The Genealogy of a Creative Anomaly: Tracing the Conflated Iconography of Mañjuśrī and Samantabhadra from Dunhuang to Late Imperial Folk Prints

Abstract

This article investigates a unique iconographic anomaly in late medieval Dunhuang silk paintings: the conflation of the bodhisattvas Mañjuśrī and Samantabhadra. Focusing on two key artifacts from the 9th and 10th centuries and tracing their legacy to later folk prints, this study argues the phenomenon is not a scribal error but a creative Anomaly—a deliberate ritual synthesis. The analysis reveals this synthesis was driven by two forces: a phonetic re-semanticization in the local dialect and a theological logic born from the integration of Huayan School doctrines with Esoteric ritual practice. The paper demonstrates how Huayan metaphysics were operationalized through condensed Esoteric invocations, turning the inscription into a functional ritual shorthand. Crucially, this study demonstrates the genealogical survival of this Silk Road variant in Ming and Qing dynasty woodblock prints. It uncovers a parallel, non-canonical lineage of visual piety, sustained through workshop copybooks rather than elite textual discourse. This trajectory challenges the linear narrative of Buddhist art history, highlighting the generative power of localized adaptations existing outside the purview of the written canon.

1. Introduction

1.1. Context and Problem Statement: An Anomaly in the Silk Road Visual Grammar

The Silk Roads served not merely as conduits for material exchange but as dynamic arenas where religious visual culture was continuously translated and transformed. As Buddhist iconography traversed from India through Central Asia to Dunhuang, local artisans operated within a complex tension between adhering to rigid canonical models and responding to the fluid ritual demands of local patrons. As Sarah Fraser notes in her study of Dunhuang workshops, the production of Buddhist art was a performative act, where the final image often reflected a negotiation between standard copybooks (粉本 fenben) and the specific liturgical needs of the donor (Fraser 2004, pp. 7–12, 195–200).

Within the standard visual grammar of Mahāyāna Buddhism, the bodhisattvas Mañjuśrī and Samantabhadra are iconographically distinct. They serve as the two primary attendants of Śākyamuni Buddha, representing the complementary virtues of Wisdom (prajñā) and Practice (caryā). Visually, they are unambiguously coded: Mañjuśrī rides a lion, and Samantabhadra rides a white elephant. While they function as a pair, they remain ontologically separate individuals in canonical texts and standard iconography.

However, a votive silk painting from late medieval Dunhuang disrupts this established visual logic. Stein painting 5 Ch.iv.0023 (Figure 1)1, dated to the 5th year of the Xiantong (咸通) era (864 CE) and housed in the British Museum, presents a baffling iconographic paradox.

The painting itself is a work of high artistic quality, executed on silk—a material indicative of significant wealth and devotion. It depicts a sophisticated Esoteric ritual assembly featuring four forms of Avalokiteśvara in the upper register. In the lower register, the visual forms appear orthodox: to the right, a bodhisattva rides a lion; to the left, a bodhisattva rides an elephant. By all visual conventions, these are clearly Mañjuśrī and Samantabhadra. Yet, the accompanying inscription for the elephant-riding figure creates a stark textual–visual dissonance. Instead of identifying the figure as Samantabhadra, the cartouche reads: “The Great Saint Mañjuśrī-Samantabhadra Bodhisattva” (大聖文殊普賢菩薩).

This conflation raises a critical set of research questions. If this were a crude sketch produced by an illiterate peasant, one might dismiss it as a clerical error. However, the donor inscriptions reveal a prominent family deeply embedded in the monastic community. The primary patron is identified as Tang Anjian (唐安諫), holding the high military rank of Yaqian Yuhou (衙前虞候). More significantly, the family lineage explicitly bridges the secular and the sacred: the male donors are led by Tang’s father, Monk Shenwei (父僧神威), and the female donors include two nuns (Bhikṣuṇī Miaoyi 比丘尼妙義 and Nun Fumiao 尼福妙) alongside his mother, Lady Qu (母麴氏), and wife, Shisanniang (阿婦什三娘).

The central puzzle of this study is thus: How could an elite family, with a monastic father and multiple ordained family members presumably guiding the ritual production, accept such a glaring theological “mistake” on a costly silk devotional object? Does this inscription represent a failure of literacy, or does it signal the emergence of a new, localized ritual logic—one where the two bodhisattvas were conceptually merged into a single dyadic entity?

1.2. Literature Review & Theoretical Reframing

Scholarship on Dunhuang Buddhist art and society has long noted the prevalence of non-standard forms in the Guiyijun (歸義軍) period (851–1036 CE). Distinguished historians such as Hao Chunwen (郝春文) (Hao 1998) and Yu Xin (余欣) (Yu 2006) have extensively documented the social life of Dunhuang monks and laypeople, revealing a borderland society where the boundaries between the monastic and secular were often porous. Within this context, anomalies in inscriptions are frequently categorized as biwu (筆誤 clerical errors) or attributed to the “diffusion” of religious knowledge, implying that practitioners in the periphery ostensibly lacked the rigorous literacy of central elites.

If this text-centric approach is strictly followed, the inscription “Mañjuśrī-Samantabhadra (文殊普賢)” on painting Ch.iv.0023 would be dismissed as a mere failure of transmission—a corrupted copy of a canonical model. However, this explanation fails to account for the intentionality of the patron, Tang Anjian (唐安諫), a high-ranking military official whose family included monastic elites. It is improbable that such a prominent donor would unknowingly accept a “defective” ritual object for his family’s private worship.

To resolve this contradiction, it is necessary to shift the analytical framework from textual verification to ritual efficacy. Eugene Wang (汪悅進), in his study of medieval Chinese Buddhist visual culture, argues that visual programs often operate on a “world-making” logic that is distinct from textual authority (E. Y. Wang 2005, pp. xv–xviii). For the devotee, the image is not merely an illustration of a scripture but a medium for creating a ritual reality. Furthermore, Robert Sharf posits that Buddhist icons should be understood as “ritual technologies”—tools designed to function within a specific liturgical context (Sharf 2001, p. 115).

Viewed through this lens, the “Mañjuśrī-Samantabhadra” inscription is not a textual corruption but a functional label. In the specific Esoteric ritual context of Dunhuang, where the two bodhisattvas were invoked together to protect the donor, the visual dyad was conceptually collapsed into a single unit of power. The inscription, therefore, faithfully records the ritual reality of the donor rather than the doctrinal constraints of scripture.

Adopting this theoretical lens, this paper argues that the “Mañjuśrī-Samantabhadra” conflation is not a random error but a creative Anomaly. It represents a deliberate ritual synthesis driven by the specific linguistic environment and theological currents of late medieval Dunhuang.

The study will demonstrate that this synthesis was made possible by a process of phonetic re-semanticization unique to the Northwest dialect, a process revealed through the textual variants found on a later Dunhuang silk painting (Ch.00224). Furthermore, it will uncover the theological logic behind this merger by situating the artifacts within the “Huayan-Esoteric Synthesis” prevalent in the mid-to-late Tang, demonstrating that the conflated title served as a ritual shorthand legitimized by contemporary doctrinal innovations. Finally, by tracing this specific iconographic template from the Silk Road frontier to the printing workshops of late imperial Northern China, this research reconstructs a resilient, non-canonical genealogical lineage of visual piety, challenging the linear narrative of Buddhist art history.

2. The Artifacts: Visual and Textual Analysis

2.1. Visual-Textual Dissonance in the 864 CE Silk Painting (Ch.iv.0023)

A detailed examination of the iconographic program of painting Ch.iv.0023 reveals the specific theological environment in which the anomaly occurred. The composition functions as a sophisticated Esoteric altar, anchored by four distinct manifestations of Avalokiteśvara in the upper register. These figures are explicitly identified by their cartouches as the “Great Compassionate Savior from Suffering (大悲救苦觀世音菩薩),” “Great Saint Savior from Suffering (大聖救苦觀世音菩薩),” “Great Compassionate Eleven-Headed (大悲十一面觀世音菩薩),” and “Great Saint Wish-Fulfilling Wheel (大聖而[如]意輪菩薩)” Avalokiteśvara Bodhisattva. The specific inclusion of the Eleven-Headed (Ekadaśamukha) and Wish-Fulfilling Wheel (Cintāmaṇicakra) forms explicitly grounds the artwork in the Esoteric tradition, intended for rituals of protection and merit accumulation.

It is against this orthodox Esoteric backdrop that the lower register presents a striking deviation. To the viewer’s right, the depiction of Mañjuśrī on a lion is iconographically standard and labeled consistently as “The Great Saint Mañjuśrī Bodhisattva” (大聖文殊師利菩薩). However, the figure to the left creates a stark dissonance. While visually possessing all the standard attributes of Samantabhadra riding a white elephant, the accompanying cartouche reads “The Great Saint Mañjuśrī-Samantabhadra Bodhisattva” (大聖文殊普賢菩薩) (see Figure 1). A close inspection of the attendants reveals the specific iconographic standard of the mid-Tang period. Both the lion and the elephant are led by dark-skinned figures with curly hair, identifiable as Kunlun slaves or Kunlun nu (崑崙奴). These figures were standard in Tang depictions, representing exotic servitude and the universality of the Buddhist dharma reaching remote lands.

In this specific inscription, the name “Mañjuśrī” functions not as a separate identifier but effectively as a prefix to “Samantabhadra,” fusing the two names into a single compound title. This textual hybridity contradicts the distinct visual form. Given the high social status of the patron Tang Anjian and the fine quality of the silk medium previously discussed, this inscription cannot be dismissed as a casual error; rather, it suggests a specific ritual logic where the identity of Samantabhadra is structurally modified by the presence of Mañjuśrī.

2.2. The Persistence of the Anomaly: The 939 CE Silk Painting (Ch.00224)

The hypothesis that the conflated title in the 864 CE painting was merely an isolated scribal error is effectively refuted by a second, later artifact: the votive silk painting Ch.00224 (also cataloged as Stein painting 41), dated to the 4th year of the Tianfu (天福) era (939 CE)2. Produced seventy-five years after the Tang Anjian painting, this work demonstrates that the conceptual merging of the two bodhisattvas had persisted and evolved within the local visual culture of Dunhuang.

In this later work, the central deity is Maitreya in Tuṣita Heaven, flanked by two attendants. While the visual iconography remains standard, the accompanying cartouches present a radical phonetic variation of the bodhisattvas’ names that mirrors the logic of the earlier artifact. The figure to the right (Mañjuśrī) is labeled “Wen-sui-shi-li Bodhisattva” (聞隨師離菩薩), while the figure to the left (Samantabhadra) is labeled “Wen-sui-bu-xian Bodhisattva” (聞隨補賢菩薩) (see Figure 2). Here, the standard Chinese characters for Mañjuśrī (文殊Wen-shu) have been replaced by the homophonous variant Wen-sui (聞隨). Crucially, this variant is not restricted to Mañjuśrī; it is also applied as a prefix to Samantabhadra (whose name is rendered as Bu-xian 補賢, a phonetic variant of Pu-xian 普賢).

This inscription structure confirms a stabilized local pattern. In both the 864 CE and 939 CE silk paintings, the signifier for Mañjuśrī (whether “Great Saint Mañjuśrī” or “Wen-sui”) is superimposed onto the identity of Samantabhadra. The recurrence of this specific naming convention—attaching Mañjuśrī’s name to Samantabhadra—across a span of seven decades indicates that the “Mañjuśrī-Samantabhadra” dyad had become a fixed concept in the visual culture of Dunhuang.

Furthermore, the donor inscription on this painting provides critical insight into the motivation behind such non-standard iconography. The text explicitly states that the offering was made because the donor was “entangled by lingering illness” (宿病纏縈) and sought relief through the power of these deities. This pragmatic urgency reinforces the argument that the conflation was driven by ritual efficacy. In the specific context of healing rituals, the distinctions between the two bodhisattvas were likely collapsed into a single, potent source of salvation, phonetically rendered as “Wen-sui.”

3. Deciphering the Mechanism: Phonetic and Theological Reconstruction

3.1. Phonetic Re-Semanticization: From “Mañjuśrī” to “Wen-sui”

The recurrence of the variant “Wen-sui” (聞隨) in the 939 CE painting provides the crucial linguistic key to unlocking the logic of the conflation. While the standard Chinese transliteration for the Sanskrit Mañjuśrī is Wen-shu (文殊), the variant Wen-sui appears anomalous. However, when analyzed through the lens of historical phonology specific to the Northwest region, this shift reveals a coherent pattern of localization.

It is important to acknowledge that in the orthodox Qieyun (切韻) rhyme dictionary system (Early Middle Chinese) (Baxter 1992, pp. 45–50), the two characters were distinct: Sui (隨) belonged to the Wei rhyme class (微韻), while Shu (殊) belonged to the Yu rhyme class (魚韻). However, phonological studies on Dunhuang manuscripts have identified a specific dialect feature of the Late Middle Chinese period known as “the merging of the Zhi and Wei rhymes into the Yu rhyme” (支微入魚). As demonstrated by the linguist Wang Junhu, who analyzed the phonetic patterns in the Jin-Shaan-Gan dialects and Dunhuang manuscripts, the actual speech habits in the Hexi Corridor during the late Tang and Five Dynasties differed significantly from the dictionary standard (J. Wang 2004).

Specifically, regarding the initials, the Guangyun dictionary categorizes Sui (隨) under the Xie (邪) initial, indicated by the fanqie spelling “Xun-Wei” (旬為切). According to Wang’s reconstruction of the Northwest dialect, this syllable likely shifted towards the pronunciation xu in the local vernacular. Meanwhile, the character Shu (殊) belongs to the Chan (禪) initial class, indicated by the fanqie “Zhi-Shu” (陟輸切). The reconstructed value for this initial is ʑiu, where the symbol ʑ represents the voiced counterpart of the voiceless alveolo-palatal fricative ɕ (equivalent to the Pinyin x). Consequently, in the acoustic environment of the period, the pronunciation of Shu also approximated xu. Thus, Sui and Shu converged towards a near-identical pronunciation, providing the precise linguistic condition for “Mañjuśrī” (Wenshu 文殊) to be transcribed as “Wen-sui 聞隨.”

Yet, this transformation was not merely a passive phonetic error; it was a process of semantic reconstruction. While the character Shu (殊) in the original transliteration is a meaningless phonetic placeholder, the characters Wen (聞) and Sui (隨) carry profound intelligible meanings in Chinese: “to hear” and “to follow.” In a devotional context, these characters evoke the concepts of “Hearing the Dharma” and “Following the Buddha’s Teachings.”

Therefore, for the local practitioner, the foreign Sanskrit name was effectively domesticated. The bodhisattva was no longer just Mañjuśrī, but the deity of Wen-sui—the one who hears and follows. Once this term established itself as a meaningful, localized signifier for the bodhisattva’s function, it became a detachable linguistic unit. This explains why, in the paintings analyzed, “Wen-sui” (or its formal equivalent “Mañjuśrī”) could be applied as a functional prefix to Samantabhadra, creating the hybrid title “Wen-sui-bu-xian (聞隨補賢)” to denote a deity who encompasses the virtues of both wisdom (hearing) and practice (following).

3.2. The Theological Logic: The Non-Duality of Essence and Function

To understand why the specific iconographic synthesis of Mañjuśrī and Samantabhadra took root in the visual culture of Dunhuang, one must first recognize the region’s status as a vibrant stronghold of Huayan devotion during the late Tang and Five Dynasties. The intellectual landscape of the Guiyijun period was not merely receiving texts from the capital; it was actively visualizing Huayan cosmology as a dominant religious paradigm. As demonstrated by Dorothy Wong in her analysis of East Asian Huayan paintings, the visual evidence from Dunhuang—ranging from monumental cave murals to portable liturgies—indicates that “the Huayan doctrine was of central importance to the Buddhist community at Dunhuang” during this period (Wong 2007, pp. 337–84). Wong notes that starting from the Mid-Tang, depictions of the Huayan bian (華嚴變 Transformation Tableaux of the Avataṃsaka) increased steadily in the cave-chapels, often occupying prime positions in the ritual space. The existence of large-scale, high-quality votive paintings on silk (similar in material to the artifact under discussion) further attests that the complex metaphysics of the Avataṃsaka were not restricted to scholastic commentaries but were integrated into the liturgical functions of the local elite.

Against this robust background of visual and doctrinal dominance, the specific anomaly in painting Ch.iv.0023 can be understood as a micro-manifestation of a broader theological paradigm. This paradigm is succinctly characterized by Chen (2024) as “Huayan serving as the Essence, and Esoteric practice serving as the Function” (華嚴為體，密教為用). This formulation rests upon a foundation of prior scholarship that has elucidated the intrinsic continuity between the two traditions. The theoretical groundwork was laid by the historian Yan Yaozhong (嚴耀中), who argued that Chinese Esoteric Buddhism and the Huayan tradition possess a fundamental “consistency in substance” (本體上的一致) (Yan 1999, pp. 82–94). Yan specifically points out that the Avataṃsaka Sutra provided the theoretical basis for Esoteric concepts such as the “Three Mysteries” (三密 sanmi) and “Consecration” (灌頂 abhiṣeka). Most crucially, Yan argues that the Esoteric veneration of Mahāvairocana is fully consistent with the Huayan concept of the Dharmakaya Vairocana, stating that “Esoteric Buddhism is merely taking this Dharmakaya Buddha Vairocana [of Huayan] and further sculpting him into a speaking, formed, and personified idol”3. This observation underscores how Esoteric iconography functions as the concrete manifestation (“Function”) of the abstract Huayan “Essence.”

This synthesis was not merely a later theoretical construction but was historically enacted by the highest authority of the Huayan School. As recorded in the “Biography of the Great Virtuous Monk Fazang of the Great Jianfu Temple in Tang” (唐大薦福寺故寺主翻經大德法藏和尚傳 Tang Da Jianfusi Gu Sizhu Fanjing Dade Fazang Hoshang Zhuan) by the Silla scholar Ch’oe Ch’iwŏn 崔致遠 (857–?), the patriarch Fazang 法藏 (643–712) himself employed Esoteric ritual technology to serve Huayan teleology. The biography records that during the Khitan rebellion in the first year of the Shengong (神功) era (697 CE), Empress Wu commanded Fazang to suppress the enemy. Fazang petitioned: “If you wish to subdue the resentful enemy, I request to employ the methods of the ‘Left Way’ (i.e., Esoteric rites)… The Master then bathed, changed his robes, and established a ritual sanctuary for the Eleven-Faced [Avalokiteśvara]…” The narrative concludes that the enemy troops saw “images of Guanyin floating in the sky” and were defeated, a victory the Empress attributed to the “blessing of Compassionate Power”4. This primary source confirms that the integration of Esoteric rituals (such as the Eleven-Faced Guanyin altar) into the Huayan repertoire was legitimized by the school’s founder himself.

Translating this doctrinal and historical synthesis into specific terms of art history, the scholar Lai Pengju (賴鵬舉) has provided crucial iconographic evidence from the Dunhuang caves. In his analysis, Lai demonstrates that the specific Esoteric forms popular in the region—such as the Thousand-Armed Avalokiteśvara—were not isolated imports but were “derived directly from the Huayan context” of the Potalaka Palace described in the Gaṇḍavyūha. Lai argues that Esoteric icons function as the practical application (Yong) of the Huayan cosmological framework, establishing a solid iconographic mechanism where distinct Esoteric deities are visually treated as emanations of the Huayan Dharmadhātu (Lai 2009, pp. 188–93).

This theological synthesis finds its most articulate expression in the biography of Chengguan 澄觀 (738–839), the Fourth Patriarch of the Huayan School, whose lifetime overlaps directly with the production of the Dunhuang artifacts in question. According to the Biographies of Eminent Monks of the Song Dynasty (宋高僧傳 Song Gaoseng Zhuan), Chengguan, while contemplating the composition of a commentary on the Avataṃsaka Sutra during the Dali (大曆) era (766–779), formulated a critical doctrinal insight regarding the non-dual relationship between the two bodhisattvas and the Supreme Buddha. The text records his reflection: “Moreover, Mañjuśrī governs Wisdom (智 Zhi), and Samantabhadra governs Principle (理 Li). The two sages combined constitute Vairocana. When the myriad practices are all-encompassing, this is the meaning of the Great Huayan”5. This statement provides the decisive theological evidence for interpreting the anomaly in painting Ch.iv.0023. In the intellectual framework of the time, Mañjuśrī and Samantabhadra were not merely independent attendants flanking a central Buddha; they were understood as ontologically inseparable components that, when fused, constituted the totality of the Dharma Realm (Vairocana). Therefore, to invoke the “Two Sages” together was, in essence, to invoke the supreme power of Vairocana himself.

This theological synthesis finds its definitive expression in the Contemplations of the Perfect Interfusion of the Three Holy Ones (San-sheng yuanrong guanmen 三聖圓融觀門), a treatise attributed to Chengguan. As Gimello (1996, p. 358) has elucidated in his translation and study of this work, Chengguan explicitly argues for the ontological non-duality of the two bodhisattvas. Gimello notes that in Chengguan’s soteriology, Mañjuśrī (representing Faith and Wisdom) and Samantabhadra (representing Principle and Practice) function together as the “two-fold causal dimension” (er-yin 二因) that leads to the “fruition” of Vairocana. Crucially, Chengguan formulates a logic of mutual implication that directly illuminates the iconographic anomaly. He states: “Subsuming both insight and principle, there is only the dharmadhātu of the one mind. Thus, to adduce just one of the holy ones is to imply them both (攝智屬理，唯一心法界，故舉一全收)”. This doctrinal assertion—that invoking one sage automatically necessitates the presence of the other—provides the precise theological license for the donor’s inscription.

However, attributing such sophisticated exegesis to a local votive painting raises a historical paradox regarding how abstract Huayan philosophy, typically reserved for the scholastic elite, could shape the visual practice of a borderland donor. The resolution may be found in the role of monastic intermediaries who operated at the interface of elite doctrine and local practice. The donor’s close familial connection to the sangha—specifically through his father, Monk Shenwei—provides a plausible social channel through which such theological concepts could permeate the lay family’s ritual life.

In the transmission from the text of the capital to the ritual of the frontier, complex metaphysical doctrines were often compressed into performative shorthands. For the local practitioner, the compound title Mañjuśrī-Samantabhadra functioned not as a scholarly citation but as an operationalized ritual formula. The monastic elite understood the theological justification of mutual implication, while the artisans and lay donors adopted the resulting hybrid title as a potent and unified invocation for healing and protection. Thus, the anomaly represents a vernacularization of elite theology wherein the doctrine was preserved not in textual treatises but in the functional logic of the ritual image.

When applied to the visual analysis of painting Ch.iv.0023, this doctrine explains the apparent dissonance between the image and the text. The inscription “The Great Saint Mañjuśrī-Samantabhadra” is not a scribal error but a literal textual transcription of Chengguan’s formula: “The two sages combined.” By visually fusing their names into a single compound title, the donor was effectively constructing a ritual invocation of the “complete” Dharma Body through its constituent powers of Wisdom and Principle. The visual iconography retains the elephant-riding figure to identify the aspect of “Practice” (Samantabhadra), which is the visible “Function” required for the ritual; however, the inscription superimposes the name of Mañjuśrī (“Wisdom”) to ensure that this Practice is grounded in the “Essence” of the Dharma. Thus, the artifact embodies the period’s theological logic: the image serves the Esoteric function of ritual presence, while the text asserts the Huayan essence of non-dual unity.

This theological unity was further concretized by the specific structure of Esoteric rituals circulating in Dunhuang. A close examination of the Ritual Procedure for Offering to Yamarāja (焰羅王供行法次第 Yanluo Wang Gongxing Fa Cidi), composed by Amoghavajra (阿謨伽三藏，即不空), reveals the precise liturgical status of these figures. The text lists a strict sequence of invocations (“Single-mindedly invite…”). Crucially, immediately following the invocation of the Buddhas (including “Maitreya Buddha of the Future”), the text invokes the Bodhisattvas in a specific hierarchy:

Single-mindedly invite the Great Saint Mañjuśrī Bodhisattva Mahāsattva.

Single-mindedly invite the Great Saint Samantabhadra Bodhisattva Mahāsattva.6

In this liturgical sequence, Mañjuśrī and Samantabhadra are isolated as the primordial pair, listed before even Avalokiteśvara or Mahāsthāmaprāpta. They function as the standard binary unit representing the head of the Bodhisattva assembly. Furthermore, the text’s sequence—positioning Maitreya immediately before this pair—offers a direct textual correlate to the iconography of painting Ch.00224, where Maitreya serves as the central deity flanked by the “Wen-sui” (Mañjuśrī) and “Wen-sui-bu-xian” (Samantabhadra) figures.

Consequently, the inscription “Mañjuśrī-Samantabhadra” on the paintings should be understood not as a confusion, but as a liturgical ligature. It reflects the ritual reality where these two figures are invariably invoked in tandem as a functional set. For the donor, engaging in a rite for “curing illness” or “protection,” the efficacy depended on summoning this specific “Wisdom-Practice Unit.” The compound title on the painting serves as a visual seal of this ritual coupling, affirming that the two bodhisattvas are present not as separate narratives, but as a unified agency of salvation.

3.3. The Theological Logic II: Resolving the Iconographic Asymmetry

While the macro-theory of “Huayan-Esoteric Synthesis” successfully legitimatizes the linguistic conflation of the names, it leaves unaddressed a deeper, structural paradox regarding the specific visual economy of painting Ch.iv.0023. A critical observation of the composition reveals a startling redundancy: If the figure on the left—the elephant-rider labeled “The Great Saint Mañjuśrī-Samantabhadra”—already encapsulates the ritual unity of the two sages via its inscription, does this render the separate, orthodox Mañjuśrī figure on the right (the lion-rider) theologically superfluous? Superficially, the painting presents an unbalanced equation: a “Pure Mañjuśrī” stands alone on the right, juxtaposed against a “Mañjuśrī-infused Samantabhadra” on the left. If the donor’s intent was simply to worship the unified dyad, the inclusion of an independent Mañjuśrī appears to be a chaotic duplication. However, this study proposes that this apparent redundancy may be understood as a precise visual articulation of the ontological priority of Wisdom (智 Zhi)—a specific soteriological position holding that Wisdom must exist as the distinct, generative source before it can be integrated into Practice.

To decipher this specific visual grammar, one must look beyond general Huayan doctrines to the specific exegetical perspective offered by Li Tongxuan 李通玄 (635–730) in his New Treatise on the Avataṃsaka Sutra (新華嚴經論 Xin Huayan Jing Lun). While previous scholarship often regarded Li as a marginal figure whose influence was confined to specific localities, recent bibliographic research has fundamentally reshaped our understanding of his reception in the late Tang. Chi (2022) has convincingly demonstrated that the widespread circulation of Li’s thought was catalyzed precisely during the Dazhong 大中 era (847–859), a period marked by the compilation of the Combined Treatise of the Avataṃsaka Sutra (華嚴經合論 Huayan Jing Helun), which fused Li’s commentary directly with the scripture.

This chronological finding is significant for our analysis. The Tang Anjian painting is dated to the 5th year of Xiantong (864 CE)—merely five years after the close of the Dazhong era. This tight chronological proximity renders it plausible to interpret the painting as a visual engagement with the theological currents exemplified by the newly circulated New Treatise. Furthermore, Li (2018) has highlighted that during this period, the veneration of Li Tongxuan was particularly intense in Northern China (such as Bingzhou 幷州/Taiyuan 太原), forming a distinct Northern Buddhist sphere that culturally encompassed the Hexi Corridor. Thus, it is historically reasonable to suggest that Li Tongxuan’s theology represents the intellectual zeitgeist accessible to a donor like Tang Anjian. It is within Li’s specific dialectic—which prioritizes the “Fundamental Wisdom (根本智慧)” of Mañjuśrī as the ontological prerequisite for all Practice—that we find a coherent explanatory framework for the painting’s complex iconography.

Using this exegetical framework, the specific arrangement of the figures—an independent Mañjuśrī on the right versus a conflated Samantabhadra on the left—can be decoded as a dynamic diagram representing the causal and temporal hierarchy between Wisdom (智 Zhi) and Practice (行 Xing). In Scroll 3 of the New Treatise, Li Tongxuan establishes a precise ontological relationship between the two sages using the metaphor of family lineage, but he imbues it with a counter-intuitive soteriological twist. He identifies Mañjuśrī as the “Small Boy” (小男 Xiaonan) and Samantabhadra as the “Eldest Son” (長子 Zhangzi). In a secular biological context, the eldest son precedes the younger; however, in Li’s mystical chronology, this order is reversed. The “Small Boy” represents the ontological “First” because he symbolizes the very inception of enlightenment.

To explicate this reversal, Li provides a definitive statement linking the “Small Boy” to the primordial moment of realization:

Mañjuśrī is called the ‘Small Boy’, because he is the head of [Vai]rocana’s initial generation of the mind (Skt. bodhicittotpāda) to realize the Buddha-nature of the Fundamental Wisdom of the Dharma Body; he is the initial realization of the Buddha-nature of Fundamental Wisdom, and is the ‘Family of Wisdom’ of the Holy Nature from which all Buddhas are first born.7

This passage defines Mañjuśrī not merely as a participant in the path, but as the “Head” (首 Shou) and the “Family/Home” (家 Jia) of the Fundamental Wisdom (本智 Benzhi). He is the “Small Boy” precisely because he represents the “newborn” moment of the initial resolve—a moment that ontologically precedes the “growth” of practice.

Li further grounds this complex reversal in the cosmological symbolism of the Yijing (易經), specifically through the spatial and temporal attributes of the Eight Trigrams. He explicitly associates Mañjuśrī with the Gen Trigram (艮卦), which corresponds to the Northeast direction. In traditional Chinese cosmology, the Northeast is the critical liminal space known as Pingdan (平旦), the break of dawn where darkness ends and light begins. Li writes: “Mañjuśrī corresponds to the Northeast… to the Gen Trigram, which is the Small Boy… It represents the break of dawn, where the appearance of darkness is gone but the full sunlight has not yet manifested; it images the head of enlightenment”8. He further specifies the attributes of this trigram: “It is Mountain (山 Shan), it is Stone (石 Shi).”9 This imagery is theologically precise. In the Yijing tradition, the quality of the Gen (艮 Mountain) trigram is defined as “Stillness” or “Stopping” (止 Zhi). Li Tongxuan appropriates this symbol of the “immovable Mountain and Stone” to characterize the nature of Fundamental Wisdom (本智 Benzhi): it is the stable, immovable source of enlightenment, existing a priori before the dynamic “movement” of practice begins. Therefore, Mañjuśrī, as the “Small Boy” associated with the stillness of the Mountain, occupies the station of the “Static Origin”.

In strict opposition to this static source, Li associates Samantabhadra with the Zhen Trigram (震卦), which corresponds to the “Eldest Son” (長男 Zhangnan). Li explicitly locates Samantabhadra in the “East” (東方 Dongfang), the direction of sunrise and renewal. He writes: “Samantabhadra is the Eldest Son… located in the East… representing the Head, the Leader, the Green Dragon (青龍 Qinglong), Celebration, and the Growth of Spring (春生 Chunsheng); he is the beginning of the establishment of the Dharma.”10

This cluster of imagery—East, Green Dragon, Spring Growth—is theologically significant. In the Yijing, the Zhen trigram symbolizes “Thunder” (雷 Lei) and possesses the attribute of “Motion” (動 Dong). By invoking the “Growth of Spring,” Li Tongxuan characterizes Samantabhadra not as a static principle, but as the Dynamic Application (用 Yong) that expands and proliferates. Thus, while Mañjuśrī (the Small Boy/Mountain) is the Immovable Origin (體 Ti), Samantabhadra (the Eldest Son/Spring) is the Active Function (Yong) that carries out the “establishment of the Dharma.”

This textual definition dictates the visual logic of the figure on the left. The figure riding the elephant represents this “Eldest Son” in his aspect of “Spring Growth” or active practice. However, as Li warns, this action must not be separated from its source.

In contrast to the static source represented by Mañjuśrī on the right, the figure on the left—the Samantabhadra on the elephant—corresponds to the “Eldest Son” (長子 Zhangzi). In the soteriological structure defined by Li Tongxuan, this figure represents the maturation and expansion of the Dharma. However, Li issues a critical theological warning regarding the danger of viewing this figure in isolation. In the same third scroll of the New Treatise, he argues that the separation of the two sages leads to a fundamental spiritual deficiency. He writes:

If there is Mañjuśrī without Samantabhadra, it signifies that one does not yet possess Practice; if there is Samantabhadra without Mañjuśrī, it clarifies that one does not see the Fundamental Wisdom of the Dharma Body.11

This dialectical statement provides the definitive logic for the iconographic anomaly. It implies that a depiction of Samantabhadra that is conceptually separated from Mañjuśrī represents a specific soteriological failure: the failure to “see the Fundamental Wisdom” (不見法身本智). For a devotee engaging in ritual practice, an image of Samantabhadra that does not explicitly invoke Mañjuśrī would risk being interpreted as practice disconnected from its ontological source. Therefore, necessitated by this theological logic, the iconographic program structurally intervenes in the depiction of the left-hand figure. By superimposing the name “Mañjuśrī” onto the Samantabhadra figure, the iconographic program creates a theological guarantee. It affirms that this specific Practice is not isolated; it is Practice that possesses the vision of Fundamental Wisdom.

Furthermore, Li Tongxuan explicitly defines the relationship between the Buddha and the two bodhisattvas as one of structural interpenetration. He states: “The three persons form a mutual body to constitute the Essence and Function of the One Dharma Realm.”12 The painting Ch.iv.0023 serves as a visual realization of this “Mutual Body” (互體 Huti) doctrine. It maps the linear progression of the believer’s path:

First, the viewer encounters the Source on the right (the independent Mañjuśrī), corresponding to the “Initial Arising” of Wisdom.

Second, the viewer encounters the Application on the left (the conflated Samantabhadra), corresponding to the “Fulfillment” of Practice.

The inscription “The Great Saint Mañjuśrī-Samantabhadra” is the final seal of this integration. It confirms that the endpoint of the path is not merely Action, but Action that has fully integrated the Wisdom of the beginning. Far from being a chaotic error, the specific arrangement of Ch.iv.0023—an independent Mañjuśrī alongside a conflated Samantabhadra—offers a highly nuanced diagram of Tang Huayan thought. It illustrates Li Tongxuan’s maxim that while the Cause (Mañjuśrī) must exist as the antecedent, the Effect (Samantabhadra) must inevitably encompass the Cause to form the One True Dharma Realm (一真法界).

4. The Genealogical Legacy: Survival in Late Imperial Folk Prints

4.1. The Material Evidence: Three Late Imperial Woodblock Prints

Any notion that the “Mañjuśrī-Samantabhadra” synthesis was merely a localized idiosyncrasy or a momentary scribal error of medieval Dunhuang is effectively challenged by the discovery of three related woodblock prints dating from the late imperial period (Ming to Qing dynasties). While the provenance of artifacts sourced from the art market is inherently uncertain, their collective evidence reveals a compelling pattern that strongly supports their historical value. Several factors corroborate their significance. First, the existence of multiple, physically distinct copies—recorded in Beijing, Shanghai, and Korean auctions—points toward standardized mass production rather than a singular modern forgery. Second, these prints consistently reproduce a highly specific and non-canonical textual anomaly, the conflated “Mañjuśrī-Samantabhādra” title, which is unlikely to be an arbitrary invention by a forger unfamiliar with its deep theological origins. Therefore, their evidentiary value lies not in a verifiable provenance but in their typological consistency, which confirms that this unique iconographic template survived and circulated widely in late imperial Northern China.

The most complete and significant specimen is a set of two woodblock prints auctioned by Beijing Taihe Jiacheng (泰和嘉成) Auction Co., Ltd. during their 2014 Spring Art Auction (Lot 863, see Figure 3). The work is a diptych printed on paper, measuring 74 cm in height and 36 cm in width, likely originally mounted as hanging scrolls for a domestic altar or a village temple. The composition strictly adheres to the standard dyadic format established in the Tang dynasty: the right sheet depicts Mañjuśrī riding a lion, while the left sheet depicts Samantabhadra riding an elephant. The artistic style, characterized by the elaborate rendering of the beasts’ manes and trappings, reflects the aesthetic lineage of the Shanxi temple wall paintings. However, a significant iconographic evolution is visible in the attendants. Unlike the Tang painting where the beasts are led by Kunlun slaves, the figure leading Mañjuśrī’s lion in the woodblock print is a Central Asian figure dressed in a long robe and a brimmed hat. This identifies him as the King of Khotan (于闐王), a standard feature of the New Style Mañjuśrī that became dominant after the tenth century.

This visual shift is analytically significant. It demonstrates that while the pictorial details were updated to match the stylistic fashions of the Song and Yuan dynasties, the archaic and anomalous inscription was left untouched. The persistence of the conflated title Mañjuśrī-Samantabhadra, even as the attendant iconography evolved from the Kunlun nu to the Khotanese king motif, confirms that the text was treated as an immutable ritual formula, resistant to the artistic changes that affected the rest of the composition.

Crucially, the inscriptions reproduce the exact theological logic found in the 9th-century Dunhuang votive painting (Ch.iv.0023). The cartouche on the right sheet follows the orthodox convention, reading “The Great Saint Mañjuśrī Bodhisattva” (大聖文殊師利菩薩). However, the cartouche on the left sheet, located directly above the figure riding the elephant, explicitly reads “The Great Saint Mañjuśrī-Samantabhadra Bodhisattva” (大聖文殊普賢菩薩)13. This textual replication is structurally identical to the Dunhuang prototype separated by nearly a millennium. It confirms that the formula of superimposing Mañjuśrī’s name onto Samantabhadra was not a momentary scribal error, but had solidified into a fixed convention within this specific lineage of religious art.

Furthermore, this print bears two critical workshop marks that allow for geographic and stylistic localization. The upper right corner of the Mañjuśrī sheet is inscribed “Painted by Zhu Haogu of Qinchang” (禽昌朱好古畫). Zhu Haogu was a renowned master of temple wall paintings active in Southern Shanxi during the Yuan dynasty (1271–1368), known for his influential style that dominated North Chinese religious art for centuries. The invocation of his name on a later woodblock print suggests that this iconographic composition was perceived as part of the prestigious “Zhu Haogu style” lineage. Simultaneously, the upper left corner of the Samantabhadra sheet bears the colophon “Carved and Printed by the Zhao Family of Xiao Xingzhou” (小興州趙家雕印). “Xiao Xingzhou” refers to a strategic location in the borderlands of modern Hebei and Shanxi, a hub for immigrants and folk religious propagation during the Ming–Qing transition. The juxtaposition of a Yuan dynasty master’s name with a Ming–Qing folk workshop mark indicates a deep historical stratification, suggesting that this “anomalous” iconography was transmitted through generations of artisans as a canonical form.

The stability of this iconographic template is corroborated by a second artifact recorded by Artron Art Network (Lot 0223, see Figure 4), auctioned in 200614. The lot is explicitly titled “Image of Mañjuśrī and Samantabhadra” (Wenshu Puxian Xiang), indicating that although the available digital image displays only a fragment focusing on Mañjuśrī, the complete artifact—like the Taihe Jiacheng specimen—was understood as a dyadic set comprising both bodhisattvas. Physically, the item consists of two sheets of hemp paper labeled as “Qing Dynasty Woodblock Prints.” Crucially, the upper right corner of the visible fragment preserves two legible characters: “Gu Hua” (古画). This is unmistakably the truncated remainder of the full signature “Qinchang Zhu Haogu Hua” (禽昌朱好古画) found on the Taihe Jiacheng specimen. This textual correspondence is reinforced by the visual identity of the primary figure: the linear execution of the Bodhisattva’s robes, the specific curvature of the drapery folds, and the distinctive posture of holding the ruyi scepter are strictly consistent with the Taihe Jiacheng print, confirming that despite being a distinct physical copy, it belongs to the exact same visual lineage.

Significantly, the geographical reach of this circulation is evidenced by a third specimen identified in the collection of Han Auction in Korea (Lot 50-145, see Figure 5)15. The lot is explicitly titled “Woodblock Print of Mañjuśrī and Samantabhadra”, confirming that it preserves the complete dyadic set. A comparative visual analysis of the Samantabhadra image reveals that the line quality, the structural rendering of the elephant, and the calligraphic style of the inscription “The Great Saint Mañjuśrī-Samantabhadra Bodhisattva” are identical to those of the Taihe Jiacheng print. Most importantly, it preserves the exact same workshop markers, including the colophon “Carved and Printed by the Zhao Family of Xiao Xingzhou” (小興州趙家雕印), which links it directly to the “Zhu Haogu” lineage [26]. However, a close examination of the paper edges and the wear patterns indicates that this is a distinct physical copy, not the same sheet. The existence of multiple distinct copies produced from what appears to be the same woodblock—or a closely related set of re-carved blocks—provides decisive evidence of standardized mass production. Unlike the singular hand-painted silk paintings of Dunhuang, these woodblock prints confirm that by the Ming–Qing period, the “Mañjuśrī-Samantabhadra” hybrid had transitioned from a manuscript tradition into a mechanically reproduced religious commodity.

4.2. The Geographic and Stylistic Lineage: From Southern Shanxi to the North China Plain

To understand how a theological anomaly originating in the frontier rituals of medieval Dunhuang could resurface and become standardized in late imperial woodblock prints, it is necessary to decode the specific geographical and historical markers embedded in the artifacts. The inscription on the upper right corner of the Mañjuśrī sheet, “Painted by Zhu Haogu of Qinchang” (禽昌朱好古畫), serves as the primary key for reconstructing the artistic lineage of these prints. A philological analysis of the toponym “Qinchang” reveals that it is a deliberate archaism, carrying significant cultural weight. As detailed in the General Gazetteer of Shanxi (山西通志 Shanxi Tongzhi), Qinchang was the name of a Commandery established during the Northern Wei dynasty (428 CE), which corresponds to the modern Xiangling (襄陵) County in Linfen (臨汾), Southern Shanxi. By the Yuan dynasty, when the historical painter Zhu Haogu was active, this place had long been known as Xiangling.

Why would a printmaker use an obsolete place name? Meng (2011) provides a penetrating explanation in her study of the Southern Shanxi mural workshops. She points out that the use of “Qinchang” by Zhu Haogu and his school was a specific strategy of “revering antiquity and valuing culture” (崇古尚文). By invoking a Northern Wei toponym, the artisan class sought to elevate their status from mere craftsmen to bearers of a classical tradition. For the woodblock print in question, the inclusion of this archaic signature performs a similar function: it claims a prestigious, orthodox heritage rooted in Southern Shanxi—a region historically recognized as a sanctuary where Northern Buddhist visual culture, including Tang and Song iconographies, was preserved in temple murals long after they had faded elsewhere.

Furthermore, the invocation of “Zhu Haogu” must be understood not as a claim of direct authorship, but as a marker of lineage branding. Meng establishes that Zhu Haogu was not merely an individual artist but the charismatic leader of a prolific “Painter’s Workshop” (畫工班子), whose style dominated the production of religious art in North China during the 14th century. His disciples, such as Zhang Boyuan (張伯淵), explicitly identified themselves as “Men of Zhu Haogu” (朱好古門人) in mural inscriptions at the Yongle Palace (永樂宮) (Meng 2011). The woodblock print, by retaining the signature “Painted by Zhu Haogu,” suggests that the template for this image was derived from a fenben (粉本preparatory sketch or copybook) originating from this prestigious workshop tradition. This implies that the specific iconographic anomaly—the conflated title Mañjuśrī-Samantabhadra—did not vanish after the Tang but was fossilized within the Yuan dynasty mural copybooks of the Zhu Haogu school. These workshop manuals served as the silent carrier, transmitting the hybrid deity across the centuries from the medieval frontier to the late imperial metropolis. This connection explains the stylistic fidelity of the print—its meticulous linear definition and iconographic precision—and, crucially, why the “Mañjuśrī-Samantabhadra” inscription was preserved. In the conservative transmission of workshop copybooks, the master’s design possesses canonical authority; errors or anomalies in the original template are often faithfully reproduced by later generations as integral parts of the sacred image.

The second marker, found on the Samantabhadra sheet, reads “Carved and Printed by the Zhao Family of Xiao Xingzhou” (小興州趙家雕印). This colophon shifts the analysis from the artistic source to the mechanics of distribution and migration. As established by Wei (2000), “Xiao Xingzhou” is not a standard administrative unit found in official dynastic geographies but a vernacular toponym deeply embedded in the collective memory of migration. Historical records and genealogies indicate that during the massive state-sponsored migrations of the early Ming dynasty, populations from Shanxi were relocated to populate the war-devastated North China Plain (modern Hebei and Henan). Many immigrants claimed to originate from “Xiao Xingzhou” (often associated with the famous Hongdong Locust Tree 洪洞大槐樹), making it a symbolic node in the diasporic network of Northern Chinese folk religion.

The intersection of these two inscriptions—“Zhu Haogu of Qinchang” and “Zhao Family of Xiao Xingzhou”—maps the precise trajectory of the image’s survival. It indicates that the Zhu Haogu style template, carrying the Tang-era “Mañjuśrī-Samantabhadra” anomaly, was physically transported from Southern Shanxi to the North China Plain by immigrant artisan families like the Zhaos (趙家). Operating within this immigrant network, the workshop produced religious prints that served as links to their ancestral home. For these communities, the image was valuable precisely because it preserved the old forms of the Shanxi temples. Thus, the theological synthesis that began in the rituals of the Silk Road was frozen into a visual template in the workshops of Shanxi, carried across North China by Ming immigrants, and mass-produced in the Qing, surviving as a living fossil of medieval piety.

5. Conclusions

The investigation of the votive silk painting Ch.iv.0023 and its associated artifacts reveals that the inscription “The Great Saint Mañjuśrī-Samantabhadra” is not a peripheral scribal error, but fossilized evidence of a sophisticated process of religious localization along the Silk Road. By deconstructing this specific iconographic anomaly, this study has reconstructed a complex “ritual logic” that operated at the intersection of linguistic adaptation, thological synthesis, and liturgical pragmatism in late medieval Dunhuang.

The mechanism behind this conflation was driven by two converging forces. First, the phonological environment of the Northwest dialect provided the condition of possibility. As evidenced by the variant inscription in painting Ch.00224, the standard transliteration Wenshu (文殊 Mañjuśrī) shifted acoustically to Wensui (聞隨), facilitating a semantic reconstruction where the deity was understood through the intelligible Chinese concepts of “Hearing” (聞 Wen) and “Following” (隨 Sui).

Second, and more fundamentally, the theological climate of the mid-to-late Tang provided the condition of legitimacy. This period witnessed a paradigm shift characterized by the structural integration of “Huayan as the Essence, and Esoteric practice as the Function.” Grounded in the scholarship of Yan Yaozhong and Lai Pengju, this paradigm posits that the abstract ontology of the “One Dharma Realm (一真法界)” was operationalized through Esoteric rituals. The biography of Chengguan (澄觀) confirms that the two sages (二聖) were understood as the constitutive powers of Vairocana, while the ritual manuals of Amoghavajra fixed their status as a binary invocation unit. The resulting compound title on the painting serves as a ritual shorthand, efficiently summoning the total protective power of the Wisdom–Practice complex.

Furthermore, the semantic asymmetry of the composition—retaining an independent Mañjuśrī alongside the conflated Samantabhadra—reflects a precise soteriological grammar derived from the exegetical tradition of Li Tongxuan. By depicting Mañjuśrī as the “Small Boy” who represents the “Initial Arising of Wisdom,” and Samantabhadra as the “Eldest Son” who embodies Practice, the iconographic program adheres to the strict Huayan admonition that Practice must never be separated from its foundational Wisdom. The superimposition of Mañjuśrī’s name onto the Samantabhadra figure acts as a theological seal, affirming that the practice depicted is not blind but is infused with the vision of the Dharma Body.

Crucially, the significance of this study extends beyond the spatial confines of Dunhuang. The discovery of the “Zhu Haogu” and “Xiao Xingzhou” woodblock prints confirms that this “creative anomaly” possesses a resilient genealogical legacy. However, it is essential to define the nature of this legacy with precision. A comprehensive survey of scholastic commentaries from the Song to the Qing dynasties reveals no mention of this specific composite deity; the “Mañjuśrī-Samantabhadra” hybrid never entered the orthodox textual discourse. This very silence in the textual record makes the material evidence—the paintings and prints—all the more significant. It demonstrates that the genealogical legacy of this image was sustained not through a process of ritual sedimentation. What began as a sophisticated theological synthesis in the Tang dynasty was gradually flattened into a visual convention, transmitted through the conservative hands of workshop artisans via copybooks (粉本 fenben) and the pragmatic needs of folk communities, such as the Ming–Qing immigrants from Shanxi. This trajectory reveals the existence of a parallel, non-canonical lineage of visual piety. It challenges the linear narrative of Buddhist art history, demonstrating that Silk Road Buddhist art was not only a transitional phase of importation, but also a generative locus where indigenous forms of visual theology were created, standardized, and transmitted through centuries, often existing outside the purview of the written canon. The “Mañjuśrī-Samantabhadra” dyad stands as a material testament to this enduring process of intercultural evolution and local agency.