-attested in Buddhist manuscripts, similar features—such as double-line glosses and color markings—also appear in other manuscripts, such as rhyme books and glossaries for Confucian classics and literary anthologies.16 Such parallels point to broader patterns of material adaptation across manuscript traditions, though further research is needed to clarify the extent of shared influence and mutual borrowing.

The flexible layouts preserved in Xuanying’s Yinyi manuscripts—ranging from canonical emulation to customized condensation—provided the foundational repertoire for later printed formats. As the shift to woodblock printing introduced new constraints and opportunities, some of these manuscript layouts persisted, while others were gradually eliminated. The next section will examine how these visual strategies were selectively retained, adapted, or standardized in the transition from manuscript to print.

4. The Triumph of the Double-Line Format: Standardization of Buddhist Block Printing

Following its inclusion in the Chinese Buddhist Canon, Xuanying’s Yinyi was continually reproduced in nearly every major block-printed edition of the Canon from the late 10th century to the 18th century.17 Scholars generally divide these editions into two main lineages: the Northern lineage, represented by the two editions of Goryeo Canons (Goryeo jang 高麗藏) and the Jin Canon (Jing zang 金藏), and the Southern lineage, exemplified by the Qisha Canon (Qisha zang 磧砂藏).18 Apart from textual differences, these two traditions also exhibit distinct layout conventions.

From the late 18th century onward, Xuanying’s Yinyi and Huiyuan’s Yinyi also attracted the attention of Qing literati, leading to the production of a series of scholarly editions. As shown in Figure 5, applying the four-category typology introduced earlier to juan 6 of Xuanying’s Yinyi reveals a stark contrast between early block-printed editions (10th–13th centuries) from the two lineages. Northern editions typically use single-line annotations with 14 characters per line, while Southern editions adopt double-line annotations with 17 characters per line. Within the Northern lineage (left side of Figure 5), the layout remained relatively stable across both officially sponsored and privately printed editions. In contrast, the Southern lineage (right side of Figure 5) underwent a gradual process of layout evolution and standardization. This section focuses primarily on that Southern trajectory, while also drawing comparisons with Northern exemplars. A full list of the printed editions referenced in this study is included in Appendix A.

While early block-printed editions largely preserved the layout diversity inherited from manuscript traditions, subsequent developments—particularly within the Southern lineage—led to a progressive standardization of layout, culminating in the dominance of the double-line format. This section examines how this shift occurred, exploring both the persistence of manuscript conventions in early block printing and the eventual triumph of the double-line format as the standard for Buddhist glossaries.

4.1. Early Block-Printed Editions: The Persistence of Manuscript Conventions

As discussed in the previous section, Buddhist glossary manuscripts exhibited significant variation in layout. The transition from manuscript to print did not immediately impose rigid standardization. Rather, early block-printed editions continued to reflect the diversity of their manuscript sources, carrying forward the inherent flexibility of handwritten transmission. As Fang (2006, p. 23) points out, manuscript Buddhist Canons, being produced by hand, inherently embody a dual characteristic of uniqueness and fluidity, which shaped the early phase of Buddhist block printing.

At the onset of block printing, technological expertise was still in its early stages, and printers often relied on existing manuscripts as models for engraved copies. The most efficient and error-minimizing strategy was simply to reproduce manuscript layouts directly rather than reformat the text. Changing the layout required recalculating spacing and adjusting column structures, which not only slowed the engraving process but also increased the risk of errors, omissions, or transpositional mistakes. As Liu (2021, pp. 174–76) argues, while the original motivation for preserving manuscript formats may have been efficiency, the result was a relatively faithful transmission of textual content. In this context, maintaining layout continuity with manuscripts was not only a natural choice but also a highly practical information management strategy.

This tendency to preserve manuscript layout conventions was not unique to Buddhist texts. A similar phenomenon can be observed in Confucian classics and literary anthologies from the 10th to the 13th centuries. Wang Guowei 王國維 (1877–1927) notes that directorate editions (jianben 監本), published under state sponsorship in the early period of woodblock printing, closely resembled Tang manuscripts in format and size.19 Liu (2021, pp. 176–79) further demonstrates that early Song-printed literary collections, such as the Wenxuan 文選, retained many formatting conventions of Tang manuscripts. The same tendency can be observed in Buddhist texts, where early block-printed Canons displayed inconsistent formatting, reflecting the diversity of their manuscript sources.

The Fuzhou Canon, the earliest representative of the Southern lineage, illustrates this phenomenon most vividly. Within its edition of Xuanying’s Yinyi, three distinct layout formats coexisted. Juans 1 and 25 employed single-line annotations with line breaks, juans 2 and 6 used double-line annotations with continuous text, while the remaining juans adopted double-line annotations with line breaks.20 Wang Guowei, who borrowed a copy of the Fuzhou edition from Luo Zhenyu 羅振玉 (1866–1940) in 1918 to compare with the 1869 scholarly edition of Xuanying’s Yinyi, had already observed these layout inconsistencies. He specifically noted that the first and sixth juans differed from the rest and suggested that juan 1’s content had been significantly abridged due to a shift from a double-line to a single-line format during printing. He reasoned that since the majority of juans used double-line annotations, the extensive omissions in juan 1 must have been caused by a deliberate reformatting decision.21

While Wang Guowei attributed these omissions to a deliberate editorial decision during the transition to block printing, a closer examination of its structure and comparison with contemporary manuscripts suggests that this inconsistency is better explained as a direct consequence of the diverse manuscript sources used for its compilation. Rather than being the result of an intentional reformatting process, the mixed layout more likely reflects the manuscript traditions they inherited.

This interpretation is supported by evidence from the Fuzhou edition, which offers no clear editorial rationale for altering the layout. Within the Fuzhou Canon, glossaries were consistently formatted with double-line annotations, making it unlikely that certain juans of Xuanying’s Yinyi would have been reformatted into a single-line layout. Apart from juans 1 and 25, all other juans of Xuanying’s Yinyi adopt a double-line format. While two of these juans use a continuous double-line format rather than breaking at each new entry, the dominant layout across the text remains double-line.

This preference for double-line annotations extends beyond Xuanying’s Yinyi. Huiyuan’s Yinyi, which appears adjacent to it in the Fuzhou Canon, also maintains double-line annotations, despite some variation between line-break and continuous styles. Phonetic annotations appended at the end of most cases in the Fuzhou Canon are also in the double-line format. This practice of including phonetic glosses at the end of canonical texts originated with the Fuzhou Canon and suggests a consistent editorial preference for double-line glossaries. Further supporting this pattern, Yamada (2009, pp. 4–5) observes that two cases in the Fuzhou Canon contain phonetic annotations that do not conform to the standard 17-character-per-line format, yet they still retain double-line annotations.

These observations also indicate that the editors were not engaged in a systematic effort to reformat the text, but were instead preserving the format of their manuscript sources. This editorial tolerance for format inconsistencies aligns with broader trends in early block-printed Buddhist Canons. Even within the Northern lineage, which predominantly employs single-line annotations with line breaks, occasional deviations appear. For instance, the Goryeo First edition sometimes arranges phonetic glosses in double-line format, particularly when annotations contain only phonetic information. The presence of such exceptions suggests that formatting choices were often dictated by practical considerations rather than a rigid editorial policy. Given this broader pattern, the mixed formatting in the Fuzhou edition of Xuanying’s Yinyi is more plausibly a product of manuscript transmission rather than deliberate intervention.

A comparison of the Fuzhou edition with both manuscripts and other printed editions of Xuanying’s Yinyi—such as the Jin and Goryeo Second Editions—further challenges the claim that the Fuzhou edition reflects a deliberate editorial intervention. The sections of the text that are “missing” or “abbreviated” in the Fuzhou editions overwhelmingly consist of explanatory annotations, particularly source citations. For instance, in juan 1, the entries for “ming e” 名遏 and “yi mu” 翳目 provide a clear example (see

Table 1. Two entries from juan 1 of Xuanying’s Yinyi.

	Ming e 名遏	Yi mu 翳目
Haneda 56R	古文閼，同。安曷反。 《蒼頡篇》：遏，遮也。 止也。	《韻集》作瞖，同。於計反。目病也。 《說文》：目病生翳也。 經文有作曀，陰而風曰曀。曀非此義也。
Fuzhou	古文閼，同。安葛反。 《蒼頡篇》：遏，遮也。止也。 亦絕也。	《韻集》作瞖，同。於計反。目病也。 《說文》：目病生翳也。 經文有作曀，陰而風曰曀。曀非此義。
Jin and Goryeo Second	古文閼，同。安曷反。 《蒼頡篇》：遏，遮也。 《詩傳》曰：遏，止也。亦絕也。	《韻集》作瞖，同。於計反。瞖，目病也。《說文》：目病生翳也。並作翳。 《韻集》作瞖，近字也。經文有作曀，陰而風曰曀。曀非此義。

).

When compared with Dunhuang manuscript Haneda 56R,22 the Fuzhou edition, the Jin edition, and the Goryeo Second edition, all of which follow a single-line layout with line breaks, it becomes evident that Haneda 56R and the Fuzhou edition exhibit striking similarities. Unlike the Northern lineage editions, the Fuzhou edition occasionally lacks certain types of information, including source citations, explanatory repetitions, and redundant characters or overly detailed explanations. These variations reflect a pattern commonly associated with customized manuscripts.

The “missing” content in the Fuzhou edition can be divided into three categories, each of which is marked differently in Table 1. Source citations, such as the Shizhuan 詩傳 citation in the “ming e” entry, are indicated by single underlining. Explanatory repetitions also disappeared; for instance, the repeated citation of the Yunji 韻集 in the “yi mu” entry is marked with double underlining. Redundant characters or overly detailed explanations are marked with wavy underlining.

This suggests that the missing portions of the text may not have been the result of active editorial intervention during the block-printing process. Rather, they were more likely inherited from a manuscript source that had already undergone selective adaptation. Given that abridgment and omission of source citations and redundant information were common practices in manuscript copies used for personal or pedagogical study, it is more likely that the Fuzhou edition simply preserved these pre-existing manuscript characteristics.

Additional textual irregularities support this conclusion. Both juans 1 and 25 employ a single-line format with 17 characters per line, conforming to manuscript formatting convention23 rather than reflecting a deliberate departure from the double-line standard. Furthermore, certain entries show evidence of textual misalignment, such as the merging of consecutive headwords. In juan 1, “mi lun” 弥綸 is not given a separate line but instead blends into the previous entry, “qun meng” 群萌. Similar instances of confusion between headwords and annotations occur, especially when the headword consists of more than two characters, such as “mi hou jiang” 獼猴江 in juan 14 and “duo jia liu xiang” 多伽留香 in juan 20. These inconsistencies point to scribal transmission errors typical of manuscript culture, further reinforcing the idea that the Fuzhou Canon was preserving, rather than modifying, its sources.

As indicated above, the presence of three distinct layouts within the Fuzhou edition aligns with pre-existing manuscript traditions, suggesting that these variations were inherited rather than deliberately introduced. Apart from the two layouts (single/new line and double/continuous) discussed in the previous section, for instance, the Shōgozō manuscript of juan 6, copied in Japan during the Tenpyō 天平 era (729–749), follows a double-line format with line breaks, which corresponds to the most common style observed in the Fuzhou edition. This continuity indicates that the double-line format had been an established convention long before the Fuzhou Canon was compiled.

Given that Buddhist manuscript transmission frequently involved piecemeal copying and supplementation, it is likely that the compilers of the Fuzhou edition assembled Xuanying’s Yinyi from multiple incomplete sources, each reflecting different scribal traditions and formatting conventions. This would explain why the Fuzhou edition preserves a mixture of single-line and double-line layouts rather than adhering to a uniform standard.

Furthermore, the two juans that diverge from the dominant pattern—juans 2 and 6—employ a double-line continuous format. Notably, these two juans correspond to glossaries for the Mahaparinirvana Sutra and the Lotus Sutra, two of the most widely studied texts in the Buddhist Canon. The layout used here closely resembles the double-line continuous format discussed in the previous section, suggesting a degree of continuity between earlier manuscript practices and the variation observed in the Fuzhou edition.

In light of this, the mixed formatting of Xuanying’s Yinyi in the Fuzhou edition is more plausibly a result of inherited manuscript diversity rather than deliberate editorial intervention. As the Canon drew on early Northern Song exemplars circulating in Fuzhou and its surrounding regions,24 the compilers appear to have retained the layouts of their source materials rather than imposing a unified format. This continuity underscores how early block printing was still closely tied to practices of manuscript transmission.

The persistence of manuscript conventions in early block-printed editions demonstrates that the transition to print did not immediately impose uniformity. However, as block-printing technology advanced and large-scale Buddhist Canon production became more systematized, a shift toward greater standardization took place. The following section will explore how this process unfolded, examining the factors that led to the eventual dominance of the double-line format and its implications for the organization and transmission of Buddhist glossaries.

4.2. The Path to Standardization: Xuanying’s Yinyi in the Southern Buddhist Editions

Having examined the two exceptional formats in the Fuzhou edition, this section shifts focus to its dominant layout—the double line with line breaks format—which was subsequently adopted in later Southern Buddhist Canons, leaving a lasting impact. Over the six centuries from the 11th to the 17th century, Buddhist block-printed texts in the Southern tradition underwent a gradual process of standardization. This evolution is reflected in two key aspects: the overall unification of layout conventions and the refinement of textual content, particularly in adjusting the spacing and alignment of annotations within the double-line format.

Successive Buddhist Canons show a clear trend toward greater uniformity in layout. As illustrated in Figure 6, aside from the four juans with exceptional formats discussed earlier, the dominant layout in the Fuzhou edition is the double line with line breaks format. Within this dominant type, two sub-variants can be identified based on whether a space is inserted between the entry head and its annotation: double/new line with space (12 juans) and double/new line without space (9 juans), yielding a ratio of roughly 4:3.

The Sixi edition, which was compiled in a similar period as the Fuzhou edition, largely retained this mixed pattern but introduced minor adjustments within the double/new line category. Specifically, juans 4 and 21, which had previously followed the without space format, were modified to include spacing. As a result, aside from juans 1, 2, 6, and 25, which followed non-double/new line formats, the remaining sections achieved a greater degree of internal consistency. In this revised structure, juans 8–14 continued using double/new line without a space, while the rest adopted double/new line with a space, improving overall uniformity.

By the early 14th century, the Qisha edition further refined the format, establishing double/new line as the universal standard across all juans. Although some inconsistencies remained—three juans retained the with space format, while the remaining 22 adopted without a space—the dominance of the double/new line format had been fully established. Subsequent editions, including the Puning Canon (Puning zang 普寧藏), Hongwu Southern Canon (Hongwu nan zang 洪武南藏), and Yongle Southern Canon (Yongle nan zang 永樂南藏) editions, closely followed the Qisha edition’s formatting conventions. Complete uniformity was only achieved with the Yongle Northern Canon (Yongle bei zang 永樂北藏) edition in the mid-15th century, where all juans were standardized to the double/new line without space format. Finally, the Jiaxing Canon (Jiaxing zang 嘉興藏) edition, completed in the late 17th century, preserved this uniformity while adapting the format to suit the transition from accordion-style to bound-book format, thereby marking the final stage of full format standardization.

Beyond the overall standardization of layout, refinements in textual content also played a crucial role in shaping the final format. Any modifications in character count inevitably impacted the page arrangement, necessitating deliberate strategies to maintain coherence within the double-line format. A close examination of early privately printed editions—the Fuzhou, Sixi, and Qisha editions—reveals how these adjustments were implemented to balance textual accuracy with spatial organization.

A clear example is the first entry of juan 6, “qi she ku shan” (Figure 7). The Fuzhou edition annotates this term using 98 characters, while the later Sixi and Qisha editions expanded the annotation to 118 characters, aligning with the Northern Buddhist Canon tradition. Despite this 20-character increase,25 the placement of entry heads (highlighted in orange) remains almost identical between the Fuzhou and Sixi editions. This raises the question: How did the Sixi edition accommodate the additional content without altering entry placement?

A careful comparison of the layouts reveals that while the Fuzhou and Sixi editions appear nearly identical in entry head placement, the annotation density differs. The Sixi edition’s annotations, when compared to the neatly arranged Fuzhou edition, appear crowded and unevenly distributed (Figure 8). This discrepancy is particularly evident when juxtaposing Sixi with the Fuzhou and Qisha editions, highlighting a gradual shift toward textual consolidation.

Analysis of the Sixi edition’s page structure suggests a consistent strategy of incorporating textual emendations without overhauling the existing layout. For instance, in one half-leaf unit from the first juan of the Lotus Sutra annotations, lines 1–2 and 7–8 conform to the standard 17-character-per-line format, while lines 3–6—those without entry heads—carry a higher character density. This suggests that the Sixi compilers deliberately preserved the positions of entry heads as anchor points and adjusted only the content of intervening annotation lines. In doing so, they avoided disrupting the overall structure while accommodating textual revisions.

While that adjustment strategy created minor inconsistencies, this method of modulating annotation density within non-headword lines allowed for localized adjustments, balancing accuracy and production efficiency. Because the majority of entries remained unchanged, a full reformatting would have been prohibitively labor-intensive. By selectively compressing content in specific sections, the Sixi edition integrated new material without extensive disruption.

This layout strategy is not limited to a single instance. Character count variations appear throughout the glossary, indicating a systematic application of this adjustment approach. While the Fuzhou edition adhered closely to the 17-character rule, the Sixi edition occasionally exceeded this limit—sometimes with even more than 20 characters per line—resulting in visibly compressed lines. These irregularities were not signs of scribal negligence, but the outcome of editorial interventions aimed at incorporating new content without reformatting the entire page.

Building on this approach, the Qisha edition introduced further refinements. It largely retained the textual emendations found in the Sixi edition but adopted a more deliberate and consistent layout strategy. While the entry heads in the Fuzhou and Sixi editions (highlighted in orange in Figure 7) begin at the topmost line and are positioned directly after the previous entry, those in the Qisha edition (highlighted in blue) are slightly indented rather than aligned flush with the margin. Annotations typically begin three-character spaces below the top of the page—except in cases where the headword contains more than two characters. These refinements not only enhanced visual clarity but also contributed to a more stable and legible presentation.

In sum, these strategies illustrate how Southern system editions gradually evolved toward an increasingly structured and reader-friendly visual format. Through a series of editorial refinements, these editions brought textual fidelity into closer alignment with visual hierarchy, laying the groundwork for the standardized layouts that would define later printed editions.

4.3. The Dominance of the Double-Line Format: Key Contributing Factors

Among the various layout styles used in Buddhist glossaries, the double-line annotation format with line breaks gradually became the dominant standard in the Southern Buddhist Canons. Its widespread adoption was not the result of a single decision or institutional mandate, but rather the product of multiple intersecting factors: its functional adaptability, its resonance with broader textual trends, and the material infrastructure that underpinned large-scale Buddhist publishing projects.

From a functional perspective, the double-line format offered a compelling balance between clarity, efficiency, and adaptability—qualities especially suited to glossarial texts like Xuanying’s Yinyi. By visually separating headwords from glosses—often through script size variation or slight indentation—and by assigning each entry to its own line, this layout enhanced readability and reduced cognitive load. Readers could scan for headwords more easily, particularly across the upper portions of a page, facilitating non-linear consultation typical of reference works. At the same time, line breaks reduced the alignment complications seen in continuous double-line formats, while preserving the compactness that single-line formats often sacrificed.

This format also facilitated production: compilers and engravers could modify individual entries without recalibrating entire pages—a key advantage for canonical editions that required updates or collations. The Qisha edition, for instance, subtly exceeded the standard 17-character line limit in some cases to prevent awkward breaks, reflecting a calculated flexibility in balancing legibility and spatial economy.

Beyond practical considerations, the rise of the double-line layout also reflects broader cultural trends toward visual stratification of textual layers. This format resonated with a growing emphasis on distinguishing primary texts from their interpretive supplements—a concern not unique to Buddhist works. As mentioned in the last session, glossary manuscripts, including both Buddhist and non-Buddhist, from Western China and Japan already employed layout strategies to separate headwords from glosses through script differentiation.

From the twelfth century onward, similar formatting practices emerged in the printing of Confucian classics, where the combined format of main texts and annotations (jingzhu heke 經注合刻) became increasingly popular.27 In these editions, commentarial and glossarial materials were visually subordinated to the classics they accompanied, often using double-line annotations to reinforce this hierarchy. While there is no direct evidence that Buddhist compilers intentionally borrowed from Confucian practices, the structural similarities—and the shared intellectual preoccupation with textual authority and clarity—suggest that the adoption of the double-line format was embedded within a broader shift toward hierarchical textual presentation.

Finally, the consolidation of this layout was materially reinforced through the institutional and artisanal infrastructure that underpinned large-scale Buddhist publishing. Once the double-line format with line breaks had been adopted, the sheer scale of Buddhist Canon engraving projects helped entrench it. Standardization was essential for the efficiency of such enterprises, and once a layout proved functional, compilers were generally reluctant to modify it. This tendency was especially evident in privately sponsored editions, where limited funding and labor rendered reformatting a costly and impractical risk. Unlike state-sponsored engravings, which possessed greater resources, private Buddhist publishers tended to preserve inherited layouts even as they continued to engage in active textual collation. This inertia became particularly apparent in later Buddhist editions, especially those succeeding the Qisha Canon.

The recurrence of some engravers’ names across different editions of Xuanying’s Yinyi—including the Chongning 崇寧, Pilu 毗盧, and Sixi Canons—suggests an interlinked network of artisans whose shared practices may have helped stabilize formatting conventions.28 Although layout decisions were typically made by compilers or scribes, engravers often worked across a range of Buddhist and non-Buddhist projects—including Confucian glossaries, historical works, and literary anthologies. This shared involvement allowed them to transmit and reinforce prevailing layout norms across genres and regions.29 Although such evidence does not directly attribute layout choices to engravers, it provides important circumstantial insight into how certain formatting strategies circulated across genres and regions, contributing to a shared repertoire of visual organization for glossarial and commentarial texts.

Overall, the triumph of the double line with line breaks format was not accidental, but the result of a long process of adaptation, efficiency-driven standardization, and intellectual evolution. While early printed editions exhibited the persistence of manuscript conventions, with multiple formats coexisting, the intrinsic advantages of the double-line format made it ideal for usability and production. Its standardization was reinforced by the inertia of large-scale engraving projects and broader textual trends across multiple genres, further legitimizing its role as the dominant format.

Yet, this triumph was not without complications. In later printed editions, efforts to incorporate textual emendations without disrupting the existing layout gave rise to new problems. Since most variants in Xuanying’s Yinyi appear in the annotations rather than in the headwords, compilers often retained the original position of entry heads while adjusting only the annotation lines. This strategy minimized labor but sometimes produced unintended effects on the page.

One such complication was spatial dislocation, referring to cases where the integration of revised annotation content—without reconfiguring the surrounding layout—resulted in visual misalignment and textual disruption. In an effort to preserve the placement of entry headings while incorporating textual emendations, compilers often adjusted only the annotations. However, this approach could distort the alignment between headwords and their glosses: newly inserted or expanded glosses might overflow into adjacent spaces, compress surrounding content, or interrupt the internal logic of entries. These layout-induced distortions sometimes affected not only individual entries but also the sequence in which entries appeared, introducing new variants or interpretive ambiguities. Rather than being a neutral backdrop, layout here shaped—subtly but consequentially—the transmission and coherence of textual content.30

Another strategy used to preserve visual uniformity was the adjustment of filler characters— final particles with little or no semantic function, such as ye 也, zhi 之, or yi 矣—into the blank space at the end of entries. In the double-line annotation format, which sometimes created uneven line endings, these characters were inserted to maintain visual balance and column alignment. While this practice produced a neater appearance, the presence and treatment of these fillers varied across editions. Later attempts to standardize them—whether by deletion, simplification, or substitution—sometimes introduced inconsistencies or gave rise to unintended misreadings.31

The triumph of the double-line format thus reflects both the achievements and limitations of standardization. While its consistency and broad applicability established a dominant framework for textual transmission, the rigidity of its structure sometimes hindered textual adaptability. These complications highlight how layout was not merely a passive visual element but a critical component of the editorial process—one that directly shaped the transmission, reception, and interpretation of Buddhist glossarial texts in the print era.

5. The Revival of the Double-Line Continuous Format: Scholarly Editions in Late Imperial China

Outside the Buddhist community, Xuanying’s Yinyi remained largely obscure, buried within the Buddhist Canon for centuries. Apart from occasional citations in graph dictionaries and encyclopedic works, Buddhist glossaries received little attention from the literati. However, this situation changed significantly in the late 18th century. With the rise of evidential scholarship, Buddhist glossaries attracted scholarly interest due to their extensive citations of earlier philological works, many of which had otherwise been lost. This changed in the late 18th century, when Buddhist glossaries attracted renewed attention for preserving earlier philological materials. Once peripheral to the Confucian classical tradition, these texts became integral to late imperial scholarly practices.

In 1786, Zhuang Xin 莊炘 (1735–1818) collaborated with his peers to collate and publish the Xuanying’s Yinyi. This edition became the base text for subsequent scholarly editions and had a profound impact. Due to sustained interest and high demand, four additional literati editions of Xuanying’s Yinyi were produced between 1786 and 1869.³² Among them, the 1869 edition by Cao Zhou 曹籀 (1800–ca. 1877) was particularly significant, serving as the culmination of literati engagement with Buddhist glossaries. Cao not only reprinted Zhuang Xin’s edition of Xuanying’s Yinyi but also included another influential Buddhist glossary, Huiyuan’s Yinyi.

In late imperial China, Xuanying’s Yinyi and Huiyuan’s Yinyi continued to receive scholarly attention and were often studied in tandem, warranting their joint discussion here. Starting with Zang Yong’s 臧庸 (1767–1811) 1799 edition of Huiyuan’s Yinyi, a total of five scholarly editions of the text appeared,33 along with one combined edition with Xuanying’s Yinyi—the Cao Zhou edition.

5.1. Reformatting for Evidential Scholarship: The Shift from Buddhist to Scholarly Editions

The scholarly editions of Buddhist glossaries were all based on southern Buddhist print editions, such as the Yongle Southern edition or the Jiaxing edition, which can be considered part of the same textual lineage. However, these scholarly editions diverged significantly from their Buddhist counterparts in both content and format. In terms of content, the scholarly editions incorporated annotations by literati, along with prefaces, colophons, and critical evaluations of the text. Some editions also introduced omissions; for instance, Zang Yong’s edition of Huiyuan’s Yinyi deliberately excluded explanations of Buddhist terms and transliterations that he deemed irrelevant to Confucian exegetical studies.

From a formal perspective, the layout of scholarly editions also underwent significant modifications. Zhuang Xin’s edition of Xuanying’s Yinyi, for example, was based on the Yongle Southern edition, but it altered the line arrangement from double line with line breaks to double-line continuous text. A comparison of paper usage between Buddhist and scholarly editions—as exemplified by the Jiaxing and Zhuang Xin editions—indicates that this shift reduced the average number of sheets per juan from 25 to 15, resulting in denser and more compact pages.34 Fewer pages also meant fewer woodblocks were required for printing, reducing production costs.

Another key modification was the consolidation of juans. Buddhist editions typically followed a one-juan-per-fascicle format, grouping ten juans into a case, with cases labeled according to the Qianzi wen 千字文. For example, in the Yongle Southern edition of Xuanying’s Yinyi, the text was divided into the three cases: Yun 云 (5 juans), Ting 亭 (10 juans), and Yan 雁 (10 juans). In contrast, the scholarly editions abandoned this Buddhist organizational structure and merged juans. Zhuang Xin’s, Ruan Yuan’s 阮元 (1764–1849), and Cao Zhou’s editions generally grouped every six juans into a single fascicle (ce 冊), resulting in four juans in total (the fourth volume comprising juans 19–25). The Guxi tang 古稀堂 and Pan Shicheng’s 潘仕成 (1804–1873) editions adopted a different approach, grouping four to five juans per volume, yielding six juans in total (except for the final volume, which contained only juans 23–25). For Huiyuan’s Yinyi, regardless of whether it followed the two-juan or four-juan edition, it was compiled into a single volume.35

These adjustments, while seemingly technical, reflect shifting emphases within the usability–production efficiency framework: the transition from double-line with line break to a continuous layout, in particular, reveals a move toward higher material efficiency, even at the cost of retrievability. Yet these layout changes were driven by more than just economic considerations; they reflect the literati’s shifting perception of Buddhist glossaries.

Buddhist texts, as carriers of both knowledge and faith, were considered sacred objects, and the act of copying or printing them was seen as an accumulation of religious merit. As Wu (2015, pp. 46–78) has noted, the Buddhist Canon itself became the object of devotional practices—a phenomenon he describes as the “cult of the canon”. While Buddhist glossaries were not scriptures in a strict sense, their inclusion in the Buddhist Canon imbued them with religious significance.

Against this backdrop, the reprinting of Xuanying’s and Huiyuan’s glossaries in Qing scholarly editions marked a significant shift: these works were extracted from the Buddhist Canon and repurposed for evidential study, effectively stripping them of their religious dimension and foregrounding their value as repositories of philological resources. This transformation in perception—away from sacred object toward scholarly reference—was also expressed through changes in layout. A comparison of layout styles (Figure 9) shows that when the literati printed Buddhist glossaries, they adopted conventions closely resembling those employed in the Jingdian shiwen 經典釋文, a late sixth-century glossary to the Confucian and Daoist classics that became highly influential in later exegetical traditions.

These changes in layout also reveal a transformation in how the literati engaged with Buddhist glossaries. Since their primary interest lay in the citations embedded within these texts, they focused less on the entry heads extracted from Buddhist scriptures and more on the sources cited in the annotations. Marginalia found in surviving scholarly editions indicate that Qing literati frequently scanned these glossaries for references to philological texts to support their own textual research.

For example, Tao Fangqi 陶方琦 (1845–1884), who worked extensively on reconstructing the Cangjie pian 倉頡篇, marked many book titles with red circles, including Guangya 廣雅 and Cangjie pian (Figure 10). The advantage of the Buddhist layout—which facilitated cross-referencing with scriptures or searching for specific entry heads—thus became irrelevant to the scholarly editions’ intended audience.

5.2. Antiquity Reconstructed: Chen Zongyi’s Editorial Approach

Chen Zongyi’s 陳宗彝 (d. 1841) edition of Huiyuan’s Yinyi stands as the sole exception to the broader shift away from the Buddhist layout.39 This edition preserved the format of the Yongle Southern edition, retaining its two-volume structure and distinctive double-line wrapping layout. The main text was printed in an elegant fangsong 仿宋 script, while the prefaces and colophons were reproduced from handwritten copies. Although comprising fewer than 130 pages, the production of this edition spanned eleven years.

The decision to adopt such a time-consuming, labor-intensive, and costly approach was not driven by abundant financial resources. On the contrary, securing funding was one of the greatest obstacles to the project. Despite receiving financial support from his mentor, peers, and even contributions from individuals motivated by Buddhist merit-based fundraising, progress remained slow. The project ultimately came to fruition due to substantial backing from Li Zhangyu 李璋煜 (1784–1857), a local official in Jiangning 江寧 at the time.

Chen’s editorial choices did not stem from a personal commitment to Buddhism or an explicit intent to uphold Buddhist textual traditions. In his preface, he underscored the glossary’s importance for preserving pre-Tang texts and its relevance to Confucian exegetical studies. He meticulously cataloged the cited works to facilitate scholarly reference, indicating that, like many of his contemporaries, he primarily regarded Buddhist glossaries as repositories of philological knowledge.

However, Chen’s approach was closely linked to his deep reverence for antiquity. Referring to himself as Qigu 耆古 (“Fond of the Ancient”), he dedicated his career to evidential studies and epigraphy, producing a significant number of scholarly editions despite his modest background. Prior to publishing this edition, he carefully collated the two-juan Yongle Southern edition against Sun Xingyan’s 孫星衍 (1753–1818) manuscript copy of the four-juan Jiaxing edition. For Chen, layout was not merely a formal or practical consideration but an essential aspect of textual authenticity and authority. His editorial strategy illustrates that layout choices serve as deliberate reflections of scholarly intent.

5.3. Layout as Editorial Vision: Navigating Functionality and Tradition

The evolving layout of Buddhist glossaries reflects a broader set of considerations beyond their original Buddhist readership. At its core, layout serves as a material expression of editorial intent, shaped by both practical needs and academic values. This intent operates on two levels: first, how editors envisioned the text’s utility for readers, and second, how they perceived its scholarly or religious significance.

For Qing literati engaged in evidential scholarship, Buddhist glossaries were not tools for understanding Buddhist scriptures but rather repositories of lost philological material. As a result, they stripped these texts of their religious framing and reformatted them into more compact layouts, prioritizing efficiency and textual density over the layout conventions of Buddhist print culture. By consolidating fascicles, reducing the page count, and eliminating formatting features designed for scriptural cross-referencing, they transformed these glossaries into scholarly reference works that fit within the broader tradition of classical Chinese philology.

In contrast, editors who prioritized the preservation of antiquarian textual forms resisted these changes, retaining Buddhist layout conventions even when producing scholarly editions. For figures like Chen Zongyi, maintaining the original formatting was not simply a matter of practicality but a conscious effort to uphold textual authenticity and historical continuity. His decision to retain the double-line wrapping format of the Yongle Southern edition illustrates how material aspects of a text could serve as markers of authority, reinforcing a commitment to ancient textual traditions even at the cost of efficiency.

These divergent approaches highlight the multiple forces shaping textual presentation in late imperial China. The transition of Buddhist glossaries from religious to scholarly domains was not merely a shift in readership but also a transformation in how texts were materially constructed, read, and valued. Whether through adaptation for philological research or the preservation of historical forms, layout choices embodied deeper intellectual and cultural priorities, underscoring the dynamic relationship between textual form, function, and scholarly tradition.

6. Conclusions: Layout as a Materialized Editorial Strategy

This study has examined layout as both a reflection of editorial intent and a material factor shaping the transmission, interpretation, and reception of texts. By tracing the evolution of Xuanying’s Yinyi from manuscripts to block-printed editions—including Buddhist and scholarly editions—it has demonstrated that layout is more than a visual feature; it is integral to textual organization, reader engagement, and knowledge management. Across different periods and formats, copyists and compilers adjusted their layout strategies based on the perceived textual function and intended readership, negotiating between material practical constraints and intellectual priorities.

A key finding of this study is that layout embodies how editors conceptualize a text’s purpose. Manuscript traditions reveal a functional spectrum: some glossaries mirrored canonical scriptures, while others prioritized usability and economy. This spectrum persisted in early block-printed editions, where layout was shaped by inherited manuscript formats, institutional practices, and production concerns. The eventual dominance of double-line annotation with line breaks in the Southern Buddhist Canon editions reflects how a layout came to prevail by offering practical flexibility, production efficiency, and alignment with broader conventions of textual organization.

The Qing scholarly editions marked a new phase in the reception and function of Buddhist glossaries. No longer used primarily for monastic scripture study, these texts were repurposed for philological scholarship. This shift in perception was directly reflected in layout changes: many Qing editions adopted formats resembling Confucian glossaries rather than preserving the Buddhist Canon’s standardized layout. These transformations show how layout was not solely governed by technical and economic efficiency but also by changing epistemological commitments—what a text was for, and how it was meant to be read.

At the heart of this evolution lies a broader dynamic of information management. While standardization enhanced readability, retrievability, and production efficiency, it also introduced risks. Textual collation in later editions—such as the Sixi and Qisha editions—often required integrating corrections into pre-existing formats, sometimes producing spatial dislocation or filler character insertion that inadvertently altered the text’s structure. These strategies, aimed at preserving visual regularity, occasionally introduced misalignments, variant readings, or partial reconfigurations of glosses. Such cases remind us that textual variants can arise not only from intentional emendation or scribal lapses but also from the material and logistical constraints of textual production. Traditional philological approaches have often prioritized textual content over layout and formatting, yet the visual structures of a page—and the demands of printing, copying, and compilation—play a decisive role in how texts are interpreted, reshaped, and transmitted across time.

By situating Buddhist glossary layout within a broader ecology of textual production, this study challenges narrowly text-focused models of transmission and proposes an integrated perspective that foregrounds material, institutional, and intellectual factors. The transition from manuscript diversity to early print standardization—and later to Qing-era reinterpretations—reveals that layout decisions often reflected more than aesthetic concerns, embodying shifting epistemological priorities and the competing demands of fidelity, usability, and authority. Understanding layout as a historically contingent and materially embedded editorial strategy allows us to reassess textual variation not as mere error, but as evidence of the evolving life of texts.

In this light, layout emerges not simply as a matter of visual design, but as a method of information organization—one that shaped how knowledge was structured, accessed, and interpreted across different historical and material conditions. As such, tracing the visual lives of Buddhist glossaries—such as Xuanying’s Yinyi and Huiyuan’s Yinyi—offers not only a history of textual transmission but also a window into the broader mechanisms through which knowledge was reproduced and redefined in pre-modern China.