The Creative South: Buddhist and Hindu Art in Mediaeval Maritime Asia, volume 1 9789814951494

This edited volume programmatically reconsiders the creative contribution of the littoral and insular regions of Maritim

540 103 33MB

English Pages 375 [377] Year 2022

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

The Creative South: Buddhist and Hindu Art in Mediaeval Maritime Asia, volume 1
 9789814951494

Citation preview

THE CREATM SOUTH

The ISEAS – Yusof Ishak Institute (formerly Institute of Southeast Asian Studies) is an autonomous organization established in 1968. It is a regional centre dedicated to the study of socio-political, security, and economic trends and developments in Southeast Asia and its wider geostrategic and economic environment. The Institute’s research programmes are grouped under Regional Economic Studies (RES), Regional Strategic and Political Studies (RSPS), and Regional Social and Cultural Studies (RSCS). The Institute is also home to the ASEAN Studies Centre (ASC), the Singapore APEC Study Centre, and the Temasek History Research Centre (THRC). ISEAS Publishing, an established academic press, has issued more than 2,000 books and journals. It is the largest scholarly publisher of research about Southeast Asia from within the region. ISEAS Publishing works with many other academic and trade publishers and distributors to disseminate important research and analyses from and about Southeast Asia to the rest of the world.

00_CreSouth V1_FM_8P_9Mar22.indd 2

9/3/22 3:10 PM

First published in Singapore in 2022 by ISEAS Publishing ISEAS – Yusof Ishak Institute 30 Heng Mui Keng Terrace Singapore 119614 E-mail: [email protected] Website: http://bookshop.iseas.edu.sg All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the ISEAS – Yusof Ishak Institute. © 2022 ISEAS – Yusof Ishak Institute The responsibility for facts and opinions in this publication rests exclusively with the authors and their interpretations do not necessarily reflect the views or the policy of the publishers or their supporters. ISEAS Library Cataloguing-in-Publication Data Name(s): Acri, Andrea, 1981-, editor. | Sharrock, Peter D., editor. Title: The creative south : Buddhist and Hindu art in mediaeval maritime Asia / edited by Andrea Acri and Peter Sharrock. Description: Singapore : ISEAS – Yusof Ishak Institute, 2022. | Includes bibliographical references and index. Identifiers: ISBN 978-981-4951-48-7 (soft cover ; Volume 1) | ISBN 978-981-4951-49-4 (pdf ; Volume 1) | ISBN 978-981-4951-51-7 (soft cover ; Volume 2) | ISBN 978-981-4951-52-4 (pdf ; Volume 2) Subjects: LCSH: Art, Medieval—Asia—History. | Buddhist art—Asia—History. | Hindu art—Asia— History. Classification: LCC N8191 A8C91 Cover design by Lee Meng Hui Typesetting by International Typesetters Pte Ltd Printed in Singapore by Markono Print Media Pte Ltd

00_CreSouth V1_FM_8P_9Mar22.indd 4

9/3/22 3:10 PM

Reproduced from The Creative South: Buddhist and Hindu Art in Mediaeval Maritime Asia, edited by Andrea Acri and Peter Sharrock (Singapore: ISEAS – Yusof Ishak Institute, 2022). This version was obtained electronically direct from the publisher on condition that copyright is not infringed. No part of this publication may be reproduced without the prior permission of ISEAS Publishing. Individual chapters are available at .

ContENTs

1. Introduction: Volume 1: Intra-Asian Transfers and Mainland Southeast Asia Andrea Acri and Peter Sharrock

1

PART I: INFLUENCES FROM THE SOUTH 2. From Melayu to Thamel and Back: The Transmigration of the Eight-Armed Amoghapāśa Iain Sinclair 3. In the Footsteps of Amoghavajra (705–774): Southern Indian Artistic Mode in Tang China and its Transmission to Tibet Yury Khokhlov 4. Heruka-Maṇḍalas across Maritime Asia Peter Sharrock

9

66

126

PART II: TRANSFERS AND INNOVATIONS IN MAINLAND SOUTHEAST ASIA 5. Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor Jinah Kim

167

6. Dancers, Musicians, Ascetics, and Priests: Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor Swati Chemburkar

192

7. Libraries or Fire Shrines? Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples from the Prism of Early Śaivism Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

222

00_CreSouth V1_FM_8P_9Mar22.indd 5

9/3/22 3:10 PM

6

Contents

8. Śaiva Religious Iconography: Dancing Śiva in Multi-Polity Medieval Campā Mai Bùi Diệu Linh

289

9. The Colossal Trà Kiệu Pedestal in Campā and its Relationship to Courtly Culture in Cambodia, East Java, and China Mya Chau

305

10. On the Chronological Interrelationship between Newly Found Inscriptions and the Temple Architecture of Campā: The Hòa Lai and Po Dam Sites Trần Kỳ Phương

334

The Contributors

359

Index

362

00_CreSouth V1_FM_8P_9Mar22.indd 6

9/3/22 3:10 PM

Chapter 1

Introduction Volume 1: Intra-Asian Transfers and Mainland Southeast Asia A n dr e a Acr i a n d Peter Sh a r rock

T

his edited volume is the first of two forming an anthology that programmatically reconsiders the creative contribution of the littoral and insular regions of Maritime Asia—Southeast Asia in the first place—to shaping new paradigms in the Buddhist and Hindu art and architecture of the mediaeval Asian world.1 Far from being a mere southern conduit for the maritime circulation of Indic religions, in the period from ca. the 7th to the 14th century those regions transformed across mainland and island polities the rituals, icons, and architecture that embodied these religious insights with a dynamism that often eclipsed the established cultural centres in Northern India, Central Asia, and mainland China. This collective body of work brings together new research aiming to recalibrate the importance of these innovations in art and architecture, thereby highlighting the cultural creativity of the monsoon-influenced Southern rim of the Asian landmass. This approach aims at favouring the integration of these findings into the mainstream academic discourse pertaining to various disciplines, such as Art History, Archaeology, Religious Studies and Textual Studies, which too often have ignored the constitutive role played by the (supposedly ‘peripheral’) maritime regions in

the genesis and development of cultural, religious, and artistic motifs across the mediaeval Indic world. Current research increasingly shows that the Southeast Asian ‘peripheries’ were freer to innovate than the traditional ‘centres’, while the flourishing trade routes provided them with the necessary flow of human and material resources. As a result, the early kingdoms of Maritime Asia became wellsprings of cults, ritual technologies, sacred art/ architecture, and the new political models they underpinned. A transforming southern stream flowed from India and Sri Lanka, across the emerging island and littoral states, touching East Asia and eventually reaching Northeastern India, Nepal, and Tibet. For instance, images in the Chinese Mogao and Yulin caves, which have been hitherto ascribed to Tibetan influence, can now be seen to be influenced by the art of the Pallava dynasty of Southern India and related to teachings introduced to China by Vajrabuddhi/Vajrabodhi and popularized by his pupil Amoghavajra (Khokhlov, this volume). The early propagation of the new maṇḍala- and mantra-based soteriological and ritual systems beyond the subcontinent begins to be recorded in the late 7th century in the journeys of learned and enterprising monks over the so-called (land) ‘Silk Routes’. It expands in the 8th century over the maritime routes, from the monasteries and 1. Adhering to the same intellectual agenda, this anthol- courts of South India and Sri Lanka to the Malay ogy can be considered a ‘sequel’ to the volume Esoteric Peninsula, Java, and then China (with disciples Buddhism in Mediaeval Maritime Asia: Networks of Masters, following on to Korea and Japan). The presence of Texts, Icons, edited by Acri (2016a). On the conceptualizasuch important Mahāyāna and Mantranaya Budtion of ‘Maritime Asia’ as a geographical and historical dhist teachers as Prajña, Vajrabuddhi/Vajrabodhi, theatre for the multi-centric spread of Esoteric Buddhism, Amoghavajra, Śubhakarasiṃha, Suvarṇadvīpīya see Acri 2016b, 2018, 2019; on the idea of ‘Maritime Silk Road’, see Kwa 2016. For an argument in favour of the Dharmakīrti, and Atiśa in the polities arising along ‘Global Middle Ages’ (and the still underappreciated role the ocean and shipping lanes turned them into that Asia played in them), see Holmes and Standen 2015 vibrant centres of innovation rather than passive and 2018; cf. Frankopan 2019.

01_Intro CreSouthV1_4P_22Jan21.indd 1

25/1/22 1:02 PM

2

Andrea Acri and Peter Sharrock

recipients of transmitted Indic ideas, as they have temple, carved with multiple giant faces, was ‘…no longer merely a devagṛha, the traditional sometimes been considered. In spite of the scantiness of the textual record Indian “house of God”, it became itself a God’ of the burgeoning southern trade and rice king- (Maxwell 2007: 86; cf. Sharrock, this volume). doms, where manuscript textual sources surviving Southeast Asian kingdoms were not only the warm and humid climate as well as onwards early adopters of artistic, religious, and cultural religious and cultural developments2 are less elements that have been putatively ascribed to abundant and relatively younger than those that ‘India’, but also innovators—or, indeed, creators. have come down to us from the Indian subcon- A specific form of Amoghapāśa, a manifestation tinent and Tibet, the scale and artistic originality of the compassionate Bodhisattva Avalokiteśvara, manifested in the monumental ritual centres in important across Asia as a saviour deity, whose Java, the Malay Peninsula, Campā, and the Khmer cult now became linked to rituals for the dead, is domains are striking evidence for the religious and now thought to have first appeared in the Malay political novelties devised and implemented in the Peninsula (Sinclair, this volume). A significant and South. The direct access Buddhist and Śaiva masters puzzling fact is that the series of karaṇa dancewere afforded to the royal courts of the emerging poses (as theorized in the Sanskrit Nāṭyaśāstra) Southern states—access which went beyond that depicted on the reliefs of Candi Śiva in Prambanan achieved from the Buddhists’ monastic confines is not only one of the most complete, but also the in the subcontinent—opened the way to the con- earliest, being about 200 years older than the first struction of monuments at large ceremonial centres known series in the Indian subcontinent, at the on an unprecedented scale. Along with the mon- Coḻa Bṛhadīśvara temple in Tanjore (Iyer 1996: umental scale came conceptual leaps that turned 170–171); this opens up to the possibility that ‘the temples into ‘meta-icons’, effectively merging idea of representing karaṇa in a serial order in Cola the Indic disciplines of vastuśāstra (architecture) times was recycled into India after Prambanan was and śilpaśāstra (statuary). For example, early-9th carved’ (ibid.: 171). Such changes were to reverbercentury Borobudur in Central Java, the largest Bud- ate across the Buddhist world for centuries and yet dhist monument on earth, is now seen to embed their origin in the southern seas has hitherto gone the Yogatantra garbhadhātumaṇḍala of deities, as mostly undetected. well as the apex of the vajradhātumaṇḍala, the two While the Gaṇḍavyūhasūtra inspired murals in cosmograms that defined a Buddhist metaphysical Central Asian Buddhist cave temples, sculptures hierarchy that would spread across Asia from the and book illustrations in China, as well as scroll 8th century (Kandahjaya, Volume Two). A few paintings in Japan, what Fontein (2021: 1) called its decades later at Prambanan in Java, some 50km ‘most monumental and by far the most elaborate to the southeast, a 47m-high temple to Śiva was set of illustrations’ was created by the sculptors of erected with approximately 1,000 ribbed, bulbous Borobudur, who were also pioneers in devising a finials which are regarded as liṅgas that carry the large-scale monumental rendering that finds no monsoon rain down through big makara-gargoyles counterparts in the coeval Buddhist world. And to periodically flood the temple courtyard below even if the depiction of the Hindu Epic Rāmāyaṇa with holy water (Sundberg, Volume Two). In the at Lara Jonggrang in Prambanan might have been Khmer kingdom, on Mount Kulen near Angkor, an early adoption from Gupta India, it remains a sahasraliṅga was carved into the riverbed of the unequalled even in the subcontinent for richness Siem Reap river that flowed down to the capital; and completeness, suggesting that Southeast Asian three centuries later, the Buddhist Bayon state masterminds inventively perfected and skilfully implemented Indic ideas. This can also be witnessed from the fact that Southeast Asia documents the 2. For instance, the shift from a Hindu-Buddhist to an earliest and most abundant epigraphical evidence Islamic paradigm in Sumatra and Java, and from a Śaiva of the use of the material zeros in the Indic world and Sanskritic Buddhist to a Pāli Buddhist paradigm in (and the entire world indeed)—their rare Indian the Khmer domains in mainland Southeast Asia.

01_Intro CreSouthV1_4P_22Jan21.indd 2

25/1/22 1:02 PM

Introduction: Volume 1: Intra-Asian Transfers and Mainland Southeast Asia attestations being at least a century later than their late-7th- and 8th-century precursors in Sumatra and Cambodia, and much less abundant than the 9th-century specimens in Java (Diller 1995; Soutif 2008). The problem of architectural prototypes was raised, somewhat preliminarily and provocatively, by Romain (2011) when discussing early temple architecture in Java and India, to ‘question the perpetuated use of influence for understanding the process of diffusion of Indian art in Southeast Asia’, for ‘India related temples in Java appear almost simultaneously with the rise of free-standing stone architecture in South Asia itself’ (ibid.: 314). The southern ‘periphery’ was evidently energetic in sponsoring and devising royal applications of tantric cults of the Buddhist Mantranaya and Vajrayāna as well as of the Atimārga and Mantramārga divisions of Śaivism. Amoghavajra, among others, made contributions to vindicating fierce rituals to support the armies of Buddhist monarchs. The southern innovations followed Vajrabuddhi/Vajrabodhi and Amoghavajra—who outsourced texts and doctrines from South India and Sri Lanka—to China in the 8th century, and Atiśa—who studied in the Malay Peninsula or Sumatra—to Tibet in the 11th century, suggesting that influence in a connected Buddhist world was bi-directional throughout. For a time in the early 11th century, shortly before the Cōḻa invasions devastated several ports in the south and Islamic armies descended from Kabul to plunder Varanasi, Śrīvijaya (with centres in Kedah in the Malay Peninsula and Palembang in Sumatra) emerged as a haven for Buddhist study, with Atiśa sitting for a decade at the feet of Buddhist guru Dharmakīrti of Suvarṇadvīpa.3 And in the ‘late phase’ of Buddhism in the Indian subcontinent, kingdoms in various regions of Southeast Asia (like Java, Sumatra, and the Khmer domains) attracted important religious personalities4 and granted support to artists and 3. See Sinclair 2021, Schoterman 2016, and the chapter by Sharrock in this volume. 4. One such intellectual was Śrī Gautamaśrī(bhadra), a paṇḍita (native of present-day Bengal) who, probably by the first half of the 13th century, engraved a short Sanskrit inscription on a cliff overlooking the sea at Pasir Panjang in Karimun Besar island (Riau archipelago); the same master was active in Northeastern India and Tibet, and taught

01_Intro CreSouthV1_4P_22Jan21.indd 3

3

artisans hailing from Bengal and the Himalayan region, thereby becoming the last bastions of Sanskritic Buddhism in Maritime Asia. *** This two-volume anthology brings together strands of research by both senior and junior scholars into these southern innovations that are visible in icons and architecture, and that contribute to a new sense of the historical role of Maritime Asia. Most of the proposed chapters stem from papers presented in two summer programmes held in East Java in 2016 (co-organized by the SOAS Southeast Asia Art Academic Programme [SAAAP] and the Nalanda-Sriwijaya Centre of ISEAS – Yusof Ishak Institute) and in Central Java in 2017 (co-organized by the SOAS SAAAP and Universitas Gadjah Mada).5 These scholarly gatherings aimed at rediscovering the influence of the Buddhist and Hindu paradigms of mediaeval kingdoms adjoining the maritime trade route. With contributions from leading local and international scholars, the anthology expands on themes of innovation and transfer in the unique monuments, icons, and rituals developed in South India and Sri Lanka, Sumatra, the Malay Peninsula, Java, and the Khmer and Cam domains that strongly impacted Buddhist and Hindu art, architecture, and religious culture in India, Nepal, Tibet, and China, and incite to deeper explorations. The present volume includes nine studies, subdivided into two parts, combining a mix of art historical, archaeological, and textual data to advance innovative interpretations of monuments and icons across Maritime Asia from a perspective that takes into account circulatory dynamics and the creative role of South Indian and Southeast Asian locales in the development of iconographic and architectural motifs. Collectively, the nine studies presented in this volume explore intra- and inter-regional connections across Maritime Asia and its ‘edges’ (for instance, Nepal and northern-central China) as well as regional foci, for instance Mainland Southeast disciples who became influential at the Sino-Mongolian Yuan court (see Sinclair 2018 and this volume). 5. Both programmes were generously (co-)funded by the Alphawood Foundation in Chicago.

25/1/22 1:02 PM

4

Andrea Acri and Peter Sharrock

Asia. They are complemented in Volume Two by nine more studies, two of which mainly dealing with Odisha,6 and the remainder focusing on Java and its translocal artistic and religious echoes.7 Part I, ‘Influences from the South’, presents three chapters highlighting the transfer of icons, artistic styles, and texts from Southern to Northern Asia, as well as their circulation across Maritime Asia. Chapter 2, ‘From Melayu to Thamel and Back: The Transmigration of the Eight-Armed Amoghapāśa’, by Iain Sinclair focuses on the enigmatic figure of Amoghapāśa with eight arms, which is prominent in the tantric Buddhist art of both the Malay-Indonesian Archipelago and Nepal. No scriptural source for it having been identified, previous scholars assumed that the figure was among those transmitted to the Archipelago as part of ‘Indo-Tibetan’ Buddhism. In the first comprehensive, cross-cultural iconographic study of this type, surveying all known examples and drawing on previously unstudied texts that are here edited and translated for the first time, Sinclair lays out a different scenario, namely that the earliest images originated within the former Melayu kingdom, soon diffusing to insular Southeast Asia, East Asia, and eventually Nepal, probably at the hands of Atiśa Dīpaṃkaraśrījñāna, while the latest icons were reintroduced from the subcontinent to East Java in the 13th century, probably by the paṇḍita Gautamaśrībhadra. This appears to be a remarkable case of ‘periphery’ culture migrating over a long period and distance to the ‘centre’, as well as of the feedback loop popularly known as a ‘pizza effect’.8 A case of long-distance transfer of iconographical themes from Southern Asia to Central and East Asia is documented in Chapter 3 by the late Yury Khokhlov, ‘In the Footsteps of Amoghavajra (705–774): Southern Indian Artistic Mode in Tang China and its Transmission to Tibet’.9 Khokhlov’s research analyses the images

from Mogao and Yulin caves, which have been traditionally ascribed to Tibetan influence and believed to represent the early Tibetan style. Contrary to this common opinion, the author shows that these images are in fact influenced by Southern Indian art and in particular by the art of the Pallava kingdom in Tamil Nadu, and closely affiliated with teachings introduced to China by Vajrabuddhi/Vajrabodhi in the first half of the 8th century and specifically propagated in Hexi by Amoghavajra in the mid-8th century. Furthermore, the chapter argues that the Tibetan conquest of the Hexi corridor enabled the still unacknowledged Tibetan appropriation of Chinese religious and artistic trends, which were popular in the region, and casts further light on the Tibetan adoption of Buddhism as a state religion following the model of the Chinese state-protective Buddhism established by Amoghavajra. Starting from the assumption that we are currently underestimating the influence from the Maritime Asian courts on wider Indic tantric paradigms, Chapter 4, ‘Heruka-Maṇḍalas across Maritime Asia’ by Peter Sharrock, discusses the royal abhiṣekas or initiations into the system of the fierce Esoteric Buddhist deity Heruka. Whereas the Heruka cults in Northern India and Tibet were confined to secret monastic settings, in the courts of the Southern Seas, political applications of the new ritual technology brought the icons and ceremonies into more public and exoteric modes. The chapter documents the manifestations of this cult in the Khmer empire, Campā, East Java, Sumatra, and China under the Mongols. Part II, ‘Transfers and Innovations in Mainland Southeast Asia’, contains six chapters dealing with aspects of Hindu and Buddhist art and architecture in the Khmer domains and Campā, and contextualize them in the religious history of the wider Southern Asian region. Chapter 5, ‘Goddess Pra-

6. By Sonali Dhingra and Umakant Mishra. 7. By Hudaya Kandahjaya, Saran Suebsantiwongse, Michel Gauvain, the late Roy Jordaan, Mimi Savitri, Jeffrey Sundberg, and Hadi Sidomulyo. 8. A Neapolitan dish that has flourished and evolved in the US and has been further exported at distance around the world and re-imported into Italy. 9. This essay was originally published as ‘Uncovering

Amoghavajra’s Legacy in the Hexi Corridor and Tibet’ in the Journal of Tibetan Studies 21 (2019). On account of the importance and pertinence of its findings, a revised and updated, and additionally peer-reviewed, version is published in this anthology. Sadly, the author untimely passed away just a couple of months after submitting the final version of the article. We would like to thank Yannick Laurent for his assistance in editing Khokhlov’s manuscript.

01_Intro CreSouthV1_4P_22Jan21.indd 4

25/1/22 1:02 PM

Introduction: Volume 1: Intra-Asian Transfers and Mainland Southeast Asia jñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor’, by Jinah Kim explores the religious and political significance of Prajñāpāramitā during Jayavarman VII’s reign (c. 1181–? 1218) from a trans-regional comparative perspective that locates Angkor in the trans-regional network of Esoteric Buddhism of the 12th–13th centuries. By contextualizing the use of Prajñāpāramitā in Angkor and the larger historical context, the chapter argues that the analogy of Prajñāpāramitā as the mother of all Buddhas was firmly materialized in Angkor more than anywhere else in the Buddhist world of the same period. By analysing epigraphic, textual, and art historical data not only from the Khmer context but also from India and Java, it demonstrates how Jayavarman VII’s political adaptation of the Prajñāpāramitā-mother analogy is reflected in the unique Khmer iconography of the Buddhist goddess Prajñāpāramitā. This further confirms that Khmer examples are not ‘outcastes’ in the iconographic genealogy of the goddess Prajñāpāramitā, but innovative yet politically sensitive creations that can help us understand the historical process behind the formation of a divine iconography in a doctrinally complex situation. Chapter 6, ‘Dancers, Musicians, Ascetics, and Priests: Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor’ by Swati Chemburkar documents and re-examines the Śaiva ritual practices enjoining performance in the mediaeval Khmer domains. Combining data mined from the Śaiva textual archive in Sanskrit and vernacular languages with hitherto neglected art historical material from the Khmer domains, as well as insights from Campā, Java, and South India, the chapter first focuses on the Pāśupata evidence of performance-based worship in the pre-Angkorian period, then documents the survival of this sect as an elite of state-sponsored ritualists through the Angkorian period, and finally describes the development of performance in temples and festivals in both Śaiva and Buddhist milieus in the late Angkorian period. Rather like the Buddhist cult of Prajñāpāramitā, Pāśupata ascetics found fertile ground among the Khmers long after they declined in the subcontinent. Chapter 7, ‘Libraries or Fire Shrines? Reinterpreting the Function of “Annex Buildings” in Khmer Śaiva Temples from the Prism of Early

01_Intro CreSouthV1_4P_22Jan21.indd 5

5

Śaivism’, by Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin, deals with a unique and enigmatic feature of Khmer temple complexes dating from the early Angkorian to the late Angkorian periods, namely the small annex building with ventilation holes built in the southeastern quarter of the main temple, which scholars have tentatively called either a ‘library’ or a ‘fire-house’. The chapter suggests that these annex buildings, which appear to have no clearly identifiable counterparts in South Asian temples, may have been intended to function as both repositories for manuscripts and sacred spaces for specific Śaiva rituals, including ash-related observances of the Pāśupata sect, as well as initiation- and homa-rituals. Chapter 8, ‘Śaiva Religious Iconography: Dancing Śiva in Multi-polity Medieval Campā’, by Mai Bùi Diệu Linh opens a series of three chapters focusing on Campā. It examines the identity and position of Śiva Naṭarāja, the ‘Lord of Dance’ manifestation, within the Cam mediaeval religious landscape, particularly in Indrapura, Amaravati, and Vijaya. Having identified seven Cam sculptures of the dancing Śiva in museums and temples in Vietnam, it asks whether Naṭarāja was ever worshiped in a Cam Śaiva temple as the main deity. Sculptures produced in different regions and in various time periods had their own unique styles and these differences suggest a dissemination process through multiple channels over a considerable period of time. The author’s analysis defines the independent political stance of each Cam polity as mirrored in its own icons and tackles the issue as to whether they evolved in a unified chronological sequence. Chapter 9, ‘The Colossal Trà Kiệu Rāmāyaṇa Pedestal in Campā and its Relationship to Courtly Culture in Cambodia, East Java, and China’, by Mya Chau reviews the existing scholarship on the Trà Kiệu pedestal from Campā within the context of the Rāmāyaṇa in Khmer and Javanese cultures during the 7th–8th century. Using textual sources and secondary literature, iconographic evidence, and recent excavations, the author suggests that the Trà Kiệu pedestal can be understood as a translocal depiction of the Rāmāyaṇa predominantly inspired by an oral narrative tradition aiming at heightening the real and imagined courtly culture for others

25/1/22 1:02 PM

6

Andrea Acri and Peter Sharrock

beyond the king and his advisors. Visual evidence such as the Rāmāyaṇa reliefs from the Đà Nẵng Museum, bas-reliefs from the Chiên Đàn towers, Khương Mỹ temple reliefs, and portable bowls depict imagery related to the Trà Kiệu pedestal that reinforces military and courtly culture, thereby promoting an imagined courtly culture of Campā displaying close relationships with Khmer and East Javanese iconography. Chapter 10, ‘On the Chronological Interrelationship between the Newly Found Inscriptions and the Temple Architecture of Campā: The Hòa Lai and Po Dam Sites’, by Trần Kỳ Phương analyses approaches to the Cam temple architectural history advanced in previous research. Through a comparative study of the newly found dated inscriptions at Hòa Lai and Po Dam temple sites, it analyses the process of structural techniques and the decorative patterns of these two temple complexes in order to propose a date for each monument. The site yields indicators of cultural transfers to the major early Khmer sites at Sambor Prei Kuk and Mount Kulen, as well as to Java and to the Malay Peninsula, reflecting a contemporaneous interaction in the region through trading connections by land, sea, and waterways.

References

Acri, Andrea. 2016a. Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons. Singapore: ISEAS – Yusof Ishak Institute. . 2016b. ‘Introduction: Esoteric Buddhist Networks along the Maritime Silk Routes, 7th–13th Centuries CE)’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons, pp. 1–25. Singapore: ISEAS – Yusof Ishak Institute. . 2018. ‘Maritime Buddhism’, Oxford Research Encyclopaedia of Religion (Online). Oxford University Press. [https://doi.org/ 10.1093/acrefore/9780199340378.013.638] . 2019. ‘Imagining “Maritime Asia” ’, in Andrea Acri, Murari K. Jha, Sraman Mukherjee, and Kashshaf Ghani (eds.), Imagining Asia(s): Networks, Agents, Sites, pp. 36–59. Singapore: ISEAS – Yusof Ishak Institute. Diller, Anthony. 1995. ‘Śrīvijaya and the First Zeros’, Journal of the Malaysian Branch of the Royal Asiatic Society 68/1 (268): 53–66.

01_Intro CreSouthV1_4P_22Jan21.indd 6

Fontein, Jan. 2012. Entering the Dharmadhātu: A Study of the Gandavyūha Reliefs of Borobudur. Leiden: Brill. Frankopan, Peter. 2019. ‘Why We Need to Think About the Global Middle Ages’, Journal of Medieval Worlds 1/1: 5–10. Holmes, Catherine and Naomi Standen. 2015. ‘Defining the Global Middle Ages (AHRC Research Network AH/K001914/1, 2013-15)’, Medieval Worlds 1: 106–117. . 2018. ‘Introduction: Towards a Global Middle Ages’, Past & Present 238, Issue supplement 13: 1–44. Iyer, Alessandra. 1996. ‘Prambanan Revisited: A Fresh Perspective on the Dance Sculptures of Candi Śiwa’, Bulletin de l’École française d’Éxtrême-Orient 83: 157–184. Kwa, Chong Guan. 2016. ‘The Maritime Silk Road: History of an Idea’, Nalanda-Sriwijaya Centre Working Paper no. 23. Maxwell, Thomas S. 2007. ‘Religion at the Time of Jayavarman VII’, in Joyce Clark (ed.), Bayon: New Perspectives, pp. 72–135. Bangkok: River Books. Romain, Julia. 2011. ‘Indian Architecture in the “Sanskrit Cosmopolis”: The Temples of the Dieng Plateau’, in Pierre-Yves Manguin, A. Mani, and Geoff Wade (eds.), Early Interactions between South and Southeast Asia: Reflections on Cross-cultural Exchange, pp. 299–316. Singapore: Institute of Southeast Asian Studies. Schoterman, Jan A. (Roy Jordaan and Mark E. Long, trans.). 2016. ‘Traces of Indonesian Influences in Tibet’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons, pp. 113–121. Singapore: ISEAS – Yusof Ishak Institute. Sinclair, Iain. 2018. ‘New Light on the Karimun Besar Inscription (Prasasti Pasir Panjang) and the Learned Man from Gaur’, NSC Highlights 11 (December 2018–February 2019): 16–17. . 2021. ‘Dharmakīrti of Kedah: His Life, Work and Troubled Times’, Temasek Working Paper no. 2. Soutif, Domnique. 2008. ‘Dénombrer les biens du dieu: Étude de la numération du vieux khmer (VIe–XIIe siècles śaka)’, Journal of Cambodia Research 10: 51–80.

25/1/22 1:02 PM

Chapter 2

From Melayu to Thamel and Back: The Transmigration of the Eight-Armed Amoghapāśa Iain Sinclair1

T

he form of Amoghapāśa that has one face and eight arms is generally regarded as part of the Indian pantheon, but the depictions from the subcontinent only began to appear after the downfall of Indian Buddhism in the early 13th century. A thorough survey of the artistic corpus and the salient textual material, which is undertaken here for the first time, shows that this form of the bodhisattva was already widespread throughout the Malayo-Javanese world half a millennium earlier. The eight-armed Amoghapāśa thereby appears to have originated in the Malay Archipelago, and to have spread from there to East Asia and the Khmer Empire. Its South Asian debut came much later, in the monasteries housing foreign monks in the Kathmandu Valley, especially the Vikramaśīla branch monastery in the district of Thamel. Then, in the late 13th century, the eight-armed form finally reappeared in Sumatra and Java, taking centre stage 1. This chapter was written during a Visiting Fellowship at the former Nalanda-Sriwijaya Centre, ISEAS – Yusof Ishak Institute, Singapore. It expands on a subject first explored in my doctoral thesis (Sinclair 2016b: 146–153). Some of its findings were subsequently presented in an invited lecture at the Asian Civilisations Museum, Singapore (Sinclair 2019c). I dedicate it to the memory of the late Buddharatna Vajrācārya, who introduced me to the praxis of the eight-armed Amoghapāśa at Lalitpur’s ancient Northern Stupa over two decades ago. I gladly thank Andrea Acri for constructive comments and encouragement to publish; E. Edwards McKinnon, Hueimin Chen, Julio Jeldres, Kwa Chong Guan, Peter Sharrock, David Templeman and the staff of ISEAS Library for locating useful material; the two anonymous reviewers for their learned comments; and Tai Yew Seng for advice on passages in Chinese. Any shortcomings are my responsibility alone.

02_CreSouthV1_8P_4Feb22.indd 9

during the Javanese Siṅhasāri dynasty’s occupation of Melayu. The Siṅhasāri images of the bodhisattva show signs of direct input from the subcontinent, which can be linked circumstantially to exiles such as the paṇḍita Gautamaśrībhadra. By this time, the bodhisattva had undergone many transmigrations, from a seafaring deity who saves devotees from the eight great fears to a bodhisattva who stands at the gateway to advanced tantric practice.

the eight-armed avalokiteśvaras from śrīvijaya The eight-armed, single-faced form of Avalokiteśvara, who is usually identified as Amoghapāśa, is well represented in the art of the Malay Archipelago. More than twenty sculptures of this form, dated to between the 7th and early 11th centuries on the basis of their style, have been discovered there. An inventory of early depictions originating from this region is provided in Appendix 2.1. Only the twoand four-armed forms of Avalokiteśvara are more frequently represented in the period corpus than the eight-armed form. What has been overlooked so far is the fact that no sculptures of the standing eight-armed form have been found in India; all such examples from the subcontinent are Nepalese. Furthermore, the many written descriptions of this form produced on the subcontinent are also much later than the period of initial diffusion in the Archipelago, and also, it seems, the Nepalese depictions. As such, the iconography of the eightarmed form is not only unusually prevalent in Maritime Southeast Asia; it appears to have originated there, or at least, to have enjoyed its earliest, most definitive success in the region. The eight-armed Avalokiteśvara can then be appraised as an extraor-

3/2/22 6:16 PM

10

Iain Sinclair

dinary development: an Austronesian innovation in an otherwise Sanskritic socioreligious environment. It represents a challenge to entrenched narratives of the region’s ‘Indianization’, which Andrea Acri (2017) and others have problematized. The fact that the early sculptures of the eightarmed bodhisattva are peculiar to the Archipelago has been noticed before. Nandana Chutiwongs observed: ‘Such manifestations of Avalokiteśvara, curiously enough, have never been found in Indian art’ (1984: 344). However, this observation has not led to the proposition that the eight-armed form could be native to the region. Instead, the prevalence of the form in Nepal, remarked on by Chutiwongs (ibid.: n. 398, 442) and others, has supported a tacit assumption that eight-armed Avalokiteśvaras existed all over the subcontinent. But all depictions of the eight-armed form found on the subcontinent are Nepalese, not Indian.2 They are also much later than the Malayo-Javanese ones. The Nepalese images are so late that they bypass ‘Indian Buddhism’ entirely, insofar as Buddhism in India is conceived as a pre13th century phenomenon.3 This is a remarkable situation in itself. There are two regional corpuses of depictions, separated from each other by a large geographical distance as well as a long interval, and so far, no direct connection between them has been found. The earliest of the Southeast Asian depictions is the bronze from Bantaeng in South Sulawesi (11; numbers in bold refer to entries in Appendix 2.1). It is dated to between the 7th and 8th centuries by Pauline Lunsingh Scheurleer and Marijke J. Klokke (1988: 111). Another seven sculptures can be assigned to about the 8th century (3, 6, 7, 10, 13, 17, 18). The remainder of the corpus appears to have been produced between the 9th and 10th centuries. The sculptures produced in the Javanese Siṅhasāri dynasty (1222–1292) are outliers and represent a revival movement, as I will argue later in the 2. For instance, Donaldson 2001: 201–202, in an otherwise exemplary survey of Buddhist sculpture of Odisha, states that eight-armed Avalokiteśvaras are ‘most popular’, but all examples or sources referred to by Donaldson originate in Nepal. 3. Buddhists persisted on the Indian subcontinent beyond the 13th century (Sinclair 2019b: 295–299), but were confined in enclaves that had little contact with each other.

02_CreSouthV1_8P_4Feb22.indd 10

chapter. While the style-based dating of the corpus is imprecise, the approximate range of dates is not currently in doubt. The findspots of the early eight-armed Avalokiteśvaras are concentrated in Malay-speaking areas of Southeast Asia. Four bronzes are from Palembang and the Sumatran highlands (1, 3, 13, 16), and another six are from the northern fringe of the Malay Peninsula, including parts of present-day Thailand (8, 9, 14, 15, 17, 18). Two bronzes were found at farther reaches of the Malay world: the above-mentioned Bantaeng bronze, and a figure from Phan Rang in former Cam country (19). Both of these are considered to be ‘imports’,4 the latter coming from Śrīvijaya. Central Java is identified as the source of just two bronzes (4, 12), and the probable but unverified source of six others (5, 6, 7, 8, 10, 21)—the smaller proportion of the corpus. A further two bronzes (2, 20) have unspecified places of origin but are generally understood to have come from the Malay Archipelago, as they belong to collections of other objects found in the region. On the whole, the distribution of these images clusters around the Malay parts of Maritime Southeast Asia, and in the middle of the 7th–10th century interval. There are minor differences in their ornamentation, hairstyles, and modes of dress, some of which will be examined further below. Sumatran, Peninsular and Javanese bronzes of the eight-armed figure are not stylistically differentiable at this stage of analysis. The corpus is homogenous enough to constitute a common ‘Śrīvijayan type’. The term ‘Śrīvijayan type’ will then be used, albeit with the caveat that not all specimens can be definitely associated with a Śrīvijayan polity. Śrīvijaya, however its territory may be demarcated, is generally understood to have centred on the city of Palembang and the river formerly called Melayu.5 While the art-historical record reveals that standing Avalokiteśvaras with eight arms first appeared in the Malayo-Javanese world, their models or typological precursors are to be sought on the 4. Chutiwongs 1984: 442. 5. Palembang once understood itself to have been the central polity of Śrīvijaya, in relation to a more fluid periphery (Manguin 2017: 99). On the location of Palembang’s Tatang River and its ancient name Melayu, see Manguin ibid.: 102 and Kitchener and Kustiarsih 2019: 7.

3/2/22 6:16 PM

From Melayu to Thamel and Back

11

Fig. 2.1a: Eight-armed Avalokiteśvara (11). Bronze, c. 7th century. Bantaeng, Sulawesi. On loan from the Asian Art Society, The Netherlands. (Public Domain)

Fig 2.1b: Avalokiteśvara. Bronze, date unknown, ‘reportedly from Sumatra’. Hand drawing by author based on Guy 1995, fig. 14.

Indian subcontinent. Śrīvijayan Avalokiteśvara statues were first associated with South Indian style by Marie Thérèse de Mallmann (1951: 573–574). Certain features of the earliest Śrīvijayan example, the Bantaeng bronze, have been likened to those of Pallava art of the 7th to 8th centuries. The bronze is believed to have been an import—that is, ‘not made on the island on which it was found’, because ‘no other bronze figurine from the area is known, resembling this one’ (Lunsingh Scheurleer and Klokke 1988: 111). However, it can be added here that the Bantaeng bronze is not a stylistic isolate in the region. It looks to have come from the same workshop as another Avalokiteśvara statuette ‘reportedly from Sumatra’, which displays common features including a ‘bejewelled waistband … a recurrent feature of bronzes of south India’ (Guy 1995: 79). Other features reminiscent of South Indian period sculptures, such as long matted hair ends, are displayed by some bronzes (1, 2).6 Contacts between the Archipelago and the South of the subcontinent

are, in general, well known from written sources such as Tang-era travelogues.7 Although there are few if any surviving Pallava bronzes that could be nominated as models for the Śrīvijayan type, it is known that bronze statues of Avalokiteśvara were being cast in South India from the 7th century onwards, if not earlier.8 A faint indication of a cult of the bodhisattva in South India comes from a story in a Chinese anthology dating to the 11th century or thereabouts, the Sanbao ganying yaolüe lu 三宝感应要 略录. One story in this anthology tells of an image of Avalokiteśvara set up on the advice of a monk from the north. It concerns the noose-wielding

6. See e.g. Eilenberg 1997: 170, fig. 8.

02_CreSouthV1_8P_4Feb22.indd 11

7. Maritime trade between Southeast Asia and the subcontinent in this period was concentrated in the ‘southern rim’, as Acri 2018: 140–141 argues. 8. Srinivasan 2013: 172, fig. 13.2 states that a bronze in the Victoria and Albert Museum, IM.300-1914, identified as Avalokiteśvara (it lacks the characteristic Amitābha in the crown), has been ‘technically finger-printed to the early Andhra Pallava period’. On this bronze see also Guy 2014: 7, vol. I, cat. 7.

3/2/22 6:16 PM

Iain Sinclair

12

form of Avalokiteśvara known as Amoghapāśa— the name by which the eight-armed form is being called at this time. The story is titled ‘A South Indian Country Makes an Amoghapāśa Icon [which] Responds [to Prayers]’ 南印度國造不空羂索像感應: A country in South India was desolate, the ruler and ministers losing health, the people already dying. The king sent an envoy to ask the Central Indian [monk] Śrīmitra to save the country from disaster. Mitra came to this country, telling the king: ‘There is a great saint, the Infallible Noose [Amoghapāśa] Avalokiteśvara, O great king. Just make [his] image, install [it in] the city’s southwestern tower.’ The king accepted the teaching, made the image right away and installed [it] in the west of the city, a pavilion in the south. Radiance from the image shone for one yojana. The king and minister kept their health. The five grain [crops] were abundant, the people had a surplus. The country was touched by that image. There was an old monastery in the south of the city; the pavilion was moved to the monastery.9

Whatever the historicity, if any, of this tale, it may well ‘originate in a Western Regions chronicle’ (chu Xiyu ji 出西域記), as is claimed.10 It is also interesting that the anthology elsewhere refers to *Suvarṇabhūmi (Jindiguo 金地國)—the very area of interest—even if only in the sense of a ‘place of legend’.11 In any case, there is no indication as to which of the many forms of Amoghapāśa is being discussed.

9. 南印度國荒廢。君臣不保壽。人民已喪。王遣使。請中天 竺尸利蜜多。欲救國災。密多來至此國。白王言:有大聖不空 羂索觀自在[。]大王。方造像安置城西南閣。王受教已。即造 形像。安置城西南閣。像放光明。照一由旬。王臣保壽。五穀 豐饒。人民從餘[。]國愍感其像。有城南故寺。即移閣爲寺矣 (T 2084 853b1–7, with modified punctuation). 10. T 2084 853a29. An Indic name such as Śrīmitra, and for that matter any detailed account of affairs in South India, is unlikely to have been invented by the Chinese compiler of this anthology. 11. Matsumura 2011: 263. The short story in question is said to be taken from an (otherwise unidentified) ‘Parable Scripture’ (Piyu jing 譬喩經); see T 2084 836a 6–10. This relatively early reference is omitted from the discussion of Suvarṇabhūmi’s toponomy in Skilling 2019.

02_CreSouthV1_8P_4Feb22.indd 12

the scriptural undergirding of the eight-armed amoghapāśa A canonical, Buddha-spoken description of a form of Avalokiteśvara or Amoghapāśa with eight arms has not yet been identified. Protocols for worshipping images of Amoghapāśa are, in general, documented in texts that were in circulation by the 7th century. The two principal works are the Amoghapāśahṛdaya (hereafter Hṛdaya), the earliest Buddha-spoken text on the bodhisattva, and its major expansion, the Amoghapāśakalparāja (hereafter Kalparāja). The Hṛdaya was widely disseminated throughout Asia. The less common Kalparāja has been transmitted in a single Sanskrit manuscript (hereafter ‘MS’12), in a Tibetan translation and a Chinese translation. The MS and the translations convey different recensions of the text.13 The Kalparāja builds a ritual edifice around the Hṛdaya, emphasizing image magic.14 Many of the manifestations of Avalokiteśvara described in the Kalparāja are informed by the appropriations of theist imagery revealed in the Kāraṇḍavyūha, a narration of Avalokiteśvara’s deeds that began circulating at around the same time.15 A common thread throughout this literature is that the bodhisattva partakes of and supersedes the powers of various Hindu gods, whom he is said to incorporate.16

12. The text of the East Indian manuscript of the Amoghapāśakalparāja kept in Tibet is being published in irregular instalments at Taisho University. The published facsimile of the MS (Mikkyō Seiten Kenkyūkai 1996) has not been used in the present study. 13. According to Tanaka 2019: 44, there are two primary recensions: one corresponding to the Sanskrit MS and the Tibetan translation, and another corresponding to the Chinese. 14. Shinohara 2014: 86 regards this ‘image worship’ as a fundamentally new development. 15. On this central theme of the Kāraṇḍavyūha, see Sinclair 2015: 435ff. 16. An important example of imagery shared with the Kāraṇḍavyūha is the eleven-faced Avalokiteśvara whose body creates and extrudes the heads of eleven Hindu gods (cf. Sinclair 2015: 436), translated from Kalparāja MS f. 69b2–4 in Sinclair 2016b: 154–155.

3/2/22 6:16 PM

From Melayu to Thamel and Back While manifestations of Avalokiteśvara with four, six, ten, sixteen and thirty-two arms are described in the Kalparāja, an eight-armed, single-faced form is not. There is only a brief description of a form with three faces and a partly dissimilar array of eight hands (noose, lotus, trident, wishing gem, club, reassurance pose, rosary, flower bunch).17 A description corresponding to the Śrīvijayan type is yet to be found in any canonical source—in the Chinese canon (Wong 2015: 48–52), in the Tibetan bKa’ ’gyur or in large extracanonical corpora such as the Dunhuang manuscripts (van Schaik 2006). Nor has it been noticed in previous longue durée studies of Amoghapāśa (e.g. Tuladhar-Douglas 2006) that the extant descriptions of the type in Sanskrit only started circulating several centuries after the advent of the first images. A conceptual basis for the eight-armed iconography must then be sought in the entire religious milieu surrounding Avalokiteśvara up to the 7th century. Many aspects of the bodhisattva’s appearance are consistent with the iconographic programme of the Kalparāja. The Śrīvijayan eight-armed form looks like a canonical four-armed form18 with two pairs of hands added, or a six-armed form with one pair of hands added, or a ten-armed form with one less pair of hands. There is an especially close resemblance to the standing six-armed form, the so-called Pretasaṃtarpita form,19 which was widely 17. don yod pa’i rgyal po ni [...] phyag brgyad pa/ dbu gsum pa/ de bzhin du spyan gsum pa/ zhi ba’i zhal du bri bar bya ste de bzhin du zhags pa ’dzin pa dang padma ’dzin pa dang mdung rtse gsum pa dang/ ril ba spyi blugs ’dzin pa dang/ yid bzhin gyi nor bu dang/ be con dang dbugs ’byin pa dang/ bgrang phreng dang/ me tog thum po ’dzin par bya’o (D 686 256b3–5). 18. The first description of a four-armed Amoghapāśa in the Kalparāja is given with instructions for making a sandalwood-paste effigy: amoghapāśaṃ kartavyaṃ sarvālaṃkāravibhūṣitaṃ jaṭāmakuṭamaṇḍitam eṇeyacarmapaśupativeśadharaṃ caturbhujam eka[ṃ] padmadharaṃ dvitīya[ṃ] pāśadharaṃ tṛtīyaṃ śūladharaṃ caturtham āśvāsanaṃ padmāsanaṃ vyāmaprabhasamalaṃkṛtam (MS, f. 42a7–b1, via Suzuki, Ōtsuka and Kimura 2000: 52–53). Note that citations from the Kalparāja are given here with regularized orthography. Other four-armed forms are described in the early literature (see e.g. Reis-Habito 1999: 59, 63–64). 19. Leoshko 1985 and Bhattacharya 2001: 34–36 have studied East Indian examples of standing six-armed

02_CreSouthV1_8P_4Feb22.indd 13

13

depicted across East India and to a lesser extent in Nusantara. The eight-armed form, seen alongside these canonically codified forms of Amoghapāśa, appears to be an improvised filling of a gap in the pantheon. From the start, the eight-armed form shows signs of being transmitted mainly as an image cult, and this mode of transmission may reveal something about its environment of origin. A distinctive feature of many early depictions of the bodhisattva, which has a literate basis, is the tiger pelt wrapped on top of a lower garment or dhoti (2, 3, 6, 8, 11, 13, 14, 18, 20). The requirement for a pelt goes back to the Hṛdaya,20 and is repeated in the Kalparāja: Amoghapāśa wears the garment of Paśupati (paśupativeśadhara), that is, a tiger pelt. The most straightforward interpretation, in the context of the Hindu Paśupati/Rudra mythos, is that a tiger pelt is worn on its own, as a ‘single garment’,21 like a kilt, as befits a ‘Lord of Sacrificial Beasts’. To interpret the paśupativeśa as an additional lower garment might be regarded as an indication of remoteness from the classical cultural context. There are also examples of this interpretation in the non-Buddhist art of the Archipelago—at Prambanan, for instance.22 The use of pelt-over-dhoti style is another indication that the Śrīvijayan Amoghapāśas are not modelled wholly on Indian images. They share in a visual idiom that was common in Maritime Asia at the time.23 Certain elements of the iconography of the eight-armed form are consistent across the Śrīvijayan corpus, but there is also variation in the corpus that is largely missing in later representations and has not been explained. In most depictions, a book, a lotus, and a waterpot are held in the Avalokiteśvaras. Bhattacharya identifies them as the Pretasaṃtarpitalokeśvara specified in Sādhanamālā 43 (ed. Bhattacharyya 1925: 89; Sakuma 2002, No. 16), but it is doubtful that this text is the direct source of the earliest depictions (cf. Woodward 2007: n. 24). 20. See ed. Meisezahl 1962: 322 9, 3275. 21. On the directive that Pāśupata ascetics should be ekavāsāḥ (this single garment can be an animal pelt), see Chakraborti 1970: 77. 22. Nik Hassan Shuhaimi 1977: 26–27. 23. Bopearachchi 2017: 49, 61 has identified a couple of Sri Lankan pelt-over-dhoti Avalokiteśvara images, which are presumably recipients of Archipelago style.

3/2/22 6:16 PM

14

Iain Sinclair

left hands, and the right hands display a rosary, a gesture of giving and a gesture of fearlessness. The objects held in the fourth pair of hands vary. In later, standardized depictions of the form, the right hand holds a noose and the opposing hand holds an ascetic’s three-branched staff (tridaṇḍa), but in the Śrīvijayan bronzes a noose or visually similar object is generally held in a right hand. These early sculptures display items that resemble a threebranched staff—a noose (5, 18), hook (14), conch (12), or trident (2). Some forms of Amoghapāśa are required by the Kalparāja to carry a trident (triśūla).24 The trident and conch would probably be seen as objects more representative of Śaivism or Vaiṣṇavism, respectively, than of Buddhism. The god Harihara, for instance, holds both a conch and a trident.25 Eight-armed forms of Nārāyaṇa and Maheśvara were certainly known to literate Buddhists,26 and the eight-armed Durgā is an early, persistent figure in the visual culture of Java. It is hard to avoid the impression that the eightarmed Amoghapāśa aims to capture some of the otherworldly grandeur of Vaiṣṇava and Śaiva religiosity, as other Avalokiteśvara forms clearly attempted to do.27 The fact that all the Amoghapāśas from the Archipelago have their crowns marked with the Buddha Amitābha—a rare element of consistency— at the same time shows a certain keenness to avoid confusion with Hindu cult objects. There are, furthermore, concrete indications that the eight-armed form invokes the age-old notion of Avalokiteśvara as the pacifier of the Eight Great Fears.

24. For one prescription of a form with a trident see e.g. note 18 above; see also Kalparāja MS f. 53b3. For magical uses of the (tri-)śūla, see MS ff. 30a 2–3, 38a 2 . Wong 2015 lists references to tridents in the Chinese translation T 1092, ff. 265, 266, 281, 304–305, 327, 346, etc. 25. Adiceam 1966. 26. See e.g. Ru lengjia jing 入楞伽經 (Laṅkāvatāra) 10.797a, viṣṇur maheśvaraś cāpi (ed. Vaidya 1963: 15924), translated as 八臂那羅延及摩醯首羅 (T 671 584b7), ‘Nārāyaṇa the Eight-armed and Maheśvara’, by Bodhiruci 菩提流支. Many more references could be given to eight-armed (ba bi 八臂) Nārāyaṇa (e.g. Tuoluoni ji jing, discussed below, T 1336, f. 591a) and other Hindu deities (tian 天) in Chinese Buddhist texts; see e.g. Reis-Habito 1999: 53. 27. Sinclair 2015: 453, 2016b: 153.

02_CreSouthV1_8P_4Feb22.indd 14

two indian eight-armed avalokiteśvaras and the eight fears Two atypical sculptures of an Avalokiteśvara with one face and eight arms were produced on the subcontinent in the first millennium. They are seated figures, unlike the Śrīvijayan icons, but they provide the closest points of comparison from outside Maritime Southeast Asia. The earlier of the two sculptures, first described by Akira Miyaji (2000), is said to have come from Swāt in present-day Pakistan. The later sculpture, as noticed by Bautze-Picron (2004: 288, fig. 40), is from Nālandā. The crosslegged pose of both sculptures is iconographically distinct from the standing pose of nearly all Malayo-Javanese images.28 One figure in Ronggowarsito Museum (21), depicted in a seated pose, is the lone exception that proves the rule. It has one leg dangling, which is again unlike the two sculptures from the subcontinent. The Nālandā sculpture, furthermore, holds a pair of hands at its heart in a pose that is not seen in the Śrīvijayan corpus and which marks it as another distinct type, a separate ontogeny, which did not extend into the Archipelago. There are, in addition, eight-armed forms of the dancing Avalokiteśvara, Padmanarteśvara, which represent an independent and unrelated iconography; the few extant South Asian depictions need not be discussed here.29

28. One specification allows Amoghapāśa to be depicted either sitting or standing: […] āryāvalokiteśvaram amoghapāśaṃ citrayitavyam | utthitaṃ vā niṣaṇṇaṃ vā paryaṇkaniṣaṇṇaṃ vā likhya (Kalparāja MS f. 45a 2, via Suzuki, Ōtsuka and Kimura 2000: 60). 29. A sculpted stone toraṇa from Bihar featuring ‘the eight-armed Avalokiteśvara’, as described by Claudine Bautze-Picron (2019: 31), plainly represents Padmanarteśvara, albeit with a non-standard iconography that has little relation to the eight-armed form described in Sādhanamālā 31 (ed. Bhattacharyya 1925: 76). In this toraṇa the entire ensemble is, unusually, dancing: Avalokiteśvara holds clappers (in the view of Bautze-Picron) and his four followers play musical instruments. Since another toraṇa at the same site depicts a dancing Mañjuvajra, its iconographic programme could be investigated in connection with Vṛttamālāstuti 119, a metrical work by Jnānaśrīmitra that is nominally dedicated to Mañjuvajra. Śākyarakṣita’s commentary on this verse (ed. Hahn 2014: 14) characterizes Padmanarteśvara as mañjuvajrasyāmnāyaviśeṣaḥ.

3/2/22 6:16 PM

From Melayu to Thamel and Back The eight-armed Avalokiteśvara sculpture from Swāt is novel in that it arranges portrayals of the classical Great Fears (mahābhaya) around the main figure. The Saddharmapuṇḍarīka, which contains early and well-known preaching on Avalokiteśvara’s salvific power, is identified by Miyaji as the primary source for the Fears depicted in this sculpture. It should, however, be pointed out that the relevant passage, Saddharmapuṇḍarīka 24.5–16, can be read as specifying up to twelve Fears.30 Sets of Eight Great Fears (aṣṭamahābhaya) are set out in other scriptural traditions, which Motohiro Yoritomi (1990: 643ff.) has identified. The Swāt sculpture shows, in spite of its poor condition, the bodhisattva reaching into a number of frightful scenes: He saves devotees from the Execution, Rapacious Beast, and Firepit Fears with his left hands, and from the Rocks and Bondage Fears with his right hands. The whole tableau of the Swāt sculpture is in a category of its own. Likewise, the bodhisattva’s visual style—his hair bouffant,31 adornment and so on—are different from the Śrīvijayan sculptures. The hands and the Fears form an integral composition in the Swāt sculpture, such that the central figure could not be decoupled from the surrounding vignettes without disrupting coherence and meaning. Nonetheless, this may be exactly what has happened at some stage in the transmission of the eight-armed form.

30. Miyaji (2000: 1258) gives the impression that the Saddharmapuṇḍarīka is the locus classicus for the aṣṭamahābhaya proper and therefore the direct inspiration for the eight-armed iconography of the Swāt sculpture, but does not clarify how the twelve verses of Saddharmapuṇḍarīka 24.5–16 reduce to eight fears. Saddharmapuṇḍarīka 24 teaches that Avalokiteśvara saves from Firepits or Fire (agnikhada, 24.5); Water (jala, 24.6); Falling off a High Plateau (merutala, 24.7); Hurled Boulders (24.8); Armed Men (śatrugaṇair…śastrahastair, 24.9); Execution (vadhya, 24.10); Bondage (bandhana, 24.11); Spells and Spirits (mantra…bhūta, 24.12); Snakes etc. (nāga…, 24.13); Rapacious (‘sharp-toothed-nailed’, tīkṣṇadaṃṣṭranakha, 24.14) Beasts; and Lightning Strikes (jvalanārcis, 24.15–16). See ed. Dutt 1953: 294–295. Yoritomi 1990: 643 identifies seven Fears in this chapter. 31. The Swāt sculpture is depicted with ‘hair bunched like a vase’ kind, which, incidentally, the Kalparāja says practitioners can acquire (ākuñcitakuṇḍalakeśaś ca bhaviṣyanti; e.g. MS f. 32b27, via Suzuki, Ōtsuka and Kimura 2000: 22).

02_CreSouthV1_8P_4Feb22.indd 15

15

Fig. 2.2 : Eight-armed Avalokiteśvara pacifying the Great Fears (top left: Execution, Rapacious Beast, Firepit; top right: Rocks, Fetters). Stone, c. 7th century. Said to be from Swāt, Pakistan. Hand drawing by author based on Miyaji 2000: 1262, 1264.

There are remarkable visual similarities between the Swāt Avalokiteśvara and the Śrīvijayan type, similar enough to foster suspicion that miscopying might be involved. Firstly, one right hand of the Swāt Avalokiteśvara grips what appears to be a rope or line extending outwards. The line runs between two adjacent vignettes, those of the Swordsman and the Rapacious Beast.32 These two Fears are being appeased by separate hands: Avalokiteśvara stays the Sword and lifts the paws of the Beast. The line then may have been connected to a drowning man, as there is no depiction of the Fear of Water, which is included in most sets of the Fears. This line-holding hand is posed in the position corresponding to the rosary-holding fist of the Śrīvijayan type, not the noose-holding hand. Secondly, the lower left hand of the Swāt Avalokiteśvara lifts the door of a prison coop to release a devotee from the Fear of Bondage.33 The lifted door of the coop has a shape similar to the body of the waterpot or kendi held 32. Namely, the fear of a tīkṣṇa[-]daṃṣṭranakha beast (Saddharmapuṇḍarīka 24.14ab, ed. Dutt 1953: 2955), following Miyaji 2000: 1258. 33. Namely, the fear of haḍinigaḍa (Saddharmapuṇḍarīka 24.11ab, ed. Dutt 1953: 264 19), following Miyaji ibid.

3/2/22 6:16 PM

Iain Sinclair

16

in the corresponding hand of the Śrīvijayan icons. Both of these elements of the Swāt sculpture have visually similar but functionally different parallels in later depictions of Avalokiteśvaras with eight arms. A form of Avalokiteśvara who placates the Eight Fears is mentioned in the early literature on Amoghapāśa. This form is one of four to be engraved around the body of an enchanted vase, as described in the Kalparāja. The relevant passage gives just a minimum of detail: Amoghapāśa, the Noble Avalokiteśvara, is to be drawn facing east; southwards, the Eight Fears [form]; north, the Pensive; west, the All-Assuring [Ekādaśamukha]. All are adorned with all ornaments, are seated cross legged on lily pads.34

monly associated with the Eight Fears on the subcontinent has two arms, not eight. Sculptures of such two-armed Avalokiteśvaras abound in the rock-cut architecture of Maharashtra; they often incorporated portrayals of the Eight Fears. The prevalence and significance of these Avalokiteśvaras have been discussed by Osmund Bopearachchi (2014). The only known South Asian depiction of the Eight Fears that centres on an eight-armed figure is the Swāt Avalokiteśvara. Giving due consideration to the artistic corpus, it is doubtful that the Eight Fears Avalokiteśvara mentioned in the Kalparāja was envisaged as having eight arms.

The terse directions of the Sanskrit text have been unpacked here by following the more verbose Chinese translation.35 The Tibetan translation conveys the quite different understanding that the sides of the vase are engraved with (the bodhisattvas’) hand gestures.36 The forms of these four Avalokiteśvaras are not described in other chapters; their iconography may have been assumed knowledge. ‘Pensive’ Avalokiteśvaras, for instance, were produced in significant numbers, both on the subcontinent37 and in Maritime Southeast Asia.38 However, the form of Avalokiteśvara most com34. amoghapāśāryāvalokiteśvaraṃ citrayitavyaṃ pūrvāmukhaṃ (ed. pūrvvā mukhaṃ) dakṣiṇato aṣṭabhayam uttaraṃ cintāmayaṃ paścimaṃ samāśvāsakaraṃ sarve sarvālaṅkāravibhūṣitaṃ paryaṅke niṣaṇṇaṃ padmapatraiḥ (MS, f. 64a1–2, ed. Mikkyō Seiten Kenkyūkai 2001: 70). 35. 東面葉上。不空羂索觀世音菩薩[…]南面葉上[。]除 觀世音菩薩[…]西面葉上。十一面觀世音菩薩[…]北面 葉上。如意輪觀世音菩薩 (T 1092 288a 27–b5). Ellipses mark

passages elided here in order to facilitate comparison with the Sanskrit text cited in the previous note. 36. ’phags pa spyan ras gzigs dbang phyug bri bar bya ste/ shar phyogs su zhal bri/ lho ru mi ’jigs pa sbyin pa’i phyag/ nub tu dbyig ’byin gyi phyag/ byang du yid bzhin nor bu rgyan thams cad kyis brgyan pa skyil mo krung du bzhugs par bya’o (D 686 f. 1191–2). This passage corresponds to a part of the MS that is currently unedited. 37. Bautze-Picron 2004: 252–254, 284–285, figs. 25–28. For another instance of the Pensive form appearing in the retinue of Avalokiteśvara, see ibid.: 286, fig. 33. 38. Chutiwongs 1994; Park 2019.

02_CreSouthV1_8P_4Feb22.indd 16

Fig. 2.3: Two-armed Avalokiteśvara displaying the gesture of unfear and holding a lotus, surrounded by vignettes of rescue from the Eight Great Fears. Left, top to bottom: Firepit, Execution, Rocks, Water (truncated). Right: Rapacious Beast, Snake, Elephant, Bondage. Stone relief, Aurangabad, Maharashtra. Photograph (‘016 Door from Side’, https://www.flickr. com/​photos/anandajoti/34044313916/) by Anandajoti Bhikkhu, 2017, CC BY 2.0.

3/2/22 6:16 PM

From Melayu to Thamel and Back

17

In a further development of saviour-bodhisattva iconography, which can be raised only in passing here, the tableau of a two-armed figure flanked by scenes of the Fears began to feature Tārā rather than Avalokiteśvara as the central figure.39 Many depictions of Aṣṭamahābhaya-Tārā appeared on the East of the subcontinent instead of the West; they were first sighted around Magadha in the mid– late 7th century by Xuanzang 玄奘 and Yijing 義 淨. The worship of Tārā soon spread to Java, while gaining little traction in Śrīvijaya. Aṣṭamahābhaya-Tārā emerges as a saviour figure at about the same time as the eight-armed Amoghapāśa—but with a distinctly different geographic footprint, which has hardly been noticed as such before. 40 A degree of regionality can then be associated with maritime saviour figures, and this supports the proposition that the eight-armed Amoghapāśa emerged as a kind of local genius before spreading to other parts of Asia.

Tuoloni ji jing is a patchwork of translations from Indic material, including pre-existing translations into Chinese and some of the Amoghapāśa literature,42 as well as original writing. The section referring to the eight-armed Avalokiteśvara, first studied by R.H. van Gulik (1935: 74–75), sets out the ‘Maṇḍala for taking up the religion of the bodhisattva Hayagrīva’. Of interest here is the layout of the maṇḍala:

east asian sources: t 901, 959, 1169, mg.23079, fanxiang juan

This maṇḍala sets the stage for the summoning of the lotus-clan bodhisattvas. Their mantras are recited and foodstuffs and flowers are burned for them in the brazier at the entrance to the maṇḍala. The mantra of the Eight-armed Avalokiteśvara is likewise specified by Atikūṭa in connection with this process. However, the distinctive element of the mantra has no obvious Sanskrit parallel44 and

At the centre of this maṇḍala make a lotus seat seating an image of Horseneck Avalokiteśvara (Hayagrīva). One should make at the eastern doorway a lotus seat seating the Eleven-faced bodhisattva (Ekādaśamukha). You should make at the northern doorway a lotus seat seating the Eight-armed Avalokiteśvara. And the maṇḍala’s southern sector has no lotus seat; [there] make the eight serpent kings. […] The maṇḍala’s western gate is adjacent to the southern boundary. A lit brazier sits [there].43

Most of the early direct references to an Avalokiteśvara with eight arms occur in Chinese sources. They give little information on the bodhisattva’s form and function. At this stage the eight-armed Avalokiteśvara is encountered on the periphery of the maṇḍalas centred on other objects of worship, as in the abovementioned passage from *Dhāraṇīsaṃgraha, which are all Nepalese anthologies, as the Kalparāja on the making of an enchanted vase. far as is known at present. On the Tuoluoni ji jing’s date The bodhisattva is not yet an independent figure. of compilation, see Shinohara 2010: 393; on its ‘general maṇḍala’, see Tanaka 2019: 43–44. There are also few indications as to the place of 42. Shinohara 2010: 404–405. A paṭavidhi from an origin, though all such indications, discussed below, ‘Amoghapāśa scripture’ (Bukong juansuo jing 不空羂索 are compatible with the hypothesis that the eight- 經), which has not yet been identified, and which Atikūṭa armed form developed in Southeast Asia rather seems to have translated himself, is cited in the Tuoloni ji jing at T 901 828a –c . See also Reis-Habito 1999: 67. than India. 43. 其壇中心作蓮華座。安置馬頭觀世音像。正當東門作蓮 An early reference to an eight-armed form 華座。 安十一面菩薩。正當北門作蓮華座。安八臂觀世音。其 occurs in Atikūṭa’s Tuoloni ji jing 陀羅尼集經 壇南方更無華座。作八龍王。[…]其壇西門如近南畔。安一火 (*Dhāraṇīsamuccaya), a compendium of invoca鑪。(T 901 838a , b ). The translation is mine. tion procedures compiled in China in 654.41 The 17

24–27

39. Bautze-Picron 2004: 239ff. 40. Yoritomi 1990: 644ff. discusses early sources for the Aṣṭamahābhaya-Tārā. 41. The Tuoluoni ji jing, T 901, has no known relation with Sanskrit manuscripts titled *Dhāraṇīsamuccaya or

02_CreSouthV1_8P_4Feb22.indd 17

28

7

44. 唵一 夜勢夜輸擔二 莎訶三 (T 901 838b12). This can be transcribed as om̐ jiaᴴ ɕiᴇiᴴ jiaᴴ ɕɨo t m svāhā, with unidentified phonemes given in reconstructed Middle Chinese. The character ɕɨo 輸 appears to convey the Sanskrit phoneme śu, as in the word *viśuddhi 毘輸提, which Atikūṭa uses four times. The character *ʈˠiᴇp̚ 輒 (Mandarin zhé) is used instead of 輸 in the version of the ritual given in the derivative handbook (discussed below) T 1074 172a18 .

3/2/22 6:16 PM

Iain Sinclair

18

is currently attested only in the Tuoluoni ji jing and its derived works.45 The obscurity of this mantra in the received Buddhist corpus could be an indication of origin in a relatively marginal culture, but it is hard to read much into this, given that so much of the early corpus has been lost. As the Tuoluoni ji jing provides the first definite notice of the eight-armed Avalokiteśvara, the provenance of the text calls for closer investigation. While more work on this is needed, it may be worth mentioning here that the Tuoluoni ji jing is one of the few tantric Buddhist texts to refer to tin, a metal that has long been associated with the Malay Peninsula.46 Tin is specified, in a passage worth quoting here in full, as one of five metals to be combined in the manufacture of vajras: If you want to make a vajra, the first thing is to select metals of the five varieties. Do not use any melted down utensils. What are the names of the five? One: gold, two: silver, three: copper, four: iron, five: tin. [They can be] alloyed to make a vajra shape. If none of these five types [are available], lightning[-stricken] jujube heart[wood] can also be used. Before the vajra has been made, mantrify the material of gold, copper etc. 108 times, saying: *om̐ mahākāla nātha kubera svāhā.47

High levels of tin in bronze have been regarded as indicators of Southeast Asian provenance.48 Certain Buddhist artefacts found in Sumatra are The character 夜 consistently transliterates ya or yā in T 901. Restorations such as *om̐ yasya śuddhaṃ svāhā, *om̐ yat satyaṃ° or *om̐ yakṣāya° could be conjectured. Van Gulik 1935: 75 made no attempt to transcribe the mantra. 45. In particular, the mantra has no identifiable counterpart in the large mantric repertoire of the Kalparāja, as collated by the Mikkyō Seiten Kenkyūkai 2004. 46. See e.g. Colless 1969: 22ff. 47. 若人欲作跋折囉者。先取金等五色之物。皆未曾經作器 用者。何名五色。金銀赤銅鑌鐵錫。合和爲作跋折囉形。若 無此五種。可用霹靂棗心。亦得。且未作其跋折囉前。先 須豫呪金銅等物一百八遍呪曰 唵一 摩訶迦囉二 那吒倶鉢 囉三 莎訶四 (T 901 803c18–23).

48. Woodward 2003: 106–107. The question of where South Indian bronzes sourced their tin is discussed by Srinivasan 2013: 168. On period Javanese bronzes with high tin content, see Mechling et al. 2018: 83ff.

02_CreSouthV1_8P_4Feb22.indd 18

rich in tin; at least one icon of the eight-armed form was cast in an alloy with a very high concentration of tin solute.49 Avalokiteśvaras with eight arms are a common sight in the art of Tang Buddhism, notwithstanding the lack of detailed information on these figures in the literature of the period. Dozens of independent depictions are extant, especially in Nara and Heian sculpture, in the related iconographic compendia, and cloth and mural paintings from Dunhuang.50 Hardly anything is known about why this form attracted such a high degree of acceptance and enthusiasm. In the beginning, demand for icons of the eight-armed form would have been driven by the praxis tradition of the Tuoluoni ji jing. Its Hayagrīvamaṇḍala was at the centre of an independent ritual.51 Towards the end of the 8th century, an eight-armed Avalokiteśvara is mentioned52 in an initiation manual, the Dinglunwang da mantuluo guanding yigui 頂輪王大曼荼羅灌頂儀軌, compiled by the monk Bianhong 弘. Once again, this form of Avalokiteśvara is named as a figure on the periphery of a maṇḍala. As Bianhong had travelled to Tang China from Java (Heling53 訶陵, *Kliṅ), it might be wondered whether he had encountered the eight-armed form in his homeland. However, it is more likely that Bianhong took this figure from the Tuoluoni ji jing, which he is known to have drawn on in writing his manual.54 49. Ery Soedewo 2012: 157. 50. On the eight-armed Amoghapāśa icons in Japan, see, among others, Meisezahl 1962: 265–266, Reis-Habito 1999: 53, and Lokesh Chandra 1999: 300–301, and on those in Dunhuang, Nakamura 2008. See also Wong 2015. 51. The Heyejielipo Guanshiyin pusa shoufa tan 何耶掲 唎婆觀世音菩薩受法壇, T 1074, is a separate treatise on entering the Hayagrīvamaṇḍala described in the Tuoluoni ji jing. Its contents are included in the Japanese Gyōrinshō 行林抄, T 2409, compiled in 1158. On other reuses of the Tuoluoni ji jing, see e.g. Reis-Habito 1999: 45. For references to the contemporaneous praxis tradition of the Tuoluoni ji jing, see Sinclair 2016a: 39. 52. Sinclair 2016a: 53, no. 45. The reading ba bu Guanyin 八臂觀音is a conjectural emendation of T’s doubtful reading ba sheng Guanyin 八聖觀音 (T 959 328b14). 53. Some Middle Chinese reconstructions take the first character of 訶陵 to be pronounced with a velar rather than a glottal fricative: xalɨŋ. Rather than reading Heling (e.g. Wade 2014: 31), therefore, it is possible to perceive the underlying word as *Kaliṅga, Kliṅ. 54. Sinclair 2016a: 38, 39, n. 55.

3/2/22 6:16 PM

From Melayu to Thamel and Back The East Asian depictions display a great deal of iconographic variation, which again points to a chronic lack of an associated textual standard. The forms of some eight-armed Amoghapāśa icons in Japan were clearly copied with the mediation of model books, not written descriptions.55 East Asian depictions with two opposing hands holding identical implements56 are unlikely to have been prescribed in texts originating in South Asia. Some depictions of the bodhisattva alternate visually similar implements, such as trident and tinkle-staff (khikkhirika). In general, such iconographic irregularities can be explained as artefacts of a primarily visual mode of transmission. They are typical of image cults that disseminate with little associated documentation. Their many variations57 are in effect routine occurrences. Informal variations of this kind do not need to be recognized as distinct formal traditions or, as Shunsho Manabe (2016: 294) contends, a ‘deep mystery’. The coarse linguistic environment in which some period artisans worked can be glimpsed in a Dunhuang banner now kept in the Musée Guimet, MG.23079. It was painted no later than 1010, according to its dedication, which has recently been studied by Henrik H. Sørensen (2020: 110–113). This banner records the first definite, if semi-literate, identification of the eight-armed form as Amoghapāśa. Its accompanying caption reads namo Bokong juansuo pusa 南无伯空卷索菩薩 in Mandarin. It contains a homophonic misspelling of Amoghapāśa’s usual Chinese name Bukong juansuo 不空羂索. The painting’s main subject resembles the Śrīvijayan type in its general appearance and in features such as the characteristic noose, waterpot, and antelope-pelt shawl. However, it idiosyncratically holds aloft two tridents. Its sponsor hoped that the painting would be beneficial for his ancestors, and ‘above all’ that ‘the people 55. On the resemblance between sketches in certain Japanese iconographic compendia and various seated eightarmed Amoghapāśa icons produced before the end of the 12th century, see Taniguchi 2002. 56. Examples are the mural of Dunhuang Cave 14 (Nakamura 2008: 227, fig. 1) and the Dunhuang banner MG.23079, discussed below. 57. Nakamura 2008: 226 itemizes the attributes of six- and eight-armed forms at Dunhuang.

02_CreSouthV1_8P_4Feb22.indd 19

19

of the country would be at peace’.58 The premise of protecting a nation conceptually extends the notion of Avalokiteśvara as a rescuer of individuals, and comes up in connection with the eight-armed form in other contexts, examined below. All East Asian images of or references to the eight-armed Avalokiteśvara can, in short, be regarded as descendants of the Śrīvijayan type, including the wide variety of typologically different figures.59 A certain proximity to the Indo-Malay world is visible in the Yunnanese Scroll of Indic Icons (Fanxiang juan 梵像卷), which features one of the last premodern East Asian depictions of the Śrīvijayan type. The Scroll, painted between the years 1172 and 117660 in the Dali kingdom, includes in its large pantheon a standing Avalokiteśvara with eight arms (frame 95)61 that bears a strong resemblance to Śrīvijayan iconography and style. The bodhisattva is portrayed in the Scroll with the expected variation in the array of arms, which again seems to involve graphic confusion. The bunch of grass held in the upper right hand could have been a hook or noose in the artist’s model. The forked branch held in the upper left hand of the figure in the Scroll, not seen in most other East Asian depictions, looks much like the tridaṇḍa codified at a late stage in South Asia.62 However, it is hard to see how a direct model for this depiction could have come from the subcontinent to Dali at this time.63 The Scroll’s eight-armed Avalokiteśvara is probably informed by much earlier contact with the Malay world, which is noticeable in Yunnan from the 9th century onwards.64 58. 先奉為國人安… (Sørensen 2020: 111). 59. Miyaji 2000: 1256. The closest non-Chinese counterpart of the Dunhuang images is not the Swat sculpture discussed by Miyaji but the Nālandā sculpture discussed in Bautze-Picron ibid. 60. Li 2010: 114. 61. Soper and Chapin 1970, pl. 35. 62. See below, p. 20ff. on the Śākyaśrī/Thaṃ Bahī tradition. 63. Megan Bryson (2018) argues that Dali ‘elites’ found it expedient, for the sake of asserting independence from China’s ruling dynasties, to give their Buddhism an ‘Indian’ pedigree. 64. Śrīvijaya has been identified as a stylistic influence on Buddhist images in the Nanzhao 南詔 kingdom, the forerunner of Dali, by de Mallmann (ibid.) and thence Soper

3/2/22 6:16 PM

Iain Sinclair

20

A form of Amoghapāśa with eight arms and four faces—an iconographically divergent but apparently related type—is described in one Chinese text. A description is given in the painting instructions of a manual on the invocation of the goddess Cundā, the Chiming zang yuqie dajiao Zunna pusa daming chengjiu yigui jing 持明藏瑜伽大教尊那菩薩大明成 就儀軌經 (*Vidyādharapiṭaka-mahāyogatantra[nibaddha-]Cundāmahāvidyā-sādhanavidhi). This manual, translated in 994, is loosely connected to the Māyājālamahātantra65 and thereby to soteriological Tantrism. However, its painting procedure is in line with the image magic of the Amoghapāśa literature: On the left side [of six-armed Cundā], on a lotus, draw Amoghapāśa, four-faced, eightarmed. He has a tiger pelt for a garment, bound around his waist. He has a deerskin for a divine garment. [Each] face has three eyes. The locks are piled up in a jewelled crown, hanging down. The first right hand makes the sign of giving; the second hand holds a rosary; the third hand holds a noose; the fourth hand makes the sign of fearlessness. The first hand holds a white lotus; the second hand holds a scripture; the third hand makes a fist, raising the index finger to make the threatening sign; the fourth hand and Chapin 1970: 36–37. Similarly, Guy 1995: 78–79 and Bryson 2018: 94, n. 32 include Campā among the probable sources of Nanzhao art, particularly the *Ajaya (Acuoye 阿嵯耶) Avalokiteśvara. 65. This little-studied text, extant only in Chinese translation, claims to be part of the great yoga doctrine (*mahāyogatantra) of Mahāvairocana tathāgata 大毘盧 遮那如來瑜伽大教 (T 1169 677b25, 684c 9). Tanaka (2016b: 46ff.) noticed that the text contains material on Cundā’s nine mantradevatās (om̐ ca le cu le cuṃ de svā hā— corresponding to, respectively, Vairocana, Cakravartin, Acala, Cintāmaṇi-Avalokiteśvara, Amoghapāśa, Cundā, Vajrapāṇi, Ekajaṭā, Vajranakhā) in which is also given in Ānandagarbha’s ṭīkā (D 2513 209ff.) on the Māyājālamahātantra. Tanaka adds that T 1169 was translated in the same year (on which see Takeuchi 1976: 45–46) as the Chinese translation of the Māyājālamahātantra itself, T 890, indicating that the two texts could have been transmitted together. It can be added here that the nine mantradevatās are discussed in another practical work on the system of the Māyājālamahātantra, a Śrīmañjuvajrodayamaṇḍalavidhi of an unidentified author (D 2590, 237b).

02_CreSouthV1_8P_4Feb22.indd 20

holds a hook. A blaze of light shines over the whole body, shimmering gloriously.66

The hand array of this four-faced, eight-armed form is much the same as that of the single-faced Śrīvijayan icons, with the exception of the threatening gesture of the third left hand, corresponding to the waterpot. This variation could have originated in a visual model in which the waterpot was missing or malformed (e.g. 7). The four-faced Amoghapāśa of this sādhanavidhi may then be drawing on knowledge of the Śrīvijayan single-faced, eightarmed form.67 However, these considerations lie outside the present scope. The four-faced, eightarmed Amoghapāśa is an obscure figure, not yet identified in the art of the Archipelago or India, although there are Khmer depictions of a similar iconography.68

the khmer eight-armed avalokiteśvaras and the kāraṇḍavyūha Several Khmer depictions of eight-armed Avalokiteśvaras were produced between the late 10th century and the 13th century. They are later than the Śrīvijayan icons, and appear to have been modelled on them. At least one Śrīvijayan bronze has been ‘found on Khmer-controlled territory’ 66. 左邊蓮華上。畫不空羂索。四面八臂。以虎皮爲衣復爲 絡腰†。復以鹿皮爲天衣。面有三目[。]頂戴寶冠髮髻下垂。右 第一手作施願印。第二手持數珠。第三手執羂索。第四手作 施無畏印。左第一手持白蓮華。第二手持經。第三手作拳。竪 立頭指作期剋印。第四手執鉤。光炎遍身照耀熾盛 (T 1169 685a 6–12). †T: 腋

67. The elaborate eighteen-armed form of Cundā described in this manual (T 1169 679a12–23) is known to have been depicted not only in India (see e.g. Paul 1987, pl. 78), but also in Java. See e.g. the cover of Heine-Geldern 1925, described on p. 21, pl. 12, and Asian Art Museum (San Francisco) 2006.48.a–c. 68. An eight-armed Khmer bronze reproduced in Seki 1996: 27 AJ-16 bears a miniature Amitābha on top of its four heads, marking it as an Avalokiteśvara, albeit one that bears a different (if hard to discern) hand array. T 1169 does not require Amoghapāśa to bear an Amitābha image in the crown or to be portrayed standing up; this opens up to investigation images that likewise lack these attributes. An eight-armed, three(?)-headed ‘Mahāvajranātha’ form is known in Nepal (Bhattacharyya 1958: 428).

3/2/22 6:16 PM

From Melayu to Thamel and Back (Bunker and Latchford 2011:193, 14) in good condition. There are trivial iconographic differences between the Śrīvijayan and Khmer depictions, such as alternations of similar-looking objects, which are typical of the variation seen elsewhere. On the whole, the Khmer icons represent a dead end in the transmission of the form, in that they are not yet known to have impacted on later visual traditions. However, they may help to shed light on the early conception of the bodhisattva, which the Khmers would have inherited, according to the present hypothesis, from their Malay neighbours. Many Khmer depictions are associated beyond doubt with the Kāraṇḍavyūha. Even so, the Kāraṇḍavyūha is not easily identifiable as a source for earlier Śrīvijayan depictions of the eightarmed Avalokiteśvara. The text does not provide exact iconographic information; its narrative envisions the bodhisattva taking on different forms in various scenarios;69 and the vagaries of its textual transmission—including the unique recension unearthed in Gilgit70—are still being worked out. The Sanskrit text has still not been satisfactorily edited,71 long after the first attempt of Bhaṭṭācāryya (1872); some of the variation in the manuscript tradition is of potential iconographic significance.72 It is not known whether the Kāraṇḍavyūha reached the Malay world in any form,73 but its narratives were

69. Sakuma 2019: 76ff. 70. Mette 1997 has edited these fragments. Facsimiles of the fragments were published by Mette et al. 2017. 71. The most recent edition is that of Buddhadev Bhattacharya (2016). Some of its misreadings are shared with the e-text of the Kāraṇḍavyūha hosted at GRETIL (http://gretil. sub.uni-goettingen.de/gretil/1_sanskr/4_rellit/buddh/ bsu019_u.htm, last accessed 1 March 2020). For instance, the typo meheśvarāya in the e-text of Kāraṇḍavyūha 2.9, which is not in GRETIL’s printed source (Vaidya 1961: 30412), is also in Bhattacharya (2016: 11119). 72. Sinclair 2015: 436, n. 6. 73. In the view of Nihom, ‘we have ample indication that this work was available in the Archipelago’ (1994: 16). Yet it is not clear whether the Kāraṇḍavyūha is uniquely and definitely indicated by Nihom’s sources—primarily, the Old Javanese Kuñjarakarṇa. The same goes for mantra phrases in the Javanese corpus that are not unique to the Kāraṇḍavyūha. See also Woodward 2007, n. 28.

02_CreSouthV1_8P_4Feb22.indd 21

21

definitely known in the Khmer Empire by the end of the first millennium.74 The various types of Khmer eight-armed Avalokiteśvaras will be briefly surveyed here. At least four images are sculpted on steles that appear to have served as boundary markers, sīmā. In these objects Avalokiteśvara embodies a protective force at the monastic perimeter. The idea that the bodhisattva energizes a sīmā with his mantra is duly articulated in the Hṛdaya,75 albeit in the narrower context of healing magic. A stele in the Walters Art Museum, Accession No. 25.194, has invocations engraved on its reverse side including the signature ṣaḍakṣarīmahāvidyā of Avalokiteśvara revealed in the Kāraṇḍavyūha,76 om̐ maṇipadme hūm̐ . The function of the stele as a spatial marker is underscored by its opening mantra directed to Indra, that is, to the eastern direction.77 A different Khmer stele featuring the eight-armed form, belonging to the collection of the Bangkok National Museum, has five figures sculpted upon it. Hiram Woodward (2015: 223–226) understands them to have comprised a ‘maṇḍalic schema’ arranged around a centre. Such boundary arrangements may be visible in clay sealings and moulds that depict temples—mostly Hevajra temples—of the Khmer

74. Miyazaki 2010: 32–33 in this regard discusses allusions to the Kāraṇḍavyūha’s soteriology in the 970 CE Chikreng inscription, K.417. See also Green 2014: 91ff. 75. Meisezahl 1962: 298. 76. Pou 2002: 129 gives a transcription and translation of the stele, inventory number K. 1154. Woodward 2007: 72–73 first drew attention to the inscription’s relation to the Kāraṇḍavyūha. See also Miyazaki 2019: 356–357. Although this mantra is uncontroversially associated with the Kāraṇḍavyūha, it may be worth noting that the Kalparāja sets out many soundalike invocations of the form om̐ … padme including the mantra om̐ maṇipadme, in one recension (Mikkyō Seiten Kenkyūkai 2004: 142, No. 326). 77. Pou (ibid.) reads om̐ kṣāhā indrāya svāha. From the facsimile provided at The Walters’ website (http://art. thewalters.org/detail/35298/stele-with-eight-armed-avalokiteshvara), the second and third syllables look to have anusvāras and may be bījākṣaras reading as follows: om̐ kṣāṃ hoṃ indraya (sic) svāhā. No such mantra has been identified in textual sources.

3/2/22 6:16 PM

22

Iain Sinclair

Empire.78 In these depictions, the eight-armed Avalokiteśvara occupies the middle of the lower section, corresponding to centre of the sīmā of the temple compound. Khmer portrayals of the eight-armed Avalokiteśvara show the lower left hand outstretched instead of the waterpot-holding hand of the Śrīvijayan icons. The ‘giving’ gesture is displayed by a hand on each side, which could be seen as an apparent redundancy. Yet, as Woodward has noticed,79 this double-handed iconography is most apt for the spectacle of Kāraṇḍavyūha 1.3, in which Avalokiteśvara ‘emits ten subterranean rivers from his ten fingers’80 to relieve thirsty hell-beings. Late Khmer interest in this episode is demonstrated by its elaborate depiction in the reliefs at Banteay Chhmar.81 The Khmer icons also wield a vajra, a weapon that is not normally associated with Avalokiteśvara and does not feature in his narratives in the Kāraṇḍavyūha. The vajra in these icons, if it replaces the trident of the earlier visual tradition, would be seem to be an ad hoc attempt to interpret an ambiguous received iconography. Two eight-armed Avalokiteśvaras are depicted in the Kāraṇḍavyūha narrative reliefs carved for the Banteay Chhmar temple complex in the late 12th century.82 One of the reliefs has been convincingly identified by Jean Boisselier83 with the episode in which Maheśvara and Umā ask Avalokiteśvara for assurances that they will be reborn as tathāgatas.84 The relief portrays 78. One such ‘temple seal’ with five eight-armed figures in the bottom section—probably representing a complete sīmā and an actual architecture—is reproduced in Bunker and Latchford 2011: 385, fig. 9.20. Another seal, National Museum of Cambodia Ga 5657, displays a single eightarmed figure in the lower centre (Conti 2014: 389, fig. 15). See also Peter Sharrock’s chapter in the present volume. 79. Woodward 2015: 222. 80. daśabhyo hastāṅgulībhyo daśa vaitaraṇīr niṣkrāmayati (ed. Bhattacharya 2016: 51 10). For a full translation, see Roberts and Yeshi 2013–2016 #118. See also Bhattacharya 2001: 23. 81. Boisselier 2008: 308–309. 82. Sharrock 2007: 69–78. 83. Boisselier 1965: 77–78, 2008: 317–320. See also Miyazaki 2019: 360. 84. See Bhaṭṭācāryya 1872: 888 –913 , Vaidya 1961: 30311–304 and Bhattacharya 2016: 11011–111 for Sanskrit texts of the

02_CreSouthV1_8P_4Feb22.indd 22

in particular the dialogue between Umā and Avalokiteśvara. Maheśvara, having just received his prophecy of enlightenment, is depicted in the form of Sadāśiva85 at the bodhisattva’s lower left side. Umā’s salutation comprises eight distinct epithets, which could be correlated with Avalokiteśvara’s eight hands: Homage to Avalokiteśvara, to the great lord (maheśvarāya), life-giver (prāṇaṃdadāya), having the best eye on Earth in hand (pṛthivīvaralocanakarāya), having the glory of a fair lotus (śubhapadmaśriye), surrounded [by followers86] (parivṛtāya), set out [towards] nirvāṇa [or the bodhisattva] stage[s] (nirvāṇabhūmisaṃprasthitāya), positive mind in hand (sucetanakarāya), holder of dharma (dharmadharāya).87

Some of these praises relate straightforwardly to the regular iconography: life-giver (prāṇaṃdada), i.e. varada; lotus (śubhapadmaśrī); book (dharmadhara). Furthermore, the sucetanakara could be the abhaya gesture; the maheśvara is recognizable by his trident; the characteristic noose or hook can be a means of nirvāṇabhūmisaṃprasthita. The remaining two epithets (pṛthivīvaralocanakara, parivṛta) have no obvious iconographic counterparts, but could correspond to the sun/disc88 and Maheśvaranirvyūha of the Kāraṇḍavyūha, and Roberts and Yeshi (2013–2016) 2.93–96 for an English translation. Note that the content of the Maheśvaranirvyūha is not clearly demarcated, in the recensions published to date, from the first half of the same prakaraṇa, 2.7, which describes Avalokiteśvara’s pores. 85. Bunker 2011 proposes that the five-faced, ten-armed Khmer icons of Maheśvara generally represent Sadāśiva. 86. In the Kāraṇḍavyūha the word parivṛta refers to a retinue of followers: bhāryāputraduhitṛbhiḥ parivṛtaḥ (2.6, ed. Bhattacharya 2016: 104 10), dārakadārikāparivṛtā (2.8, ed. Bhattacharya 2016: 11520). 87. See ed. Bhaṭṭācāryya 1872: 9011–13; Vaidya 1961: 304 12–13; Bhattacharya 2016: 11119–20. 88. There is no counterpart for Umā’s praises in the Chinese translation of the Kāraṇḍavyūha, but the same epithet occurs in the Avīciśoṣaṇa chapter, for which the Chinese understands pṛthivīvararocanakarāya 大地爲眼 明踰日 (T 1050 53a 18) rather than pṛthivīvaralocanakarāya. However, the latter agrees with Tib. sa’i spyan mchog mdzad pa (D 116 242b3).

3/2/22 6:16 PM

From Melayu to Thamel and Back rosary carried by Avalokiteśvara in the relief. The architects of Banteay Chhmar in any case chose the eight-armed form, out of many forms in the local pantheon, to represent the moment at which Umā meets Avalokiteśvara, greets him with eight sobriquets and is assured of her future Buddhahood. As the Khmers must have understood that the bodhisattva’s form is related to his soteriological function—certainly an old idea89—the form with eight arms could be associated more generally with the winning over of non-Buddhists. The second Banteay Chhmar relief that portrays the eight-armed Avalokiteśvara has not had its corresponding narrative identified yet, but if it depicts the asura king Balin90 then it conveys a similar theme. Regrettably, neither relief is now available for direct study following ‘the largest looting operation of Cambodian cultural property in recent history’, as Keiko Miura (2016: 35–38) has called it.91 The final form assumed by Avalokiteśvara in the Khmer Empire is known as Jayabuddhamahānātha (Fig. 2.4). The iconography of Jayabuddhamahānātha fuses three elements: the Khmer eight-armed type, the cosmic imagery of the Kāraṇḍavyūha, and the physiognomy of king Jayavarman VII (r. c. 1182–1220). The eight-armed base form is much the same as earlier Khmer depictions. The visual element from the Kāraṇḍavyūha is the suffusion of the body with small figures representing the myriad bodhisattvas, pratyekabuddhas and

89. Apart from Avalokiteśvara’s shapeshifting power, set out in the Saddharmapuṇḍarīka, the notion of appearing in different forms for different gati was being taught by the 6th century (Reis-Habito 1999: 41). Similarly, there is the conception of the six-armed Avalokiteśvara as the saviour of the entire ṣaḍgati (Miyaji 2000: 1261, fig. 9). 90. One reason for supposing that this relief depicts the meeting between the warrior Balin and Avalokiteśvara (Kāraṇḍavyūha 1.11) is that the figure to the right of Avalokiteśvara holds a prominent weapon. This episode involves another ‘winning hearts and minds’ encounter and the recital of a stotra containing similar epithets. Before Balin meets with Avalokiteśvara, he is trapped by a multi-armed form of Nārāyaṇa, portrayed as eight-armed in one Nepalese illuminated manuscript (Sakuma 2007: 179, fig. 3). 91. Sharrock 2015: 77 provides visual identification of the looted reliefs.

02_CreSouthV1_8P_4Feb22.indd 23

23

buddhas residing in Avalokiteśvara’s pores.92 As for the facial features of these statues, they resemble the presumed portrait statues of Jayavarman VII.93 This hybrid figure has been identified by Woodward with the Jayabuddhamahānātha icons that Jayavarman VII had installed in twenty-five locations named in the Preah Khan inscription (K. 908).94 This identification has been questioned,95 but there are no better explanations for the Khmer buddha-permeated Avalokiteśvara statues. Several of these statues were found at the particular spots of Jayabuddhamahānātha installations specified in the Preah Khan inscription.96 Jayavarman VII had 92. Previously the visual element of suffusion had been seen as Buddhas ‘radiating’. This element instead represents the Buddha-filled pores of the bodhisattva spoken of in the first part of Kāraṇḍavyūha 2.7, trans. Roberts and Yeshi (2013–2016 #2.80–89). 93. National Museum of Cambodia Ka.1703 (for bibliography, see e.g. Dalsheimer 2001: 157–160) is one of the quintessential portrait statues of Jayavarman VII. All such portrait statues have missing arms but are assumed to have posed in a seated meditation posture. 94. For recent identifications of the locations named in connection with Jayabuddhanātha in K. 908 vv. 104–121, see Wyatt 2001: 15–17. To explain K. 908 first mentioning twenty-three installation sites (trayoviṃśatideśeṣv […] atiṣṭhipat, v. 120cd), then twenty-five (pañcaviṃśatideśakāḥ, v. 159b)—as noticed by Jacques 2005: 14—Maxwell 2007: 95 opines that ‘two more [sites] were set up while work on the inscription proceeded.’ 95. The elements of the name Jayabuddhamahānātha have been discussed by Woodward and Douglas 1994–1995: 105 and Maxwell 2007: 82. With regard to the -buddha- element, Jacques (2005: 13) objects: ‘La difficulté ici est qu’il n’est pas possible de confondre Buddha et Avalokiteśvara’. In the Kāraṇḍavyūha, of course, Avalokiteśvara incorporates Buddha bodies in a typical example of late Mahāyāna recursive imagery. And Avalokiteśvara icons themselves include the figure of the Buddha Amitābha (explicitly labelled as such in the case of the Candi Jago pantheon, discussed below). 96. The findspots of the buddha-pore-filled Avalokiteśvaras include Phetchaburi i.e. Jayavajrapurī (K. 908 v. 117a, cf. Woodward and Douglas 1994–1995: 108, 110, n. 12 #15); Ratchaburi i.e. Jayarājapurī (K. 908 v.116c, cf. Woodward 1994–95: 107, 109, n. 12 #11; Miyazaki 2010: Map 1-2); Prasat Muang Singh in Kanchanaburi (Miyazaki 2010: Map 1-1), identified with Jayasiṃhapurī (K. 908 v. 116d; cf. Woodward 1994–95: 105; Wyatt 2001: 15, n. 23); and Preah Khan itself (Miyazaki 2010: Map 1-10), where inscription K. 908 was installed. For additional

3/2/22 6:16 PM

24

Iain Sinclair these images installed in state-supported temples to effect an ‘exact reversal of the kind of Hinduisation of Buddha images that used to take place before Jayavarman’s reign’, in the view of Thomas Maxwell (2007: 83). By the time the Jayabuddhamahānātha sculptures were being produced, no new images of the eight-armed Avalokiteśvara appear to have been made in the Malay Archipelago for almost two centuries. If this form had once enjoyed the support of Śrīvijaya’s rulers, it might likewise have been pushed out of favour by some high-level disruption. The 1025 Cōḻa invasion could well be the event that put an end to the production and veneration of eight-armed Avalokiteśvaras among the Malays. This invasion had long-term effects on the region, including its sudden and permanent removal from certain Buddhist pilgrimage circuits.97 Little can be said about how much of the bodhisattva’s Śrīvijayan cult carried over to the Khmer milieu. With the destruction of Mahāyāna Buddhism in Thailand and Cambodia after the death of Jayavarman VIII (r. 1243–1295),98 the eight-armed form reached its end of life in the Angkor Empire. Nonetheless, by this time it had already been revived in the Malay Archipelago through a transmission coming back from South Asia.

śākyaśrībhadra, vibhūticandra and thamel

Fig. 2.4: Jayabuddhamahānātha, ehe Eight-armed Amoghapāśa with the face of Jayavarman VII and buddha-filled pores over the torso. Sandstone, Bangkok National Museum. Findspot: Lopburi. Photograph (‘050 Avalokitesvara’, https://www. photodharma.net/Thailand/​National-Museum/ images/050-Avalokitesvara-13c-Lopburi-Original.jpg) by Anandajyoti Bhikkhu, CC BY 2.0.

02_CreSouthV1_8P_4Feb22.indd 24

The first sightings of the eight-armed Avalokiteśvara on the subcontinent occur almost six centuries after the advent of the form in Maritime Southeast Asia. They are conveyed by Śākyaśrī, also known as Śākyaśrībhadra (d. 1225), and Vibhūticandra (fl. 1204–1248). These two paṇḍitas fled the dediscussion see Woodward (2003: 211–213). Many sites named in K. 908 remain unidentified (Wyatt 2001: 15, n. 24), leaving ample room for further identifications of findspots and installation sites. Note also that not all Jayabuddhamahānāthas are of the ‘buddha-permeated’ type. 97. On the disappearance of Śrīvijaya and Kedah from sites portrayed in South Asian illuminated manuscripts, see Peter Sharrock’s contribution in the present volume. On other such effects, see Sen 2009: 68–69. 98. For some background, see Woodward 2003: 204–205.

3/2/22 6:16 PM

From Melayu to Thamel and Back struction of Vikramaśīla-mahāvihāra in Eastern India and escaped to Tibet in 1204. Śākyaśrībhadra describes the eight-armed form in two original works: a full-length sādhana edited by Vibhūticandra and an abbreviated sādhana. Vibhūticandra also composed an independent visualization handbook. These three works are extant only in Tibetan translation—Q 4840, 3682 and 4841, respectively.99 The Tibetan translations of Śākyaśrī’s and Vibhūticandra’s sādhanas provide vital information on the movement of the eight-armed Amoghapāśa into South Asia. They are associated in particular with Thaṃ Bahī, the monastic complex that gives present-day Thamel in Kathmandu its name.100 Thaṃ Bahī incorporates the Nepalese branch of Vikramaśīla founded by Atiśa (982–1055) in 1042.101 It often served as a base for non-Newar monks, like other monasteries of the bahī class. The colophon of the translation of Vibhūticandra’s sādhana records that it was completed at Thaṃ Bahī; Śākyaśrī’s long sādhana was most likely translated there as well.102 99. Śākyaśrībhadra’s ’Phags pa don yod zhags pa’i sgrub thabs (*Amoghapāśalokeśvarasādhana), Q 4840, has been conveyed as an ‘abstract in English’ by Pal (1966: 235) and translated in full by Meisezahl (1967: 477–481) together with the iconographic section of Vibhūticandra’s sādhana of the same title, Q 4841 (ibid.: 482). Meisezahl did not notice the shorter work attributed to Śākyaśrī, translated into Tibetan with the title Don yod zhags pa’i sgrub thabs mdor bsdus pa, Q 3682 / D 2861. 100. Thaṃ Bahī comprises both Dharmadhātu-vihāra, a royal monastery predating the mid-11th century, and the local branch of Vikramaśīla-mahāvihāra. The word bahī was not originally synonymous with vihāra; it designates an ‘outside(r)’ (bahis) institution (Sinclair 2016b: 49). The Newar name of the complex (cognate with *uttama-bahis) indicates its position relative to old Kathmandu (ibid.: 136). 101. As the name ‘Atiśa’ is used consistently, with this spelling, in Tibetan sources—which provide most of the primary sources for his life—this will be the spelling adopted here. On the Sanskrit alternative *Adhīśa, see Sinclair (2016b: 54, n. 54). 102. The colophon to the translation of Vibhūticandra’s sādhana states: […] dpal bhi bu ti tsandra’i zhal snga nas/ blo(glo) bo loccha ba dge slong shes rab rin chen gyis bal yul thang gi bhi har gyi gtsug lag khang du […] (Q 4841, Dpe bsdur ma ed. 624–5). Vibhūticandra and his collaborator, the Mustang Translator (NB: the correction of blo bo to glo bo, conjectured by Schoterman 1994: 158, is supported by the Narthang edition of the text), worked together on

02_CreSouthV1_8P_4Feb22.indd 25

25

Śākyaśrī’s descriptions conform to the Śrīvijayan type in almost all details. He says that the bodhisattva has one face and eight arms, holds a rosary and noose, makes unfearing and giving gestures with the four right arms, and holds in the left arms a lotus with its stalk, a book, a threebranched staff, and a waterpot.103 Śākyaśrī places the noose on the left rather than the right side, while the object that looks like a trident or hook in some Śrīvijayan images is replaced with a threebranched staff (dbyug pa rtze gsum, tridaṇḍa). This specification eliminates the possibility that the bodhisattva holds a trident, an attribute that is seen in a few early depictions and is consistent with the iconographic programme of the Kalparāja. Instead, Śākyaśrī’s visualization has a certain consistency, in that the bodhisattva’s attributes reflect the attributes of the retinue figures. The three-branched staff, waterpot, rosary, and abhaya gesture are displayed by Bhṛkuṭī; Tārā displays varada and lotus; Sudhana holds the book; the tiger pelt is seen on Hayagrīva; the noose remains Amoghapāśa’s signature object.104 In other respects, Śākyaśrī’s and Vibhūticandra’s visions of Amoghapāśa follow the Śrīvijayan type. They all refer to the unusual tiger pelt wrapped over the dhoti.105 Both authors would the translation of Śākyaśrī’s Amoghapāśasādhana at an unspecified place (Q 4840, Dpe bsdur ma ed. 626–7). The place of translation is more likely to have been Thaṃ Bahī than Pu rang, another workplace of the Mustang Translator (cf. Stearns 1996: 135, 160), located in Tibet. This is because Śākyaśrī’s translation was transmitted with interlinear references to the sādhana of Vibhūticandra translated at Thaṃ Bahī: e.g. g.yon gyi bla ma bhi bhu ti’i phyag dpe la… (Q 4840, Dpe bsdur ma ed. 5421). 103. zhal gcig phyag brgyad pa/ g.yas kyi phyag bzhi na phreng ba dang zhags pa dang/ mi ’ jigs pa sbyin pa dang/ mchog sbyin pa’o/ g.yon gyi phyag bzhi na/ pad ma dkar po’i sdong bu dang/ po ti dang/ yal ga gsum pa’i dbyug pa dang/ spyi blugs ’dzin pa’o (Q 4840, Dpe bsdur ma ed. p. 5418 –555). See also Meisezahl (1967: 479). 104. Cf. trans. Meisezahl 1967: 479–480 and Pal 1966, and for comparison, the present Appendix 2.1.2–6. According to Q 4840, Dpe bsdur ma ed. p. 5511–12, Sudhana holds ‘a skullcup with left and right’ (g.yas dang ... g.yon pas ... ka pā li dang), which Meisezahl translates literally and Pal overlooks. Presumably the ‘skullcup’ corresponds to the cupped hands (*puṭa) of Sudhana’s kṛtāñjalipuṭa. 105. According to Śākyaśrī, the pelt is ‘under’ the dhoti: na bza’ dkar po zung gis spras pa’i stag gi lpags pas smad

3/2/22 6:16 PM

26

Iain Sinclair

have been describing objects that they could see, not reworking the old sources for Amoghapāśa. For instance, they identify the scripture held by Amoghapāśa as a book of the Prajñāpāramitā, but to be consistent with the iconographic programme of the Kalparāja, this object should be a book of the Hṛdaya.106 Śākyaśrī reports that he encountered the eightarmed Amoghapāśa at Bodhgaya. He had a vision of the bodhisattva surrounded by attendants during the fasting day (poṣadha) to recover from an illness.107 This incident, first discussed by Jan Schoterman (1994: 158), took place during Śākyaśrī’s final flight out of India around the end of the 12th century. As Śākyaśrī was describing a form that had been in existence for centuries, his ‘vision’ (zhigs, *darśana) appears to have involved the sighting of a physical object. This interpretation is compatible with his claim to have been inspired by a ‘vision in a dream’ (gzims nas bzhugs pa’i tshe). Yet no eightarmed Amoghapāśa has survived at Bodhgaya or anywhere else in India. Any depiction of the form there would most likely have been a late arrival. Pilgrims from Maritime Southeast Asia were, in this regard, becoming more frequent at Bodhgaya during Śākyaśrī’s lifetime. According to Tāranātha (1575–1634), ‘their number went on increasing so that during the time of the four Senas about half the monks of Magadha were from Ko-ki’108—that is,

from ‘Cochin’, a metonym for mainland Southeast Asia.109 Tāranātha’s report is borne out to some extent by the enthusiastic projects to replicate Bodhgaya in Pegu and Siam after the 14th century.110 It is then possible that the image seen at Bodhgaya by Śākyaśrī had been brought there by visitors from mainland or maritime Southeast Asia. Vibhūticandra also apparently describes an image seen with his own eyes, in that he gives detail not provided by Śākyaśrī or in any pre-existing source, but which some Śrīvijayan icons display. His ideal form has a lotus stalk held at the chest with the flower at ear level (gnyis pa na rna ba la rig pa’i padma dkar po’i sdong bu thugs kar bzung ba; cf. 5, 14, 20), and no ornaments (rgyan med ma; cf. 3, 8, 9).111 The model for his description may have been located at Thaṃ Bahī, as Vibhūticandra spent much of his time in exile there, eventually becoming its abbot.112 As for Śākyaśrī, it is likely that he knew that the eight-armed form had not been codified before, and this understanding may have motivated him to write his sādhana. There is no reference to an Amoghapāśa with eight arms in the visualization digests that were circulating at the time, particularly in the Sādhanasamuccaya, a manuscript of which Śākyaśrī had copied in northern Lalitpur in 1216.113 Likewise, the Abhisamayamuktāmālā of Mitrayogin (Q 5022), compiled no

bkris pa (Q 4840, Dpe bsdur ma ed. p. 555–6). This seems to mean that the tiger pelt is positioned below (smad, *adhas) the dhoti, i.e. vertically lower than the knot, as is seen on early images with pelt depictions. The abbreviated sādhana, Q 3682, simply mentions the two garments: stag gi pags pas smad dkris pa. Vibhūticandra refers only to the tiger pelt: stag lpags kyi sham thabs bkris pa (Q 4841, Dpe bsdur ma ed. p. 616–7). See also Meisezahl op. cit. 106. The sixteen-armed form is described as […] pāṇinā […] tṛtīyena amoghapāśapustakam (Kalparāja MS f. 45a3, via Suzuki, Ōtsuka and Kimura 2000: 61). 107. For clarity, the whole account is given here: rdo rje gdan na bzhugs pa’i tshe sku khams bag tsam ma bde nas dpyid zla ’bring po’i tshes brgyad la gso sbyong la gnas pa dang nam gyi brgyad cha tsam lus pa na gzims nas bzhugs pa’i tshe/ ’phags pa don yod zhags pa ’khor dang bcas pa gzims nas bzhugs pa zhal mngon sum du gzigs nas bzhengs pa dang de bzhengs par zhal bstan te sku khams kyang dangs nas sgrub thabs ’di mdzad do (Q 4840, Dpe bsdur ma ed. p. 5810–15). 108. Chattopadhyaya and Chimpa 1967: 330, with the authors’ asterisk marks removed.

109. Tibetan Ko ki is the old toponym Jiaozhi 交趾, ‘Cochin’, pronounced Giaochỉ in Sino-Vietnamese (Ferlus 2009: 4). The countries that Tāranātha understands to comprise Cochin are discussed in brief by MacDonald 1970: 196. 110. McKeown 2010: 57–110. 111. Q 4841, Dpe bsdur ma ed. p. 61, trans. Meisezahl 1967: 482. According to Vibhūticandra, the eight-armed Amoghapāśa also has flowers in the hair (dbu la me tog gi phreng ba can), an unusual detail that is only seen in the Varendra Research Museum MS illumination discussed in section 1.7. 112. Stearns 1996: 137. 113. The sponsor of MS Cambridge Add.1648 is identified as śrīkumbhatīrthamahāvihārīya śākyaśrīsthaviraānandaśrījñāna. This part of the colophon (f. 338r5) was omitted from Bendall’s description (1883: 154–155) and has been overlooked in studies of Śākyaśrī’s life (he is supposed to have returned to Kashmir in 1214; cf. Stearns 1996: 135). For further discussion of this colophon and Kumbhatīrtha, i.e. present-day Kumbheśvara in Lalitpur, see Sinclair 2017: viii–ix §23.

02_CreSouthV1_8P_4Feb22.indd 26

3/2/22 6:16 PM

From Melayu to Thamel and Back

27

later than the early 1200s,114 has no record of the form. As these digests are wide-ranging in scope, bringing together hundreds of visualization procedures, they provide a compelling argument from silence, namely, that the eight-armed Amoghapāśa had not been noticed on the subcontinent before this time. From the artistic corpus it is apparent that the eight-armed form began to be depicted in Nepal only from the beginning of the 13th century. The earliest known Nepalese portrayal is a painting on a manuscript cover (HAR 73801) that was probably produced no later than this time.115 There is also a manuscript illumination which, although East Indian in its style, could have been produced or procured in Nepal in the same period (see the next section). From the 14th century onwards, statues

Fig. 2.6: Eight-armed Amoghapāśa with retinue figures (Sudhana and Tārā, left; Bhṛkuṭī and Hayagrīva, right), vignettes and donor portraits of aṣṭamīvrata devotees. Painting on cloth, damaged, c. 17th century. On public display at Thaṃ Bahī (Thamel), Kathmandu, Nepal. Photograph by Iain Sinclair, 2007.

Fig. 2.5: Seated Eight-armed Amoghapāśa. Pigments on wood, Nepal, c. late 12th or early 13th century. Hand drawing based on HAR 73801. 114. Sinclair 2016b: 282. 115. This depiction of Amoghapāśa (figure 5) is seated, presumably in order to fit the vertically restricted format of the manuscript cover. However, its pose is visually dissimilar to the only seated Malayo-Javanese sculpture (21) and it holds odd objects in two hands. Pratapaditya Pal (2001: 196) asserts ‘a twelfth-century date’ for this painting ‘because of the palaeography’ of the MS kept within the covers, but it is doubtful that the date could be narrowed to within a century on the basis of the writing style alone. See also Sinclair 2016b: 151.

02_CreSouthV1_8P_4Feb22.indd 27

and paintings of the eight-armed form were produced in Nepal in great numbers.116 The earliest dated depiction was painted in 1361.117 Could the eight-armed form have reached Thaṃ Bahī before the 13th century, in the wake of Atiśa’s famous journey to the Malay Archipelago? This seems unlikely in view of the invisibility of the form in the pre-existing literary and artistic corpora. Nonetheless, the possibility of a connection with Atiśa is worth exploring further. Another 116. Lokesh Chandra 1999: 307 lists some examples, not including a number of gilt bronzes dated to the 14th and 15th centuries that were recently auctioned (e.g. HAR 8322, 12948, 13090, 13509, 41061). 117. Lokesh Chandra ibid.; Tuladhar-Douglas 2006: 183.

3/2/22 6:16 PM

Iain Sinclair

28

Fig. 2.7: Eight-armed Amoghapāśa. Miniature, pigments on palm leaf, c. early 13th century, Eastern India or Nepal. Varendra Research Museum MS 851 f.255b. Photograph by Tanaka Kimiaki, reproduced with permission of Md. Abdul Kuddus.

form of Avalokiteśvara distinctively associated with Atiśa was worshipped at Thaṃ Bahī for a time. This form, called ’Phags pa bcu gcig zhal ’jigs pa brgyad skyobs in Tibetan, incorporates the iconography of the Eight Great Fears and is the subject of a painting (HAR 5018) consecrated at the monastery in 1509.118 The eight-armed Amoghapāśa was already established in the local performance tradition by this time, as it occupies the central position at the bottom of the painting. A cloth painting of the eight-armed Amoghapāśa, painted in the 17th century and now in a poor state of preservation, is still displayed at Thaṃ Bahī on special occasions.

a bengali/nepalese manuscript illumination of amoghapāśa The first, last, and only known depiction of the standing eight-armed Avalokiteśvara that can be associated with India appears in an illuminated Aṣṭasāhasrikā Prajñāpāramitā manuscript119 now kept in the Varendra Research Museum, Rajshahi, Bangladesh (No. 851, f. 255b120; Siddhanta 1979: 392). There is a distinct possibility, however, that this lone ‘Indian’ depiction was produced in Nepal. 118. The painting, consecrated at thaṃ bahīri deśa, depicts a thousand-armed Avalokiteśvara (sahasrabhūjaḥ lokyeśvara; sic) described in an ‘oral instruction of Atīśa’ (Willson and Brauen 2000: 269). See also Sinclair 2016b: 159. 119. This manuscript is catalogued in Siddhanta 1979: 385–400 and is described in Melzer and Allinger 2012: 263, P16. More recently see (especially on its iconographic programme) Indian Museum and McCullough 2015; and Tanaka 2016a. 120. Tanaka 2016a: 125.

02_CreSouthV1_8P_4Feb22.indd 28

The manuscript has been assigned to the late Sena period, namely, the early 13th century, on the basis of the style of its miniatures.121 However, the manuscript was an object of worship in Nepal and had a 16th-century Nepalese postcolophon added to it.122 The miniature of the eight-armed Avalokiteśvara is instantly recognizable as the Śrīvijayan type. Some of the bodhisattva’s hands lack the usual attributes (book, staff, noose), indicating that the miniature once again seems to be an eye copy of a visual model painted without reference to a written standard. As some of the figures depicted in the manuscript remain hard to identify, the question arises as to where the illuminator had been working. Kimiaki Tanaka (2016a: 127–128) has noticed that the format of the manuscript is horizontally much narrower (33.5 cm in width) than most Pāla-Sena manuscripts (which measure up to 58 cm in width). It is rather more typical of what Tanaka calls ‘Nepalese domestic substitutes’ for the long palm-leaf writing surfaces of Eastern India. These ‘substitutes’ came into wider use after the 12th century, as communications with Eastern Indian monasteries dwindled to their end. Even though the illuminator of the Varendra manuscript is painting in Bengali style, he could well have been working in the Kathmandu Valley. This is a typical case of uncertainty in determining ‘the exact provenance 121. Tanaka 2016a: 119, 127. 122. Melzer and Allinger 2012: 257 transcribe the postcolophon of the MS, dated 1575–1576. It was once in the custody of a non-celibate monk (gṛhādhivāsitaśākyabhikṣu; cf. Sinclair 2016b: 243) called *Jayalakṣa from Tarumūla-mahāvihāra in central Kathmandu.

3/2/22 6:16 PM

From Melayu to Thamel and Back

29

of many North Indian manuscripts belonging to this period’ (Delhey 2012: 58).

candi jago, gautamaśrībhadra, and the indo-malay kings The last images of the eight-armed Amoghapāśa made in the Malayo-Javanese world are copies of a funerary statue of the Siṅhasāri king Viṣṇuvardhana (r. 1248–1268). The primary stone image of Viṣṇuvardhana was enshrined at Candi Jago in East Java.123 The images modelled after this archetype—produced in the same period and as part of the associated cult—include five small bronze plaques and a sculpture with two inscriptions set up at Padang Roco, West Sumatra, in 1286.124 The creation of this set of images is credited by name to Viṣṇuvardhana’s son and dynastic successor Kṛtanagara (r. 1268–1292). These eight-armed Amoghapāśa depictions constitute some of the most distinctive and accomplished Buddhist art produced in the Malay Archipelago. The literature on them is vast. Yet the question at the core of their conception has hardly been considered: why was the eight-armed Amoghapāśa chosen, at that point in time, to personify the king of Java? Although the eight-armed form had been widely depicted across the Archipelago before the Siṅhasāri period, it was not necessarily a natural choice for identification with the Javanese ruler. The regnal name Viṣṇuvardhana is straightforwardly evoked by the figure of the aṣṭabāhu Viṣṇu, but there seem to have been few immediate opportunities for further identification with the eightarmed Amoghapāśa. Some of the Śrīvijayan statues are large, sumptuous objects (13–18) that appear to have been state-sponsored; however, none of these statues come from Java. The Śrīvijayan corpus shows that no new images had been made in the region for well over two centuries before Candi Jago was built. The cult associated with the Śrīvijayan images was defunct by this time, if the extant corpus can be taken as a guide. The regime that 123. Schoterman 1994; Reichle 2007. 124. The base inscription was first deciphered by N.J. Krom. The text of the inscription published by Ferrand 1922: 123 is referred to here.

02_CreSouthV1_8P_4Feb22.indd 29

Fig. 2.8: King Viṣṇuvardhana as the Eight-armed Amoghapāśa, together with Tārā and Sudhana (left), Hayagrīva and Bhṛkuṭī (right), and the five Tathāgatas and their consorts. Stone with inscriptions and separate inscribed base, Eastern Java, 1285. National Museum, Jakarta, found at Padang Roco, Sumatra. Photograph (‘142 Unknown’, URL: https://flickr.com/ photos/64337707@N07/23121472509) by Anandajyoti Bhikkhu, CC BY 2.0.

propitiated the eight-armed Amoghapāśas across the Malay world seems to have been long gone by the 13th century. The Jago figures have distinct commonalities with the iteration of the eight-armed form that had recently spread in South Asia. While Candi Jago’s statues and statuettes display some stylistic continuity with the Śrīvijayan bronzes, as seen in details such as the tiger pelt and jewelled waistband, they also differ markedly from them. In particular, the Candi Jago icons all feature the tridaṇḍa rather than the triśūla or other alternatives seen in the early corpus. As such, the iconographic programme of Candi Jago is not an unmodified continuation of a local visual tradition. Aspects

3/2/22 6:16 PM

30

Iain Sinclair

of it appear to have been either innovative or imported from outside the region. This possibility of importation is consistent with findings that the Siṅhasāri sculptures were crafted under the ‘influence of sculpture made during the Pāla and Sena dynasties’ (Lunsingh Scheurleer 2008: 297). The sādhana composed by Śākyaśrī has been nominated as the direct source for the Viṣṇuvardhana memorial Amoghapāśas by Schoterman. Schoterman managed to establish that the sādhana was written some decades before the construction of Candi Jago, but offered no explanation as to how this visualization tradition could have reached Indonesia.125 Furthermore, Śākyaśrī’s vision diverges from that of the Jago figures. The Jago tradition depicts a fourteen-deity tableau, which is described as such (*caturdaśātmaka) in the Padang Roco statue base inscription. Here Amoghapāśa is conceived in connection with the five tathāgatas, their four consorts, and the four attendants Sudhana, Hayagrīva, Tārā and Bhṛkuṭī. All fourteen appear in all the Jago-related representations. By contrast, neither the five tathāgatas nor the four consorts feature in the visualization texts of Śākyaśrī and Vibhūticandra. The tathāgatas and consorts are not mentioned by name in these texts,126 nor are their mantras given. The iconographic details of the whole retinue could not be taken ‘as read’ in the full-length meditation programme set out by Śākyaśrī. As the Candi Jago pantheon consistently depicts Amoghapāśa in a fourteen-deity tableau, it is doubtful that Śākyaśrī’s tradition could be its direct model. The Siṅhasāri depictions can be connected to the praxis of the subcontinent by their display of a tridaṇḍa, an object that is codified for the first time in the Śākyasrī visualization tradition. However, the tridaṇḍa is idiosyncratically represented in the 125. ‘The date of Śākyaśrībhadra […] does not impede a possible connection with […] Jago […] but the possibility of such a connection remains to be proven or at least made plausible’ (Schoterman 1994: 158). 126. In Śākyaśrī’s sādhanas, Amitābha in Amoghapāśa’s crown shimmers with the light of the five tathāgatas (Q 4840, trans. Meisezahl 1967: 479; D 2861 199b4). The tathāgatas’ consorts are not mentioned, nor could their details be taken as read in a formal fourteenfigure tableau.

02_CreSouthV1_8P_4Feb22.indd 30

Fig. 2.9: The idiosyncratically shaped tridaṇḍa held by Amoghapāśa in the Candi Jago pantheon. Line drawing based on the bronze Amoghapāśa plaque (detail) of the Rijksmuseum voor Volkenkunde, Leiden, RV-2630-1.

Siṅhasāri figures as three rods held together like a fasces, whereas other depictions represent it more intuitively as a three-branched staff. The offbeat representation of the tridaṇḍa in the Candi Jago figures could not have resulted from the miscopying of a visual model. Instead, it looks to have come out of a naïve attempt to interpret a written source. From this it would seem that new textual material related to Amoghapāśa had somehow reached Java, but was inexpertly interpreted in the creation of the Candi Jago pantheon. One person was especially well positioned to transmit Śākyaśrī’s tradition to the Malay Archipelago: Gautamaśrībhadra (fl. 1248–1255), also known as Gautamaśrī. This Bengali mahāpaṇḍita is located in the Malay world by an inscription on the island of Karimun Besar, Riau, which gives his name, Mahāyāna religiosity and title.127 The inscription was probably written after the 1255 earthquake that damaged Gustala-vihāra,128 Gau127. Sinclair 2018 and forthcoming. 128. Sinclair ibid.; Graldi 2019: 260–262.

3/2/22 6:16 PM

From Melayu to Thamel and Back

31

Fig. 2.10: Eight-armed Amoghapāśa housed behind glass in the dharmārama of Guitaḥ Bahī, Lalitpur, Nepal. Painted clay figure consecrated in 2010. (Photograph by Aurora Graldi, 2012)

tamaśrī’s base in Lalitpur, Nepal. Gustala-vihāra is still part of the eponymous Guitaḥ Bahī129 monastic complex located there. It is likely that Vibhūticandra and Gautamaśrī had been in contact in this very place; a 1248 inscription at Guitaḥ mentions Vibhūticandra.130 According to Tibetan lineage documents, Gautamaśrī studied Raviśrījñāna’s Amṛtakaṇikā commentary with Vibhūticandra,131

who had written his own subcommentary, the Udyotanibandha, which is still extant in Sanskrit. Gautamaśrī also studied with Vibhūticandra’s student Vijayarakṣita.132 Gautamaśrī had other circumstantial connections with the nascent eight-armed Amoghapāśa cult in Nepal. Within the monastic complex at Guitaḥ there is an alcove for religious praxis (dharmāgāra) that traditionally houses an

129. The Newar word gusatala is an archaic form of guitaḥ. It appears to have originally referred to flat ground (tala) on a hill (old Newar gu-sa, contemporary gu-y). This is an apt shrī (sic; Sinclair forthcoming). See also Stearns (1996: 138) description of the monastery’s location. On the antiquity on Vibhūticandra’s studies of the Kālacakra ṣaḍaṅgayoga. of Buddhism at the site, see Graldi 2019. 132. A line of teaching on the rnam rgyal rtog pa (the 130. Sinclair 2016b: 236. *Vijayakalpa) reached Bu ston (c. 1360) via Gautamaśrī by 131. According to the c. 1360 gsan yig of Bu ston (1290– way of Vibhūticandra’s disciple Vijayarakṣita: dzi na shrī/ 1364), this commentarial tradition stemmed directly from bi bhu ti tsandra/ bi dza ya rakṣi ta/ gau ta ma shrī/ bla ma the commentator Raviśrījñāna: re wi shrī/ bi bhu ti/ go tam nyi ma’i zhabs las (Sinclair forthcoming).

02_CreSouthV1_8P_4Feb22.indd 31

3/2/22 6:16 PM

32

Iain Sinclair

Amoghapāśa image.133 The alcove at Guitaḥ Bahī is large enough to accommodate the tall wooden icons of the eight-armed Amoghapāśa that began to be produced in Nepal from the 13th century onwards.134 It is still used for housing large Amoghapāśa icons; a new image was recently installed there. The inscription mentioning Vibhūticandra can still be seen on the base of a caitya just a few metres from the alcove. Gautamaśrī would have known about the ritual that was becoming tied to the eight-armed Amoghapāśa—namely, the tantric poṣadha—through his knowledge of the Kriyāsamuccaya.135 The tantric poṣadha ritual set out in this huge digest does not involve Amoghapāśa directly, but it is structurally very similar to that of the Poṣadhavidhāna, a dedicated handbook for the ritual used at Thaṃ Bahī in the late 13th century.136 The ritual of the Poṣadhavidhāna, as codified in its unique palmleaf witness, concludes with the recitation of Amoghapāśa’s mantra.137 The Poṣadhavidhāna provides the earliest documentation for the widespread Nepalese custom of worshipping Amoghapāśa on the poṣadha day, which still continues in Newar Buddhism. 133. Vajrācārya 1999: 14ff. (fig.[5], śrī amoghapāśa lokeśvara sahita dhaṃlyāhiti). 134. Aurora Graldi, personal communication, 2018. The ‘life-size’ wooden Amoghapāśas constructed with eight arms (all examples are now damaged) include Smithsonian Institution F2000.5, Los Angeles Museum County of Art M.77.19.28, and MMoA 1982.247. 135. Gautamaśrī is named (gotmaśrībhadrapāda) in a list of the Kriyāsamuccaya’s teachers preserved in Sanskrit MS Kaiser Library 110, f. 339v4 . The same list was transmitted to Tibet and is conveyed (with embroidering and errors) in the Deb ther sngon po (Sinclair 2016b: 225). 136. The Poṣadhavidhāna is transmitted in a unique MS at the Royal Asiatic Society (Cowell and Eggeling 1876: 46, No. 70). It has been transcribed and edited in Sinclair 2016b: 323–370 in work that is currently being prepared for publication. 137. The ritual of the Poṣadhavidhāna concludes with a distinct step called the āryāmoghapāśavrata, involving the transmission of one of the bodhisattva’s dhāraṇīs. This dhāraṇī originates in the Amoghapāśakalparāja (MS f. 86v1, ed. Sasaki, Itō and Yamamoto 2010: 192, cf. Sinclair 2016b: 163): [om̐ ] amoghaśīla mahāśuddhasattva bhara bhara sambhara sambhara padmavibhūṣitabhuja dhara dhara samantāvalokiteśvarāya (Poṣadhavidhāna MS f. 6r6–v1, ed. Sinclair 2016b: 358).

02_CreSouthV1_8P_4Feb22.indd 32

Whereas the first images of the bodhisattva produced in Nepal are aligned with the ‘outsider’ (bahī) monastic tradition, as distinct from the royal monasteries that sheltered local monks,138 the Siṅhasāri Amoghapāśas are definite products of royal patronage. The inscription on the base of the Padang Roco Amoghapāśa sets out an imperial agenda. It records that the statue was transported from Java accompanied by courtiers at the behest of mahārājādhirāja Kṛtanagara. Both the icon (bharāla) and the king are referred to with the honorific pāduka. This object was purpose-built to be looked up to by the Malay people: The virtue of this monument is fit for the joyous salutations [of] all natives in Melayu country (saka praja di bhūmi malāyu)—Brahmins, Kṣatriyas, Vaiśyas, Śūdras, Āryas—especially the śrīmahārāja and śrī-Tribhuvana King Maulivarmadeva.139

A striking association is being forged here between the eight-armed Amoghapāśa and Melayu, ‘Malay’, country. The phrase praja di bhūmi calls to mind the present-day nativist terms pribumi or bumiputera. Its use in the context of the inscription implies that the eight-armed bodhisattva had a special relationship with the people of the Malay world, and ‘especially’ (pramukha) with their ruler. How the statue’s Javanese backers came to perceive such a relationship is not known, but at this point they had presumably occupied former Śrīvijayan strongholds where such images were enshrined. The official and internal Javanese view, as conveyed in Mpu Prapañca’s Deśavarṇana, is that this very 138. All sites associated with early Amoghapāśa images in Nepal belong to monasteries of the bahī class, namely, Thaṃ, Guitaḥ, and Pūco Bahī (Sinclair 2016b: 151). On the meaning of bahī as ‘outside’ (the cosmopolitan/metropolitan centre), see Sinclair 2016b: 49–50. 139. kunuṅ punyeni yogya di-anumodanāñjaleḥ saka praja di bhūmi malāyū, brāhmaṇaḥ kṣatriya vaiśya sūdra, āryyā māddhyāt, śrī mahārāja śrīmat tribhuvanarāja maulivarmmadeva pramukha (Krom 1917: 326). The literature on the inscription (e.g. Ferrand 1922: 123) gives varying translations. The interpretation of words such as punya, di, madhyāt (skipped to avoid forced wording in this translation) and pramukha would benefit from more information on the language of the inscription (classified as ‘Old Malay’ by Griffiths 2018: 282).

3/2/22 6:16 PM

From Melayu to Thamel and Back devamūrti140 was instrumental in the ultimate defeat of the Malay lands.141 Kṛtanagara began his pamalayu campaign in 1275, over ten years before the installation of the statue, and Prapañca leaves no doubt that this conflict ended with Melayu’s submission.142 The Javanese king, in the view of his court, co-opted and absorbed the divine protector(s) of the Śrīvijayan–Melayu polity along with the polity itself. The Padang Roco statue formalized the annexation of Maulivarmadeva and his people at more than one level. Its inscription not only dictates Maulivarmadeva’s status as a vassal, a lesser mahārāja than mahārājādhirāja Kṛtanagara; it sets up the ‘murtified’ Javanese king as a figurehead for the whole Malay people. The figure of Viṣṇuvardhana– Amoghapāśa enshrines the sacrality of Indo-Java over and above that of Indo-Malaya. Tribhuvanarāja Maulivarmadeva, insofar as he can be identified with Sang Sapurba Tribhuvanarāja in the Sejarah Melayu literature,143 is indeed a ‘Hindustani’ king. The Indic pedigree of Maulivarmadeva’s kingship is clearly being acknowledged in the Padang Roco inscription. His Javanese overlords place him at the apex of the four-caste hierarchy, which is rarely associated with the classical Malay social order.144 140. In the understanding of Schlüter 2010, the term devamūrti would refer to Kṛtanagara himself (‘Intense was going to be their fear because of His being a divine incarnation’). 141. nagāsyabhawa śāka [1197] saṅ prabhū kumon dumona rikanaṅ tanah di malayū / ḷwĕs mara bhayanya saṅka ri khadewamūrttinira ṅūni kālahan ikā (Deśavarṇana 41.5cd, ed. Pigeaud 1960: 31). See also trans. Robson 1995: 54. 142. Melayu is described as bowing to Kṛtanagara’s feet in Deśavarṇana 42.2b (ed. Pigeaud 1960: 32). See also trans. Robson 1995: 55. Such frank assertions of dominance make it hard to see how ‘the relationship between the two realms may have appeared (at least officially) friendly’ (Reichle 2007: 122). 143. Sinclair 2019a. Note, however, that the various Sejarah Melayu recensions offer significantly differing accounts of Sang Sapurba’s career. His status as raja melayu and bangsa hindustan is specified only in recension II (ed. Ahmat Adam 2016: 3625, 3915), as defined by Chambert-Loir 2017: 169. 144. This attitude can be contrasted with the low consciousness of four-caste norms in Java at the time. Following Hall 2000: 61ff., it seems that generic ‘clergy’ and ‘royalty’ were more compelling social categories than brāh-

02_CreSouthV1_8P_4Feb22.indd 33

33

The non-caste society of the Sumatrans is all but passed over in the inscription, with only the word ārya, perhaps, alluding to a Buddhist or ‘Indian’ segment of the population.145 The Padang Roco Amoghapāśa in this way presents itself as a totem of normative Hindu–Buddhist unity, embodying its Javanese creators’ worldview. Whereas the Padang Roco Amoghapāśa was oriented towards the then centre of the ‘Malay’ world, the other Siṅhasāri Amoghapāśas are part of a still larger and more cosmopolitan project. The Candi Jago stone sculptures have their names engraved in a script classifiable as Proto-Bengali-cum-Proto-Maithili,146 which at the time was used in the northeast of the subcontinent, and is otherwise almost unknown in Indonesia. The bronze plaques of the Candi Jago Amoghapāśas are inscribed in similar yet again distinct and foreign types of script.147 The designers of Candi Jago were making ‘Indic’ icons, meaningful to the wider Sanskritic world. But to whom, exactly, the Jago icons were meant to speak is not yet known. The inscribed Amoghapāśas seem to anticipate an event of the kind narrated in the Sang Sapurba episode of the Sejarah Melayu, in which an Indian xenarch arrives in Sumatra and has his claim to rule accepted.148 The artisans of Candi Jago were themselves in close proximity to people from the subcontinent, as can be seen from the distincmaṇa and kṣatriya in 13th- and 14th-century Java. The question of why the Javanese displayed concern about the Malay social order will have to be explored elsewhere, but see in brief Sinclair 2019a: 7. 145. The word āryya is not easily construed with the brāhmaṇa or kṣatriya mentioned in the same sentence (cf. n. 137 above) in the Padang Roco inscription. As such, these āryas are at least potentially Buddhists, if not part of the āryasaṅgha. The Buddhist king Ādityavarman, who later ruled this area, identifies as āryavaṃśa in his Jago inscription of 1343 (cf. Kozok and van Reijn 2010: 139). 146. Sinclair 2016b: 326, Table 3. 147. These remarks are made with reference to the following plaques: Berlin MIK II 233 (Moeller 1985: 23, pl. 8b); MMoA 2014.517; Rijksmuseum voor Volkenkunde, Leiden, RV-2630-1 (Bernet Kempers 1959, pl. 258 et al.); and two referred to by Speyer 1904a: 140, 1904b: 255. Identifications of the regional affiliations of the scripts used to engrave these plaques will require a separate study. 148. See, in brief, Sinclair 2019a.

3/2/22 6:16 PM

34

Iain Sinclair

tive writing engraved on the backs of the votive plaques. The powerful cosmopolitan aura exuded by the inscribed Siṅhasāri Amoghapāśas persisted well beyond the time and place of their production. Two generations after the Padang Roco Amoghapāśa was set up, in 1347, the statue was again at the centre of attention in Melayu country. It was moved from its initial location, reconsecrated and re-inscribed in Sanskrit by Maulivarmadeva’s grandson, Ādityavarman (c. 1293–1375).149 This extraordinary act was an ‘appropriation’150 of the object that had once enshrined the subordinate status of Ādityavarman’s forebear and his people. Ādityavarman’s repurposing of the Padang Roco Amoghapāśa marked the assumption of sovereignty over his grandfather’s domain, the Sumatran highlands, as reflected in his new and lasting use of the title mahārājādhirāja.151 This and other Sanskrit epigraphs of Ādityavarman give the collective impression of reviving the zeitgeist of the late 13th century,152 during which the Indic and Malayo-Javanese poles of Hindu-Buddhist civilization enjoyed a final round of contact. With Ādityavarman’s added epigraph, it becomes apparent that the worship of the eight-armed Amoghapāśa spans almost the entire history of Malay Buddhism, from its Śrīvijayan beginnings to its endgame in the 14th-century enclaves ruled by Maulivarmadeva’s descendants.153 Candi Jago is, incidentally, still a low-key site of worship, many centuries after the 149. The text of Chatterjee and Chakravarti 1933: 81–82 is followed here. On the relation of Ādityavarman’s regnal title maulimaṇivarmadevamahārājādhirāja (Chatterjee and Chakravarti ibid.) to Maulivarmadeva / Sang Sapurba, see in brief Sinclair 2019a. 150. Reichle 2007: 127 uses the term ‘appropriation’ in this context. 151. Sinclair 2019a. 152. Hendrik Kern’s assessment of the inscription as ‘not much more than gibberish’ (trans. Kozok and van Reijn 2010: 144ff.) could hardly be expected to shed much light on Ādityavarman’s project. For some references to previous work on Ādityavarman’s epigraphic corpus, see Sri Ambarwati Kusumadewi 2012. 153. We might wonder, along with Kwa 2010, whether an eight-armed Amoghapāśa had become a tutelary deity of Maulivarmadeva’s line, and came to be installed in Singapura by Sang Sapurba’s son Sang Nila Utama.

02_CreSouthV1_8P_4Feb22.indd 34

collapse of the religious and dynastic institutions associated with it.154 It can further be seen that the Viṣṇuvardhana-Amoghapāśa images followed a trail blazed by the Jayabuddhamahānātha cult of the Khmer Empire. The similarities between the Khmer and Javanese power-projection projects are remarkable, starting with the choice of the eight-armed Amoghapāśa as the sovereign’s avatar. Both Jayabuddhamahānātha and the murtified Viṣṇuvardhana were realized as portrait statues, which were distributed in the form of miniature bronzes as well as stone sculpture. A possible pathway for the spread of the devamūrti cult is the close contact between the Malay and Khmer worlds that existed during the reign of this or another Maulivarmadeva.155 It has been ventured that the stocky portrait statues of Kṛtanagara, the Joko Dolog and its replicas, were inspired by the Khmer Jayabuddhamahānātha statues (Reichle 2007: 25). One other stone sculpture from the Archipelago displaying eight arms has proven to be difficult to fit into a chronology. This sculpture is currently housed in the Pura Puseh of Kutri, Buruan village, Bali.156 It displays definite stylistic affinities with the Candi Jago images, such as the jewelled waistband. It displays wafting scarves that look like the tall lily pad stalks of Siṅhasāri sculpture. However, the Kutri figure, unlike the Candi Jago Amoghapāśa, carries a hook and conch. These objects are common in the Khmer iconography of the bodhisattva, as seen in the bronzes of Jayabuddhamahānātha, but the conch is most appropriate for the eight-armed Viṣṇu. The iconography may 154. Waterways near Candi Jago are reportedly still treated as tirtha by the local Tengger community (see e.g. ‘Begini Prosesi Ruwatan di Desa Candi Jago Tumpang’, Malang Today, 5 November 2016). 155. De Casparis (1967: 34–35) argued for the identification of Maulivarmadeva with the Maulibhūṣaṇavarmadeva who commissioned the so-called Buddha of Grahi (on which see e.g. Krairiksh 2012: 352–353, pl. 2.376). Although de Casparis’ identification is historically neat, it is paleographically improbable that the digits 1 1 0 0 6[?] inscribed on the Buddha of Grahi’s pedestal can yield a date falling within the expected dates of Maulivarmadeva’s reign. Maulibhūṣaṇavarmadeva must then be regarded as a different, earlier king. 156. See e.g. Reichle 2007: 111.

3/2/22 6:16 PM

From Melayu to Thamel and Back

Fig. 2.11: Eight-armed deity, usually identified as Amoghapāśa, holding a hook in the upper right hand and a conch in the upper left hand. Stone, c. 10th–13th century. Pura Puseh, Kutri, Buruan village, Gianyar, Bali. (Photograph by Andrea Acri, 30 December 2018)

then have some conceptual connection with Viṣṇuvardhana. As Bali is known to have been conquered by Siṅhasāri before Melayu, the Kutri statue may represent a kind of precursor to or prototype of the Siṅhasāri Amoghapāśas. As the Kutri statue has only been photographed in a damaged state, without vital identifying features such as the crown, it cannot be conclusively identified as Amoghapāśa and is discussed no further here.

the persistent significations of the eight-armed amoghapāśa Throughout the transmigration of the eight-armed Amoghapāśa, the question arises: what distinctive

02_CreSouthV1_8P_4Feb22.indd 35

35

task is performed by the bodhisattva with eight arms? Why could it not be performed by the forms with four arms, six arms, ten arms and so on? Why worship one instead of another? Problems of meaning and interpretation, which were peripheral in charting the bodhisattva’s lightly documented earlier career, come to the fore again in appraising Amoghapāśa’s final incarnation. How much was retained from the early religiosity surrounding Avalokiteśvara, which promoted the hope of rescue from the Eight Great Fears?157 Avalokiteśvara images that were thought to have offered protection from drowning were prominent in port cities and maritime regions, as has been shown by Bopearachchi (2014, 2017). Yet if safe seafaring was the bodhisattva’s original impetus, it is lost in the technical literature on Amoghapāśa produced in the Kathmandu Valley. At least three visualization traditions of the eight-armed Amoghapāśa have been transmitted in Sanskrit. They are not known to have been translated into Tibetan and are distinct from Śākyaśrī’s and Vibhūticandra’s sādhanas.158 They occur in anonymous handbooks for visualizing Amoghapāśa witnessed in a large number of manuscripts, none older than the 16th century. Most are in poor textual condition and have remained unstudied. They describe the visage of the eightarmed Amoghapāśa without substantial differences in posture, complexion, or the objects held in the hands. As such, they reflect a single late phase in the development of the type, rather than a protracted inflow that covers all of the variation seen across other parts of Asia. Altogether they fit the picture gleaned from the art-historical corpus: the bodhisattva first comes into view on the subcontinent relatively suddenly, in Nepal, and in connection with an exodus of Buddhist specialists from India. One visualization of the eight-armed Amoghapāśa circulating in Nepal provides an 157. Bautze-Picron 2004: 241; Acri 2019: 54, n. 6. 158. Three Sanskritic traditions of visualization are represented in the following texts: (1) an Amoghapāśapūjāvidhi partly transcribed by Pal (1966: 234); (2) an Amoghapāśapūjā conveyed in MS NGMPP H 314/10; (3) the Amoghapāśamaṇḍalarājāgrī nāma samādhi transcribed by Badrīratna Vajrācārya (2005:[2]–[7]) and reedited in this chapter.

3/2/22 6:16 PM

Iain Sinclair

36

especially clear picture of the late conception of the bodhisattva. This maṇḍalarājāgrī samādhi procedure forms part of an elaborate ritual of the ‘eighth-day vow’, the aṣṭamīvrata. In Newar Buddhism, the aṣṭamīvrata has become synonymous with the fasting day, the poṣadha, which is practised in a monastic or semi-monastic setting. Its practitioners are often people who are undergoing serious distress and hope for the bodhisattva’s intervention.159 The samādhi text appended to this chapter sets out a complete visualization of the bodhisattva’s maṇḍala as well as his physical appearance and retinue. It was first published by Badrīratna Vajrācārya (2005) without notes or references to sources and has received little further attention. It is re-edited here (Appendix 2.2) using the readings of three manuscripts and various textual parallels. A text-critical appraisal of this maṇḍalarājāgrī samādhi of Amoghapāśa reveals extensive borrowing from a mature Sanskrit corpus, which points to it having been composed after the early 13th century. The universe-generating meditation is lifted from a sādhana of the Harihariharivāhanodbhavalokeśvara. The emanation of the bodhisattva and the description of his retinue is cribbed from the Khasarpaṇasādhana by Padmākaramati. The passage on maṇḍala architecture is common boilerplate in the praxis literature of the Guhyasamājatantra going back to Nāgabuddhi.160 All of these works were circulating in Nepal before the mid-13th century. The only distinctive part of the visualization is the couple of sentences on the objects held by the eight-armed Amoghapāśa. The text can altogether be regarded as a typical ‘post-Indian’ Buddhist work that is driven by a practical and unpretentious need to document a religious inheritance. Its vision of Amoghapāśa unfolds as follows: Next is the concentrated state called Maṇḍalarājāgrī. Now, having cultivated the four Brahmic states—friendliness, compassion, delight, equanimity—[repeat]:

159. Sinclair 2016b: 115. 160. See Appendix 2.2, notes a–d.

02_CreSouthV1_8P_4Feb22.indd 36

om̐ , all dharmas are inherently pure, i am inherently pure. i have the inherent nature of the vajra of knowledge of emptiness. Then, emanated by means of the syllable yam̐ , is a bow-shaped wind maṇḍala marked with a blue flag; on top of it, arising from a ra[m̐ ], a red fire maṇḍala, triangular, marked with a red ra glyph; above that, transformed from the syllable vam̐ , a water maṇḍala, round, clear in colour, marked with a white bell; above that, arising from the syllable lam̐ , an earthen maṇḍala, four-sided, yellow, blazoned with yellow threepronged vajras in the four corners; above that, emanated from the syllable sum̐ , Sumeru, made of four jewels, boasting eight peaks underneath.   The residence, the crystalline cavern of blessed Potalaka mountain, is a pagoda adorned [with] four doors [and] four tympana, having sun-and-moon emblems at all four corners of the points of contact between the doorway and the pediments (niryūha) and decorated with pearl and half-pearl [strings], enclosed by four cords, beautified by banners and streamers, girded by piping and parapets, the vase-capital columns studded with vajra gems. Also, the sides have a great vajra parapet festooned with a net of tinklebells. [The maṇḍala] is adorned with eight flagstaffs, is well decorated with yaktail fans; the sun outside is in the correct place. 1.  In the internal maṇḍala transformed from the syllable paṃ, upon a lotus with all-spreading petals, is the full moon maṇḍala generated from the syllable am̐ . Having visualized upon it the syllable hrīḥ, extremely bright, [its] concentration of shimmering rays pervading the sky, he should look upon the nucleus of his syllable changed into an outspread lotus adorned with a stalk. Then he should imagine all of that transformed into the persona of the Bhagavat. His body the dazzling white of a crore of snow-glints, he has a smiling face, about twice eight years [old], looking compassionate and affectionate, attended by passionate sentiment, most peaceful, complete with all [auspicious] marks, bereft of all non-[auspicious] marks, single-faced, wearing ascetic garb; having a crown of piled-up locks, Amitābha is fashioned in the crest; he is adorned with divine ornaments of gold, ruby, pearl, lapis and jewels of all

3/2/22 6:16 PM

From Melayu to Thamel and Back kinds. He is eight-armed,161 on the right [side] bestowing a stream of ambrosia with the first hand, holding a rosary with the second hand, holding a noose with the third hand, holding the jewel of unfear with the fourth hand; on the left, holding a thousand-petalled lotus with stalk with the first hand, with the second hand satiating the needlemouthed [preta] below him tasting the to-and-fro of the stream of downpouring nectar, holding a book of the Prajñā with the third hand, holding a waterpot with the fourth hand—in this way should he conceive the persona of Infallible Noose (Amoghapāśa), the glorious Lokeśvara. He should think that the persona of Vajradharma, red in colour, seated on a peacock, is emanated from a hrīm̐ syllable on the Bhagavat’s crest. 2.  On his outside, in front of the Bhagavat, is Amoghāṅkuśa, red in colour, a hook held in his right hand, holding a noose in his left. 3.  On the Bhagavat’s right is Tārā, dark [green], in the left hand grasping with the stalk a blue waterlily being opened up with the right hand. [She is] replete with a variety of jewelry, bearing breasts of freshest youth. 4.  And in the southeast is prince Sudhana, cupping [his hands] in salutation; a book tucked under his left arm, he has the lustrous sheen of gold, maintaining the appearance of a prince laden with all ornaments. 5.  On the left is Bhṛkuṭī, four-armed, of golden aura, her locks fanned out, holding in hand a triple staff and waterpot on the right, the arms on the left waving in greeting and holding a seed rosary. [She is] three-eyed. 6.  In the northeast is Hayagrīva, red in colour, squat and potbellied, having flamelike orange hair, an anaconda for a sacred thread, tawny facial hair lined around the circle of his mouth, three round red eyes, frown-furrowed brows, and a tiger-pelt garment. Armed with a club in his right hand, he waves in greeting. […]

This is the mature form of the eight-armed Amoghapāśa and his four followers, as seen in 161. A similar description is translated by Sakuma 2015: 132.

02_CreSouthV1_8P_4Feb22.indd 37

37

the Kathmandu Valley, the Candi Jago pantheon and, with slight differences including the omission of the optional retinue figures, the Śrīvijayan artistic corpus. One distinctive feature of this visualization text is the addition of the bodhisattva Amoghāṅkuśa, which may reflect Nepalese interests. Amoghāṅkuśa is identifiable with the red Trailokyavaśaṃkara form of Avalokiteśvara, who in turn had been identified with Nepal’s ‘national deity’ Buṅgadyaḥ.162 However, the addition of Amoghāṅkuśa in this visualization tradition had little impact on the artistic corpus,163 and is visible in just one Amoghapāśamaṇḍala painting consecrated (possibly in Thamel) in 1516.164 The maṇḍalarājāgrī samādhi text goes on to detail the rest of Amoghapāśa’s maṇḍala, which contains the eight great bodhisattvas (7–14), the four offering goddesses (15–18) and the four bodhisattvas of attraction (19–22), generally the common property of Vajradhātu-system maṇḍalas. The outer part is occupied by the eight gods of the directions at the periphery (23–30). This maṇḍala— together with other Amoghapāśamaṇḍalas circulating in Nepal165 that have no particular scriptural source166—are both formulaic and arbitrary in a way that indicates a late, isolated context, unconcerned with high punditry. They provide no further leads on the early throughline of the eight-armed form. The eight-armed Amoghapāśa has been worshipped as the central figure of the poṣadha ritual since its Nepalese debut. Of course, there was a long-standing link between the fasting day and the cult of Amoghapāśa going back to the Hṛdaya.167 However, other forms of Avalokiteśvara had been 162. Sakuma 2006. 163. One example is Amoghāṅkuśa’s presence in the aṣṭamīvrata (poṣadha) ritual conveyed in MS Shakya 2010 No. 18, exposure 36. Here Amoghāṅkuśa is to be invoked with his characteristic gesture (abhinaya) immediately after the eight-armed Avalokiteśvara. 164. The partly legible dedication of HAR 58964 could be read … vaṃtha bahāra or …-vaṃ tha bahāra, i.e. Thaṃ Bahī. 165. Sakuma 2017. 166. Tanaka 2019: 55. 167. The poṣadha is the main occasion for reciting the Hṛdaya (cf. ed. Meisezahl 1962: 3275).

3/2/22 6:16 PM

38

Iain Sinclair

worshipped on poṣadha days before the eightarmed form became the main object of worship in Nepal.168 The eight-armed form was introduced at a time when the observance of the poṣadha had acquired another purpose: it had become a prerequisite for initiation into the Hevajratantra. This Tantra asked prospective initiates to observe the poṣadha before taking up the Hevajra tantric system.169 Previously poṣadha celebrations had been monastic affairs that upheld conservative Buddhist norms, but by the 13th century they were increasingly conceived as introductions to tantric Buddhism. This shift is evident in the early handbooks for the tantric poṣadha, such as the aforementioned Kriyāsamuccaya and Poṣadhavidhāna; its many and varied aspects are traced elsewhere.170 The eight-armed Amoghapāśa was therefore positioned in Nepal at a potential point of entry to tantric praxis, though not overtly and not in a manner that diminished the traditional integrity of poṣadha observances. Associations between monastic life, Hevajra praxis and the worship of the eight-armed Amoghapāśa can be traced back to Southeast Asia. It is well known that Atiśa was a Hevajra yogin171 before being ordained and embarking on his voyage to the Malay Archipelago. In his later career Atiśa famously insisted on the priority of monastic vows. He called for regular performance of the poṣadha at Thaṃ Bahī, and set up a Hevajra shrine there at its founding.172 The ritual outlined in the Poṣadhavidhāna, which codifies Thaṃ Bahī’s praxis tradi168. The eight-armed form displaced older forms of Avalokiteśvara that had been worshipped in Nepal, especially Ekādaśamukha (Sinclair 2016b: 154–159) and the twelvearmed form (ibid.: 159–162), which might have been called Avalokitanātha (Sinclair 2017: xii). 169. The directive poṣadhaṃ dīyate prathamam (Hevajratantra II.8.9a, ed. Tripathi and Negi 2006: 183) and its impact on Buddhist tantrism in Nepal is discussed in Sinclair 2016b: 85ff. 170. The affiliation of the Poṣadhavidhāna to Haivajrika tantrism is partly but strongly indicated by the verse beginning ātmatattvasya prasādakiraṇaiḥ … (MS f. 2r8 –v1), is discussed in Sinclair 2016b: 330, 360–361. Its ritual is similar to that of the Kriyāsamuccaya, which is Haivajrika in orientation. 171. Vetturini 2007 I: 52. 172. Sinclair 2016b: 318–319.

02_CreSouthV1_8P_4Feb22.indd 38

tion, reveals definite commonalities with Atiśa’s religiosity.173 In the Khmer Empire,174 the eightarmed Amoghapāśa had been appearing alongside Hevajra in institutional settings long before the 13th century. The joint veneration of Amoghapāśa and Hevajra practised in Ādityavarman’s kingdom has definite precedent on the subcontinent and elsewhere in the region. The Amoghapāśa with eight arms is an apt symbol for the poṣadha, given that it entails an ‘octopartite fast’ (aṣṭāṅgapoṣadha) of eight overnight vows held on the eighth day of the lunar month (aṣṭamīvrata). The worship of Amoghapāśa is further said to bestow eight rights (aṣṭau dharmān) to a dignified death.175 The relation between these eightfold sets and Amoghapāśa was already established in the Hṛdaya, which remained in wide circulation in Nepal. Yet no known interpretive tradition identifies the eight arms of the bodhisattva with the eightfold aspects of the fasting day.176 The old meaning of the expression aṣṭāṅga, as used in connection with the poṣadha, was lost in Nepal soon after the arrival of the eight-armed form.177 To this day there is no authoritative native interpretation of the significance of the eight arms in Nepal. The eightfold correspondences linked to the eight poṣadha vows seem to have had nothing to do with the form of the eight-armed Amoghapāśa. Alternatively, the uptake of the eight-armed bodhisattva in the Sanskritic world can be contrasted with the prevalence of non-Buddhist eight-armed deities. The attraction of potential converts to Buddhism is a well-known aim of 173. On the bodhicittotpādana verses originally from the Mañjuśrībuddhakṣetraguṇālaṃkāravyūha, which were not widely circulated but which both Atiśa and the Poṣadhavidhāna commend for use (MS f. 5r7 –v5), see Sinclair 2016b: 143–144, 366. 174. Sharrock 2013: 50 likewise speaks of the ‘Lokeśvarization’ of Khmer Hevajra statues. 175. Hṛdaya, ed. Meisezahl 1962: 3185–6. See also Reis-Habito 1999: 43 and Sonali Dhingra’s chapter in the second volume of the present edited collection (Dhingra 2022). 176. Sinclair 2016b: 151. 177. On the aṣṭāṅga of the poṣadha interpreted as prostrating with the eight limbs of the body (aṣṭāṅgapraṇāma) in Nepalese avadānamālā literature, see Sinclair 2016b: 152, n. 94.

3/2/22 6:16 PM

From Melayu to Thamel and Back poṣadha celebrations.178 Nepalese depictions of the bodhisattva, for instance, are often confused with Lakṣmīnārāyaṇa, who also has eight arms, a standing posture and a light complexion, and who began to be depicted widely in Nepal from the 13th century onwards.179 This hermaphroditic Vaiṣṇava deity had a certain prominence in the Newar visual culture of the period. He/She has been portrayed in quasi-Buddhist maṇḍala compositions and is often confused with Amoghapāśa. One visualization tradition asserts differentiation by having Lakṣmīnārāyaṇa trampled on by Vajrahūm̐ kāra.180 In spite of these strong visual commonalities, it is difficult to see how the eightarmed Avalokiteśvara could have been promoted with the hope of being accepted accidentally by devotees enculturated in Hinduism. The most compelling explanatory rubric for the eight-armed iconography remains the discourse of the Eight Great Fears. Sanskrit sources on the Fears are propagated in the Nepalese corpus and were available to devotees throughout the bodhisattva’s Nepalese career. The power of Avalokiteśvara to avert drowning is an old trait of the bodhisattva, as summed up in an oft-quoted verse of the Saddharmapuṇḍarīka: He might fall into the vault of the ocean— abode of serpents, sea-monsters, antigods and spirits; after remembering Avalokiteśvara, he never sinks into the king of waters.181

Similarly, the reciter of the Hṛdaya prays to be saved from ten fears, including the Fear of Water

178. This was argued at length in Tuladhar-Douglas 2006. See also Sinclair 2016b: 157ff. 179. Sinclair 2016b: 152–153. 180. lakṣmīnārāyaṇam ākrāntaṃ, vajrahūm̐ kāraṃ bhāvayet (Mañjuvajramukhākhyāna, ed. Tanaka 2018: 54 9, regularized text). The mukhākhyāna genre appears to be unknown outside Nepal (Sinclair 2016b: 180, n. 214). Lakṣmīnārāyaṇa is also mentioned in the Nepalese Yogasiddhāntabuddhasiddhitantra. 181. saci sāgaradurgi pātayen nāgamakarasurabhūtaālaye | smarato avalokiteśvaraṃ jalarāje na kadāci sīdati (Saddharmapuṇḍarīka 24.6, ed. Dutt 1953: 2959–10), cited also by Bopearachchi 2014: 164 (with another, similar passage from the same chapter) and Acri 2019: 58, n. 11.

02_CreSouthV1_8P_4Feb22.indd 39

39

(udaka).182 The Kalparāja refers to the Eight Great Fears as a set (MS f. 24a2). It is clear that the Malayo-Javanese world, as well, looked upon Avalokiteśvara as a saviour from the Fears. The only Sanskrit work related to Avalokiteśvara that has survived in the region, the Aṣṭamahābhaya Kliṅ transmitted in Bali,183 is nominally concerned with the Eight Great Fears. It in fact gives mantras to quell ten fears, as well as to avert harm from poisons and other dangers. While textual parallels to the Aṣṭamahābhaya Kliṅ have not yet been located in the wider Sanskritic tradition,184 its spells are broadly similar to those given in the literature on Amoghapāśa. It would be natural in this context for the noose held by Amoghapāśa to be recognized intuitively as a lifeline for sailors in peril.185 The eight-armed form therefore embodies in an almost self-evident manner the capacity for rescue from diverse dangers, in which the fear of maritime disaster is foremost. It is the natural form for Avalokiteśvara to take on in the Malay Archipelago, where Buddhism and sea journeys had longstanding linkages.186 While the earlier association with the protection of seafarers is lost in the late sādhana literature, it is still visible in the visual culture accompanying the bodhisattva in Nepal. Several Nepalese paintings of Amoghapāśamaṇḍalas—for instance Yale University Art Gallery 1988.76.14— include vignettes of the Great Fears. Among these vignettes are horrific portrayals of drowning men being devoured by swarms of marine life. These are remarkable not only for their variety but for the fact that they were painted in landlocked Nepal. 182. vadha-bandhana-tāḍana-tarjana-rāja-taskara-agny-udaka-viṣa-śastra-parimocaka (Hṛdaya, after ed. Meisezahl 1962: 3235–6). 183. Goudriaan and Hooykaas 1971: 309–312, No. 504. 184. Sylvain Lévi’s claim (quoted in Goudriaan and Hooykaas 1971: 310) that ‘any Buddhist pandit [in Nepal] can recite’ the Aṣṭamahābhaya Kliṅ has not been corroborated by the discovery of a substantially similar text in the Nepalese corpus. 185. This observation was ventured by Lokesh Chandra to Mr Kwa Chong Guan (personal communication, September 2018), who has long been pondering the significance of Amoghapāśa in the region (cf. Kwa 1967–1968). 186. Acri 2019.

3/2/22 6:16 PM

Iain Sinclair

40

Fig. 2.12: Eight-armed Amoghapāśamaṇḍala surrounded by Buddhas, vignettes of the Fears and donor portraits. Second row from left: Buddha; Fears of Lightning, Snake, Fetters, Rapacious Beast, Execution, Water; Avalokiteśvara. Painting on cloth, Nepal, c. 15th century. Yale University Art Gallery 1988.76.14 (HAR 58867). (Public Domain)

02_CreSouthV1_8P_4Feb22.indd 40

3/2/22 6:16 PM

From Melayu to Thamel and Back

41

Fig. 2.13: Vignettes of the Fear of Water in Eight-armed Amoghapāśamaṇḍalas, depicting drowning men devoured by marine life. Hand drawings based on paintings on cloth, 14th–15th centuries, Nepal. Left: two vignettes from V&A 1S58-1977 (after MacDonald and Vergati Stahl 1979: 128 pl. V). Centre: Yale University Art Gallery 1988.76.14 (Figure 12), detail. Upper right: HAR 77041, detail. Lower right: HAR 18346, detail.

Fig. 2.14: The Fear of Water, represented by a drowning man being devoured by marine life and surrounded by serpents. Wood carving with original pigments and contemporary overpainting, Ta Bāhāḥ, Lalitpur, Nepal. (Photograph by Iain Sinclair, 21 February 2010)

02_CreSouthV1_8P_4Feb22.indd 41

3/2/22 6:16 PM

Iain Sinclair

42

In the Kathmandu Valley, then, the eight-armed Amoghapāśa preserves informal but tangible reminders of its past life as a seafaring deity.187

conclusions The first diffusion of the eight-armed Amoghapāśa has been traced here to the Malay Archipelago. It spanned the early 7th to the early 11th centuries, lasting as long as Śrīvijaya itself. This initial spread built on an old culture of devotion to Avalokiteśvara as a saviour of seafarers while also drawing on a cultus of Amoghapāśa that had begun making its way across Asia. Soon after its debut in the Archipelago, the eight-armed form began spreading to China, bearing only indirect traces of its maritime origins. It persisted for some centuries, without discernible scriptural support, on the fringes of East Asia. Towards the end of the first millennium, the eight-armed form branched into the Khmer Empire, most likely through contacts with the Malay world. The Khmers repurposed the form to embody the religiosity of the Kāraṇḍavyūha. The bodhisattva’s penultimate iteration, which remains a living part of what is now called Newar Buddhism, debuted at the beginning of the 13th century. It is not yet known how the bodhisattva’s cult first reached South Asia, but the major upheavals then occurring in transregional Buddhist networks may have been involved. It was this transmission that forked back to the Malay Archipelago, causing Amoghapāśa to be again venerated, under very different circumstances, as the totem of the rulers of the Malays. The worship of the eight-armed Amoghapāśa spread in this way from Melayu to Thamel and back. During each leap into a new environment, the bodhisattva lost much of its established identity but kept some of its earlier momentum; it transmigrated, so to speak. Amoghapāśa’s eight arms are intuitively associated in Mahāyāna contexts with the wish to be rescued from eight great fears, especially the fear of drowning. This association, having been ingrained 187. Ashar Murdihastomo (2019) identifies Southeast Asian ‘patron deities of seafaring’, partly with reference to Nepalese sources, in addition to Avalokiteśvara and Tārā, who are the focus of Bopearachchi 2014.

02_CreSouthV1_8P_4Feb22.indd 42

into the non-tantric cult of Avalokiteśvara in Western India, persisted over a long period and distance in the visual culture of Indo-Nepalese tantric Buddhism. Whereas the worship of Amoghapāśa offered personal salvation for Newars observing the fasting ritual, the notion of rescuing people in distress was readily extended to the defensive protection of a territory or a polity. The eight-armed Amoghapāśa manifested in this capacity in East Asia, the Khmer Empire and the Malayo-Javanese world up to the 14th century. Another role that recurs throughout the bodhisattva’s transmigrations—again, without any known, explicit textual justification—is that of an exoteric precursor to the esoteric praxis of Hevajra. Ultimately, the standing eight-armed Amoghapāśa, an enduring and well-travelled figure in the Sanskritic world, never occupied a place in ‘Indian’ Buddhism. Likewise, many vital aspects of the bodhisattva’s cultus are invisible in, and unknowable from, the Indo-Tibetan praxis literature. Much of what is recorded in the sādhanas of the bodhisattva penned by celebrated pundits such as Śākyaśrībhadra and Vibhūticandra, and for that matter in the handbooks of the anonymous Nepalese codifiers who succeeded them, is no more than surface detail. The fact that the eight-armed Amoghapāśa was worshipped so widely and for so long as an ‘undocumented’ figure underlines the limitations of Sanskritist, Indologist and textualist paradigms for meaningfully conveying the realities of Himalayan and Southeast Asian religious environments. Classical textual sources, however important they may be, are clarified primarily in the light shed by other vital inheritances of the past, such as artistic corpora, informal culture and living praxis traditions.

Abbreviations

CC BY 2.0: Creative Commons Attribution 2.0 Generic license. URL: https://creativecommons. org/licenses/by/2.0/. CC BY-NC-SA 2.0: CC BY-NC-SA 2.0. URL: https:// creativecommons.org/licenses/by-nc-sa/2.0/. CC BY-SA 4.0: Creative Commons Attribution-ShareAlike 4.0 International license. URL: https:// creativecommons.org/licenses/by-sa/4.0/deed. en.

3/2/22 6:16 PM

From Melayu to Thamel and Back D: Ui, Hakuju, Munetada Suzuki, and Enshō Kanakura. 1934. A Complete Catalogue of the Tibetan Buddhist Canons (Bkaḥ-ḥgyur and Bstan-ḥgyur). Sendai: Tōhoku Imperial University. Dpe bsdur ma: Zhongguo Zang xue yan jiu zhong xin. 1995–2007. Bstan-’gyur. Beijing: Krung-go’i Bod rig pa dpe skrun khang. HAR: Himalayan Art Resources. URL: https://www. himalayanart.org. ISEAS: Institute of Southeast Asian Studies, Singapore. Q: Suzuki, Daisetz T. (ed.). 1962. The Tibetan Tripitaka: Peking Edition kept in the Library of the Otani University, Kyoto. Catalogue & Index. Tokyo: Suzuki Research Foundation. T: The SAT Daizōkyō Text Database (SAT 大正 新脩大藏經テキストデータベース). URL: http://21dzk.l.u-tokyo.ac.jp/SAT/satdb2015.php.

References

Acri, Andrea. 2017. ‘“Local” vs. “Cosmopolitan” in the Study of Premodern Southeast Asia’, 수완나 부미 (SUVANNABHUMI): Multi-disciplinary Journal of Southeast Asian Studies 9: 7–52. URL: https://www.earticle.net/Article/A306798. . 2018. ‘The Place of Nusantara in the Sanskritic Buddhist Cosmopolis’, TRaNS: Trans -Regional and -National Studies of Southeast Asia 6/2: 139–166. http://doi.org/10.1017/trn.2018.5. . 2019. ‘Navigating the “Southern Seas”, Miraculously: Avoidance of Shipwreck in Buddhist Narratives of Maritime Crossings’, in Marina Berthet, Fernando Rosa, and Shaun Viljoen (eds.), Moving Spaces: Creolisation and Mobility in Africa, the Atlantic and Indian Ocean, pp. 50–77. Leiden: Brill. http://doi. org/10.1163/9789004410992_004. Adiceam, Marguerite E. 1966. ‘Les images de Śiva dans l’Inde du Sud [V. Harihara]’, Arts Asiatiques 13: 83–98. http://doi.org/10.3406/ arasi.1966.948. Ahmat Adam (ed.). 2016. Sulalat u’s-Salatin. Kuala Lumpur: Yayasan Karyawan. Ashar Murdihastomo. 2019. ‘Identifikasi dewa-dewi agama Hindu-Budha sebagai dewa pelindung pelayaran (Identification of Hindu-Buddhist Gods and Goddesses as Patron Deities of Sea-

02_CreSouthV1_8P_4Feb22.indd 43

43

faring)’, Naditira Widya 13/2: 87–104. http://doi. org/10.24832/nw.v13i2.397. Bambang Budi Utomo. 2013. ‘Arca-Arca Berlanggam Śailendra Di Luar Tanah Jawa’, Amerta: Jurnal Penelitian dan Pengembangan Arkeologi 31/1: 1–80. http://doi.org/10.24832/amt.v31i1.148. . 2016. Pengaruh Kebudayaan INDIA dalam Bentuk ARCA di Sumatra. Jakarta: Yayasan Pustaka Obor Indonesia. Bautze-Picron, Claudine. 2004. ‘The Universal Compassionate Bodhisattva: Miscellaneous Aspects of Avalokitasvara/Avalokiteśvara in India’, Silk Road Art and Archaeology 10: 225–290. . 2019. ‘A Short Notice on a Recently Discovered Lintel in Lakhisarai’, Journal of Bengal Art 24: 29–38. Béguin, Gilles. 2009. Buddhist Art: An Historical and Cultural Journey. Bangkok: River Books. Bendall, Cecil. 1883. Catalogue of the Buddhist Sanskrit Manuscripts in the University Library, Cambridge, with Introductory Notices and Illus. of the Palaeography and Chronology of Nepal and Bengal. Cambridge: Cambridge University Press. Bernet Kempers, A.J. 1959. Ancient Indonesian Art. Amsterdam: van der Peet. Bhattacharya, Buddhadev. 2016. Kāraṇḍavyūha sūtra: A Bi-lingual Critical Edition for the First Time from Sanskrit-Tibetan Manuscripts with an Introduction. New Delhi: Kaveri Books. Bhattacharya, Gouriswar. 2001. ‘Pretasantarpita-Lokeśvara’, Journal of Bengal Art 6: 21–44. Bhattacharya, Benoytosh (ed.). 1925–1928. Sādhanamālā. 2 vols. Gaekwad’s Oriental Series 26, 41. Baroda: Oriental Institute. . 1958. The Indian Buddhist Iconography, Mainly Based on the Sādhanamālā and Other Cognate Tāntric Texts of Rituals. Calcutta: Mukhopadhyay. Bhaṭṭācāryya, Satyavrata Sāmaśrami (ed.). [1872.] ‘Kāraṇḍavyūhaḥ. Mahāyānasūtram. Bauddhaśāstram. Mahāsatvāryyāvalokiteśvarabodhisatvasya prabhāvavarṇanāmukhena Śrīmadbhagavannāmaka-tathāgata-kathito dharmagranthaḥ. ‘Pratnakamranandinī’-sampādaka-śrīSatyavrata-Sāmaśrami-Bhaṭṭācāryyeṇa bahutarayatnataḥ sampādya nāgarākṣaraiḥ

3/2/22 6:16 PM

44

Iain Sinclair

prakāśitaḥ’, The Hindu Commentator: A Monthly Sanskrit Journal V(1). Śrīrāmapura. Boisselier, Jean. 1963. La statuaire du Champa: recherches sur les cultes et l’iconographie. Paris: École française d’Extrême-Orient. . 1965. ‘Précisions sur quelques images khmères d’Avalokiteśvara. Les bas-reliefs de Bantãy Čhmàr’, Arts Asiatiques 11/1: 73–89. http://doi.org/10.3406/arasi.1965.916. . 2008. ‘Identification of some Khmer images of Avalokiteśvara in the Bantãy Chmàr basreliefs’, in Natasha Eilenberg and Robert L. Brown (trans. and eds.), Studies on the Art of Ancient Cambodia: Ten Articles by Jean Boisselier, pp. 303–326. Phnom Penh: Reyum Publishing. Bopearachchi, Osmund. 2014. ‘Sri Lanka and Maritime Trade: Bodhisattva Avalokiteśvara as the Protector of Mariners’, in Parul Pandya Dhar and Upendra Singh (eds.), Asian Encounters: Exploring Connected Histories, pp. 161–187. New Delhi: Oxford University Press. . [2017]. ‘Sri Lanka and the Maritime Trade: Bodhisattva Avalokiteśvara as the Protector of Mariners’. PowerPoint presentation. URL: http://www.academia.edu/download/53316351/ Sri_Lanka_and_the_Maritime_Trade._Bodhisattva_Avalokitesvara_as_the_Protector_of_ Mariners.pdf. Bryson, Megan. 2018. ‘Nation Founder and Universal Saviour: Guanyin and Buddhist Networks in the Nanzhao and Dali Kingdoms’, in Christoph Anderl, Carmen Meinert, and Ann Heirman (eds.), Buddhist Encounters and Identities Across East Asia. Dynamics in the History of Religions 10, pp. 81–107. Leiden/Boston: Brill. http://doi. org/10.1163/9789004366152_005. Budi Istiawan. 2012. ‘New Finds of the Classical Period in West Sumatra’, in Mai Lin TjoaBonatz, Andreas Reinecke, and Dominik Bonatz (eds.), Connecting Empires and States: Selected Papers from the 13th International Conference of the European Association of Southeast Asian Archaeologists, pp. 43–52. Singapore: NUS Press. Bunker, Emma C. 2011. ‘Five-headed Shiva Images’, in Emma C. Bunker and Douglas Latchford,

02_CreSouthV1_8P_4Feb22.indd 44

Khmer Bronzes: New Interpretations of the Past, pp. 503–512. Chicago: Art Media Resources. Bunker, Emma C. and Douglas Latchford. 2011. Khmer Bronzes: New Interpretations of the Past. Chicago: Art Media Resources. Bu ston rin chen grub. c. 1360. Bu ston gsan yig (Bla ma dam par rnams kyis rjes su bzung ba’i tshul, bKa’ drin rjes su dran par byed pa). Collected Works 26: 7–149 (Buddhist Digital Resource Centre W1934). de Casparis, J.G. 1967. ‘The Date of the Grahi Buddha’, Journal of the Siam Society 55/1: 31–40. URL: www.siamese-heritage.org/jsspdf/1961/​ JSS_055_1c_Casparis_​DateOfGrahiBuddha.pdf. Chakraborti, Haripada. 1970. Pāśupata sūtram with Pañcārtha-bhāṣya of Kauṇḍinya. Translated with an Introduction on the history of Śaivism in India. Calcutta: Academic Publishers. Chambert-Loir, Henri. 2017. ‘The History of a History: Variant Versions of the Sulalat al-Salatin’, Indonesia 104: 121–177. http://doi. org/10.1353/ind.2017.0013. Chatterjee, Bijan Raj and Niranjan Prasad Chakravarti. 1933. India and Java. Part II (Inscriptions). Calcutta: Greater Indian Society. Chattopadhyaya, Alaka and Lama Chimpa. 1967. Atīśa and Tibet. [Calcutta]: Indian Studies: Past & Present. Chutiwongs, Nandana. 1984. The Iconography of Avalokiteśvara in Mainland South East Asia. Leiden: D. Litt. Diss. . 1994. ‘An Aspect of the Bodhisattva Avalokiteśvara in Ancient Indonesia’, in Marijke J. Klokke and Pauline Lunsingh Scheurleer (eds.), Ancient Indonesian Sculpture, pp. 98–114. Leiden: KITLV Press. Colless, Brian E. 1969. ‘Persian Merchants and Missionaries in Medieval Malaya’, Journal of the Malaysian Branch of the Royal Asiatic Society 42/2: 10–47. URL: https://www.jstor.org/ stable/41491992. Conti, Pia. 2014. ‘Tantric Buddhism at Prasat Hin Phimai: A New Reading of its Iconographic Message’, in Nicolas Revire and Stephen A. Murphy (eds.), Before Siam: Essays in Art and Archaeology, pp. 374–395. Bangkok: River Books and the Siam Society.

3/2/22 6:16 PM

From Melayu to Thamel and Back Cowell, E.B. and J. Eggeling. 1876. ‘Catalogue of the Buddhist Sanskrit Manuscripts in the Possession of the Royal Asiatic Society (Hodgson Collection)’, Journal of the Royal Asiatic Society (New Series) 8/1: 1–52. Dalsheimer, Nadine. 2001. Les collections du musée national de Phnom Penh: L’art du Cambodge ancien. Paris: École française d’Extrême-Orient and Magellan & Cie. Delhey, Martin. 2012. ‘Textual Sources of the Mañjuśriyamūlakalpa (Mañjuśrīmūlakalpa), With Special Reference to Its Early Nepalese Witness NGMPP A39/4’, Journal of the Nepal Research Centre XIV: 55–75. Dhingra, Sonali. 2022. ‘Saviour “at the Time of Death”: Amoghapāśa’s Cultic Role in Late First Millennium Odishan Buddhist Sites’, in Andrea Acri and Peter Sharrock (eds.), The Creative South: Buddhist and Hindu Art in Medieval Maritime Asia, Volume 2: Odisha and Java. Singapore: ISEAS – Yusof Ishak Institute. Diskul, M.C. Subhadradis. 1980. ‘II: Śrīvijaya Art in Thailand’, in M.C. Subhadradis Diskul (ed.), The Art of Śrīvijaya, pp. 21–43. Petaling Jaya: UNESCO Oxford University Press. Donaldson, Thomas E. 2001. Iconography of the Buddhist Sculpture of Orissa. 2 vols. New Delhi: Indira Gandhi National Centre for the Arts. Dutt, Nalinaksha. 1953. Saddharmapuṇḍarīkasūtram. With N. D. Mironov’s Readings from Central Asian MSS. Calcutta: The Asiatic Society. Edwards McKinnon, Edmund, and Ichwan Azhari. Forthcoming. ‘An Eight-armed Bronze Amoghapasa Image from Kota Rentang, Northern Sumatra’. Unpublished draft. Eilenberg, Natasha. 1997. ‘A South Indian Śiva Naṭarāja Bronze in the Collection of Robert Hatfield Ellsworth’, in Natasha Eilenberg, M.C. Subhadradis Diskul, and Robert L. Brown (eds.), Living a Life in accord with Dhamma: Papers in Honor of Professor Jean Boisselier on his Eightieth Birthday, pp. 162–175. Bangkok: Silpakorn University. Ery Soedewo. 2012. ‘Persamaan Dan Perbedaan Unsur Logam Benda-Benda Perunggu Sumatera Bagian Utara Dengan Benda-Benda Perunggu Jawa Tengah: Antara Kemandirian Teknis Dan

02_CreSouthV1_8P_4Feb22.indd 45

45

Pemenuhan Konsep (Similarities and Differences on Metal Attribute of Bronze Artifacts from Northern Sumatera and Central Java: Between Technical Autonomy and Conceptual Completion)’, Berkala Arkeologi 32/2: 151–166. http://doi. org/10.30883/jba.v32i2.54. Ferlus, Michel. 2009. ‘Formation of Ethnonyms in Southeast Asia’. Paper presented at the 42nd International Conference on Sino-Tibetan Languages and Linguistics, Payap University, Chiang Mai, Thailand, November 2009. URL: https://halshs.archives-ouvertes.fr/halshs01182596/. Ferrand, G. 1922. L’empire sumatranais de Çrīvijaya. Paris: Paul Geuthner. Fontein, Jan. 1990. The Sculpture of Indonesia. Washington, D.C. and New York: National Gallery of Art and H.N. Abrams. Goudriaan, T. and C. Hooykaas. 1971. Stuti and Stava (Bauddha, Śaiva and Vaiṣṇava) of Balinese Brahman Priests. Amsterdam/London: North-Holland Publishing Company. Graldi, Aurora. 2019. Transforming the Buddha: A Biography of the Monumental Buddha of Guita Bahī, Lalitpur, and its Implication for the Nepalese School of Sculpture in Metal. PhD dissertation, University of Vienna. Green, Phillip S.E. 2014. ‘The Many Faces of Lokeśvara: Tantric Connections in Cambodia and Campā between the Tenth and Thirteenth Centuries’, History of Religions 54/1: 69–93. http://doi.org/10.1086/676513. Griffiths, Arlo. 2018. ‘The Corpus of Inscriptions in the Old Malay Language’, in Daniel Perret (ed.), Writing for Eternity: A Survey of Epigraphy in Southeast Asia, pp. 277–283. Paris: École française d’Extrême-Orient. van Gulik, R.H. 1935. Hayagrīva: The Mantrayānic Aspect of Horse-cult in China and Japan. Leiden: E.J. Brill. Guy, John. 1995. ‘The Avalokiteśvara of Yunnan and some South East Asian Connections’, in Rosemary Scott and John Guy (eds.), South East Asia & China: Art, Interaction & Commerce. Colloquies on Art & Archaeology in Asia No. 17, pp. 64–83. London: School of Oriental and African Studies, Percival David Foundation of Chinese Art.

3/2/22 6:16 PM

46

Iain Sinclair

. (ed.). 2014. Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia. Bangkok and New York: River Books and the Metropolitan Museum of Art. Hahn, Michael. 2014. ‘Śākyarakṣita’s Vṛttamālāvivṛti on Jñānaśrīmitra’s Vṛttamālāstuti (2)’, South Asian Classical Studies 9: 1–64. Hall, Kenneth R. 2000. ‘Personal Status and Ritualized Exchange in Majapahit Java’, Archipel 59: 51–96. http://doi.org/10.3406/arch.2000.3554. Harris, Mark and Zainuddin Zainal. 1990. National Museum, Kuala Lumpur: History and Culture of Malaysia. Kuala Lumpur: The Pepin Press. Heine-Geldern, Robert. 1925. Altjavanische Bronzen: aus dem Besitze der ethnographischen Sammlung des Naturhistorischen Museums in Wien. Artis thesaurus Band 1. Wien: C.W. Stern. Hṛdaya: Amoghapāśahṛdaya nāma mahāyānasūtra, ed. Meisezahl (1962). Indian Museum and Theresa McCullough. 2015. Treasures from Asia’s Oldest Museum: Buddhist Art from the Indian Museum, Kolkata. Singapore: Asian Civilisations Museum. Jacq-Hergoualc’h, Michel. 2002. The Malay Peninsula: Crossroads of the Maritime Silk Road (100 BC–1300 AD). Leiden: Brill. http://doi. org/10.1163/9789047400684. Jacques, Claude. 2005. ‘Deux problèmes posés par l’inscription de la stèle de Praḥ Khan K. 908’, Aséanie 15: 11–31. http://doi.org/10.3406/ asean.2005.1844. Kalparāja: Amoghapāśakalparāja. Sanskrit MS: Mikkyō Seiten Kenkyūkai 1996. Partial transcripts include: Mikkyō Seiten Kenkyūkai 2001; Sasaki, Itō and Yamamoto 2010. Tibetan translation: D 686/Q 365. Chinese translation: T 1092. Kitchener, Darrell J. and Heny Kustiarsih. 2019. Ceramics from the Musi River, Palembang, Indonesia: Based on a Private Collection. Fremantle: Australian National Centre of Excellence for Maritime Archaeology. Kozok, Uli and Eric van Reijn. 2010. ‘Ādityawarman: Three Inscriptions of the Sumatran “King of all Supreme Kings”’, Indonesia and the Malay World 38/110: 135–158. http://doi. org/10.1080/13639811003665488. Krairiksh, Piriya. [1980.] Art in Peninsular Thailand prior to the Fourteenth Century A.D. Bangkok: The Fine Arts Department.

02_CreSouthV1_8P_4Feb22.indd 46

Krairiksh, Piriya; Narisa Chakrabongse (trans.). 2012. The Roots of Thai Art. Bangkok: River Books. Krishnan, Gauri Parimoo. 2016. ‘The Roots and Legacy of the Art of Nalanda as seen at Srivijaya’, in Gauri Parimoo Krishnan (ed.), Nalanda, Srivijaya and Beyond: Re-exploring Buddhist Art in Asia, pp. 153–200. Singapore: Asian Civilisations Museum. Krom, N.J. 1917. ‘Een sumatraansche Inscriptie van Köning Kṛtanagara’, Verslagen en Mededeelingen der Koninklijke Akademie van Wetenschappen 5de reeks, deel II, Amsterdam, 1916: 306–344. Kwa, Chong Guan. 1967–1968. ‘The Bidor Bronze Once More’, Journal of the Historical Society, University of Singapore, 1967–8: 62–64. . 2010. ‘The Buddhist Origins of “Singapura”: Rereading Chapters I–VI of the Sejarah Melayu on Singapore’. Presentation at the Nalanda-Sriwijaya Centre, Institute of Southeast Asian Studies, 4 March 2010. Lee, Yong-Hyun (이용현). 2004. The Niṣpannayogāvalī by Abhayākaragupta: A New Critical Edition of the Sanskrit Text. Seoul: Baegun Press. Lefèvre, Vincent. 2007. Chefs-d’oeuvre du delta du Gange—Collections des musées du Bangladesh: Établissement Public du Musée des Arts Asiatiques Guimet, 24 octobre 2007–3 mars 2008. Paris: Éd. de la Réunion des musées nationaux. Leoshko, Janice. 1985. ‘The Appearance of Amoghapāśa in Pāla Period Art’, in A.K. Narain (ed.), Studies in Buddhist Art of South Asia, pp. 127–135. New Delhi: Kanak Publications. Li, Yu-min (李玉珉). 2010. ‘Dali guo Zhang Shengwen Fanxiang juan luohan hua yanjiu 大理國張勝 溫《梵像卷》羅漢畫研究 (A Study of the Arhat Paintings in “Handscroll of Buddhist Images” by Zhang Shengwen of the Dali Kingdom)’, Taida Journal of Art History 29: 113–140. Lokesh Chandra. 1999. Dictionary of Buddhist Iconography. Vol. 1. New Delhi: International Academy of Indian Culture and Aditya Prakashan. Lunsingh Scheurleer, Pauline. 2008. ‘The Wellknown Javanese Statue in the Tropenmuseum, Amsterdam, and its place in Javanese Sculpture’,

3/2/22 6:16 PM

From Melayu to Thamel and Back Artibus Asiae 68/2: 287–332. URL: https://www. jstor.org/stable/40599602. Lunsingh Scheurleer, Pauline and Marijke J. Klokke. 1988. Divine Bronze: Ancient Indonesian Bronzes from A.D. 600 to 1600. Leiden: Brill. Macdonald, Alexander W. and Anne Vergati Stahl. 1979. Newar Art: Nepalese Art during the Malla Period. Warminster: Aris & Phillips Ltd. Macdonald, Ariane. 1970. ‘IX. Le Dhānyakataka de Man-luṅs guru’, Bulletin de l’École française d’Extrême-Orient 57: 169–213. http://doi. org/10.3406/befeo.1970.5416. Mallmann, Marie Thérèse de. 1951. ‘Notes sur les bronzes du Yunnan représentant Avalokiteśvara’, Harvard Journal of Asiatic Studies 14/3–4: 567–601. http://doi.org/10.2307/2718187. Manabe, Shunsho (真鍋 俊照). 2016. ‘Fukūkenjaku Kannon no hyōgen to sono fuhensei 不空羂索 観音の表現とその普遍性 (The Representation of Fukūkenjaku kannon and Its Universality)’, (Abstract.) Journal of Indian and Buddhist Studies 65/1: 292–294. http://doi.org/10.4259/ ibk.65.3_1317. Manguin, Pierre-Yves. 1992. ‘Excavations in South Sumatra, 1988–90: New Evidence for Sriwijayan sites’, in Ian Glover (ed.), Southeast Asian Archaeology 1990: Proceedings of the Third Conference of the European Association of Southeast Asian Archaeologists, pp. 63–73. Hull: Centre for South-East Asian Studies, University of Hull. . 2017. ‘At the Origins of Sriwijaya: The Emergence of State and City in Southeast Sumatra’, in Noboru Karashima and Masashi Hirosue (eds.), State Formation and Social Integration in Pre-modern South and Southeast Asia: A Comparative Study of Asian Society, pp. 89–114. Tokyo/New Delhi: Toyo Bunko/Manohar. URL: http://id.nii.ac.jp/1629/00003625/. Matsumura, Hisashi (松村 恒). 2011. ‘Sanbō kanō yōryaku roku no kōsei to kōzoku bunken no sono riyō 『三宝感応要略録』の構成と後続 文献のその利用 (The Structure of the Sanbao ganying yaoluelu and its Influence on Later Literature)’, Journal of Indian and Buddhist Studies 60/1: 260–267. http://doi.org/10.4259/ ibk.60.1_260. Maxwell, Thomas S. 2007. ‘The Stele Inscription of Preah Khan, Angkor. Text with Translation

02_CreSouthV1_8P_4Feb22.indd 47

47

and Commentary’, Udaya: Journal of Khmer Studies 8: 1–113. McKeown, Arthur. 2010. From Bodhgayā to Lhasa to Beijing: The Life and Times of Śāriputra (c.1335–1426), Last Abbot of Bodhgayā. PhD dissertation, Harvard University. Mechling, Mathilde, Brice Vincent, Pierre Baptiste, and David Bourgarit. 2018. ‘The Indonesian Bronze-Casting Tradition: Technical Investigations on Thirty-Nine Indonesian Bronze Statues (7th–11th c.) from the Musée National des Arts Asiatiques – Guimet, Paris’, Bulletin de l’École française d’Extrême-Orient 104: 63–139. http:// doi.org/10.3406/befeo.2018.6270. Meisezahl, R.O. 1962. ‘The Amoghapāśahṛdaya-dhāraṇī: The Early Sanskrit Manuscript of the Reiunji Critically Edited and Translated’, Monumenta Nipponica 17/1: 265–328. http://doi. org/10.2307/2383264. . 1967. ‘Amoghapāśa: Some Nepalese Representations and Their Vajrayānic Aspects’, Monumenta Serica 26: 455–497. http://doi.org/ 10.1080/02549948.1967.11744976. Melzer, Gudrun and Eva Allinger. 2012. ‘Die nepalesische Palmblatthandschrift der Aṣṭasāhasrikā Prajñāpāramitā aus dem Jahr NS 268 (1148 AD), Teil II’, Berliner Indologische Studien 20: 249–276. Mette, Adelheid (ed.). 1997. Die Gilgitfragmente des Kāraṇḍavyūha. Swistal-Odendorf: Indica et Tibetica Verlag. Mette, Adelheid, Noriyuki Kudo, Ruriko Sakuma, Chanwit Tudkeao, and Jiro Hirabayashi (eds.). 2017. Gilgit Manuscripts in the National Archives of India, Facsimile Edition. Further Mahāyānasūtras vol. II. 4. New Delhi and Tokyo: The National Archives of India, The International Research Institute for Advanced Buddhology at Soka University. Mikkyō Seiten Kenkyūkai (密教聖典研究会) (ed.). 1996. Fukū kenjaku jinpen shingonkyō bonbun shahon ei’inban 不空羂索神変真言経梵文写本 影印版. Tokyo: Taishō Daigaku Sōgō Bukkyō Kenkyūjo 大正大学綜合仏教研究所. . 2001. ‘Transcribed Sanskrit Text of the Amoghapāśakalparāja Part IV’, Annual of the Institute for Comprehensive Studies of Buddhism, Taisho University 23: 1–73.

3/2/22 6:16 PM

Iain Sinclair

48

. 2004. Fukūkenjaku jinben shingongyō jumon shūsei: Bon-Kan taishō 『不空羂索神變 眞言經』呪文集成: 梵漢対照. Tokyo: Nonburu (ノンブル). Miyaji, Akira. 2000. ‘A Bas-Relief of Eight-Armed Avalokiteśvara as Saviour of Men, probably from Swāt, Pakistan’, in European Association of South Asian Archaeologists, Maurizio Taddei and Giuseppe De Marco (eds.), South Asian Archaeology, 1997: Proceedings of the Fourteenth International Conference of the European Association of South Asian Archaeologists, held in the Istituto Italiano per l’Africa e l’Oriente, Palazzo Brancaccio, Rome, 7–14 July 1997, pp. 1249–1266. Rome: Istituto Italiano per l’Africa e l’Oriente. Miyazaki, Akiko (宮﨑 晶子). 2010. ‘The Roles of Bodhisattva Avalokiteśvara in Angkor: Focusing on Descriptions on the Kāraṇḍavyūha-sūtra’, The Journal of Sophia Asian Studies 28: 27–48. URI: http://digital-archives.sophia.ac.jp/repository/view/repository/​00000026490. . 2019. ‘Ankōru no Kannon zō ni miru kyōten to zuzō no kankeisei: Kārandavyūha shahon to no hikaku kara (アンコールの観音像に見る経 典と図像の関係性―『カーランダヴューハ』写 本との比較から)’, in Takashi Koezuka 肥塚隆

(ed.), Ajia Bukkyō bijutsu ronshū: Tōnan Ajia (アジア仏教美術論集 東南アジア), pp. 353– 374. Tokyo: Chūō Kōron Bijutsu Shuppan 中央 公論美術出版社. Moeller, Volker. 1985. Javanische Bronzen. Berlin: Mann. Mori, Masahide (森 雅秀). 1997. ‘Orissa shūritsu hakubutsukan no mikkyō bijutsu オリッサ州 立博物館の密教美術’, Bulletin of the Research Institute of Esoteric Buddhist Culture 10: 191–232. Murphy, Stephen A., Naomi Wang, and Alexandra Green. 2019. Raffles in Southeast Asia: Revisiting the Scholar and Statesman. Singapore: Asian Civilisations Museum. Nakamura, Kayō (中村 夏葉). 2008. ‘Chūgoku Tonkō ni okeru Fukūkenjaku Kannon ni kan suru kenkyū (中国敦煌における不空羂索観 音像に関する研究)’, Μεταπτυχιακά: Annual of the Office for the Promotion of Graduate Educational Research (名古屋大学大学院文学 研究科教育研究推進室年報) 2: 224–228. http:// doi.org/10.18999/meta.2.224.

02_CreSouthV1_8P_4Feb22.indd 48

Nihom, Max. 1994. Studies in Indian and Indo-Indonesian Tantrism: The Kuñjarakarṇadharmakathana and the Yogatantra. Publications of the De Nobili Research Library XXI. Vienna: Sammlung De Nobili Institut für Indologie der Universität Wien. Nik Hassan Shuhaimi. 1976. Buddhist Sculpture from Sumatra, Peninsular Malaysia and Peninsular Thailand during the Srīvijayan period (7th–14th c. A.D.). M.Phil thesis, University of London, School of Oriental and African Studies. . 1977. ‘The Significance of Tiger Skin on the Avalokitesvara Images’, Malaysia in History XX/1: 24–31. URL: http://myrepositori.pnm.gov. my/handle/123456789/3605. Pal, Pratapaditya. 1966. ‘The Iconography of Amoghapāśa Lokeśvara—I’, Oriental Art 12: 234–239. . 1967. ‘The Iconography of Amoghapāśa Lokeśvara—II’, Oriental Art 13: 20–28. . 2001. Desire and Devotion: Art from India, Nepal, and Tibet in the John and Berthe Ford Collection. Baltimore: Walters Art Museum. Park, Hyounggook (朴亨國). 2019. ‘Tōnan Ajia ni okeru tapi Kannon no zuzō ni tsuite 東南アジ アにおける多臂観音の図像について [On the Iconography of Multi-armed Avalokiteśvaras in Southeast Asia]’, in Takashi Koezuka 肥塚隆 (ed.), Ajia Bukkyō bijutsu ronshū: Tōnan Ajia (アジア仏教美術論集 東南アジア), pp. 287– 319. Tokyo: Chūō Kōron Bijutsu Shuppan 中央公論美術出版社. Paul, Debjani. 1987. The Art of Nālandā: Development of Buddhist Sculpture 600–1200 A.D. Leiden: Quick Service Drukkerij (proefschrift). Pigeaud, Theodore G. 1960. Java in the 14th Century: A Study in Cultural History. The Nāgara-Kĕrtāgama by Rakawi Prapañca of Majapahit, 1365 A.D. Vol. I: Javanese texts in transcription. 3rd ed. The Hague: Martinus Nijhoff. Pou, Saveros (ed. and trans.). 2002. Nouvelles inscriptions du Cambodge, II & III. Paris: École française d’Extrême-Orient. Quaritch Wales, H.G. 1940. ‘Archaeological Researches on Ancient Indian Colonization in Malaya’, Journal of the Malayan Branch of the Royal Asiatic Society XVIII Part I: i–xiii, 1–85.

3/2/22 6:16 PM

From Melayu to Thamel and Back . 1976. The Malay Peninsula in Hindu Times. London: Bernard Quaritch Ltd. Reichle, N. 2007. Violence and Serenity: Late Buddhist Sculpture from Indonesia. Honolulu: University of Hawai‘i Press. Reis-Habito, Maria. 1999. ‘The Amoghapāśa Kalparāja: A Historical and Analytical Study’, Studies in East and Central Asian Religions 11: 39–67. Roberts, Peter Alan and Tulku Yeshi. 2013–2016. ’phags pa za ma tog bkod pa zhes bya ba theg pa chen po’i mdo: The Noble Mahāyāna Sūtra ‘The Basket’s Display’. Āryakāraṇḍavyūhanāmamahāyānasutra. 84000: Translating the Words of the Buddha. Degé Kangyur, vol. 51 (mdo sde, pa), folios 200.a-247.b. v 2.41. URL: http:// read.84000.co/translation/toh116.html. Robson, Stuart O. (trans.); Mpu Prapañca. 1995. Deśawarṇana (Nāgarakṛtāgama). Verhandelingen van het Koninklijk Instituut voor Taal-, Land- en Volkenkunde, 169. Leiden: KITLV Press. Sakuma, Ruriko (佐久間 留理子). 2002. Sādhanamālā: Avalokiteśvara section. Sanskrit and Tibetan texts. Delhi: Adroit. . 2006. ‘A Historical Background of the Trailokyavaśaṅkarabhugma Lokeśvara sādhana in the Sādhanamālā’. Nagoya Studies in Indian Culture and Buddhism: Saṃbhāṣā 25: 1–13. http://doi.org/10.18999/sambh.25.1. . 2007. ‘Kanjizai bosatsu no shosetsu to bijutsu: Dai-Ei Toshokan zō Kārandavyūha sōshuku shahon o chūshin ni tsuite 観自在菩 薩の説話と美術―大英図書館蔵『カーランダ˙ ヴューハ』装飾写本を中心として’, in Miyaji Akira sensei kenti ronbunshū henshū iinkai 宮 治昭先生献呈論文集編集委員会 (eds.), Han ajia no bukkyō bijutsu 汎アジアの仏教美術, pp.

172–195. Tokyo: Chūōkōron Bijutsu Shuppan (中央公論美術出版). . 2015. Kannon bosatsu: hengen jizai na sugata o toru kyūsaisha (観音菩薩: 変幻自在な 姿をとる救済者). Tokyo: Shunjusha 春秋社. . 2017. ‘Nepāru no Fukūkenjaku Kanjizai mandara—giki to nunosei kaiga ネパールの不 空羂索観自在マンダラ―儀軌と布製絵画―

(Maṇḍala of Amoghapāśa-lokeśvara in Nepal:

02_CreSouthV1_8P_4Feb22.indd 49

49

Its Ritual Texts and Cloth Paintings)’, Journal of Indian and Buddhist Studies 65/2: 892–886. http://doi.org/10.4259/ibk.65.2_892. . 2019. ‘Appearances of Avalokiteśvara in the Kāraṇḍavyūha-sūtra: Their Functions and Religious Backgrounds’, Nagoya Studies in Indian Culture and Buddhism: Saṃbhāṣā 35: 73–97. Sasaki, Daiju (佐々木 大樹), Gyōkan Itō (伊藤 堯貫), and Shoichiro Yamamoto (山本 匠一 郎). 2010. ‘Transcribed Sanskrit Text of the Amoghapāśakalparāja Part VI’, Annual of the Institute for Comprehensive Studies of Buddhism, Taisho University 32: 170–207. Schaik, Sam van. 2006. ‘The Tibetan Avalokiteśvara Cult in the Tenth Century: Evidence from the Dunhuang Manuscripts’, in Ronald M. Davidson and Christian Wedemeyer (eds.), Tibetan Buddhist Literature and Praxis. Proceedings of the Tenth Seminar of the IATS, 2003, Vol. 4, pp. 55–72. Leiden: Brill. Schlüter, Lisa. 2010. ‘Materials for the Medieval History of Indonesia: Java1365.Nagara-Kertagama.Canto.38.3-49’. Hamburg: Universität Hamburg. URL: http://www.spaetmittelalter. uni-hamburg.de/java-history/JavaNK/Java1365. Nagara-Kertagama.Canto.38.3-49.html. Schoterman, Jan A. 1994. ‘A Surviving Amoghapāśa Sādhana: Its Relation to the Five Main Statues of Candi Jago’, in Marijke J. Klokke and Pauline Lunsingh Scheurleer (eds.), Ancient Indonesian Sculpture, pp. 154–175. Leiden: KITLV Press. Seki, Chisato (関千里). 1996. Tōnan Ajia no kobijutsu: sono miryoku to rekishi 東南アジアの古 美術: その魅力と歴史 (South East Asian Arts: From Ancient Time to 17th Century). Tokyo: Mekon (めこん). Sen, Tansen. 2009. ‘The Military Campaigns of Rajendra Chola and the Chola-Srivijaya-China Triangle’, in Hermann Kulke, K. Kesavapany, and Vijay Sakhuja (eds.), Nagapattinam to Suvarnadwipa: Reflections on the Chola Naval Expeditions to Southeast Asia, pp. 61–75. Singapore: Institute of Southeast Asian Studies. Shakya, Miroj (ed.). 2010. Catalogue of Digitized Rare Sanskrit Buddhist Manuscripts. Vol. 1: 1–200. Rosemead: University of the West.

3/2/22 6:16 PM

50

Iain Sinclair

Sharrock, Peter. 2007. ‘The Mystery of the Bayon Face Towers’, in Joyce Clark (ed.), Bayon: New Perspectives, pp. 230–281. Bangkok: River Books. . 2013. ‘The Tantric Roots of the Buddhist Pantheon of Jayavarman VII’, in Marijke J. Klokke and Véronique Degroot (eds.), Materializing Southeast Asia’s Past: Selected Papers from the 12th International Conference of the European Association of Southeast Asian Archaeologists. Vol. 2, pp. 41–55. Singapore: NUS Press. . 2015. Banteay Chhmar: Garrison-temple of the Khmer Empire. Bangkok: River Books. Shinohara, Koichi. 2010. ‘The All-Gathering Maṇḍala Initiation Ceremony in Atikūṭa’s Collected Dhāraṇī Scriptures: Reconstructing the Evolution of Esoteric Buddhist Ritual’. Journal Asiatique 298/2: 389–420. http://doi.org/10.2143/ JA.298.2.2062439. . 2014. ‘Dhāraṇīs and Visions in Early Esoteric Buddhist Sources in Chinese Translation’, Bulletin of the School of Oriental and African Studies 77/1: 85–103. http://doi. org/10.1017/S0041977X13000931. Siddhanta, Sachindra Nath. 1979. A Descriptive Catalogue of Sanskrit Manuscripts in the Varendra Research Museum Library. Vol. I. Rajshahi: Varendra Research Museum. Sinclair, Iain. 2015. ‘The Creation of Theism Personified: A Conceptual History of the God-Maker Avalokiteśvara’, in Andrea A. Di Castro and David Templeman (eds.), Asian Horizons: Giuseppe Tucci’s Buddhist, Indian, Himalayan and Central Asian Studies, pp. 431–478. Melbourne: Monash University Publishing. . 2016a. ‘Coronation and Liberation According to a Javanese Monk in China: Bianhong’s Manual on the abhiṣeka of a cakravartin’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia, pp. 30–66, 389–391. Singapore: ISEAS – Yusof Ishak Institute. http:// doi.org/10.1355/9789814695091-005. . 2016b. The Appearance of Tantric Monasticism in Nepal: A History of the Public Image and Fasting Ritual of Newar Buddhism, 980– 1380. PhD dissertation, Monash University. http://doi.org/10.4225/03/58ab8cadcf152.

02_CreSouthV1_8P_4Feb22.indd 50

. 2017. Addenda and Errata. [s.p.] http://doi. org/10.6084/m9.figshare.13387175.v3. . 2018. ‘New Light on the Karimun Besar Inscription (Prasasti Pasir Panjang) and the Learned Man from Gaur’, NSC Highlights 11: 16–17. URL: http://hdl.handle.net/11540/9401. . 2019a. ‘Sang Sapurba/Maulivarmadeva, First of the Last Indo-Malay Kings’, NSC Highlights 12: 6–8. URL: http://hdl.handle. net/11540/9986. . 2019b. ‘Sanskritic Buddhism as an Asian Universalism’, in Andrea Acri, Kashshaf Ghani, Murari K. Jha and Sraman Mukherjee (eds.), Imagining Asia: Networks, Actors, Sites, pp. 275–333. Singapore: ISEAS – Yusof Ishak Institute. http://doi.org/10.1355/9789814818865-012. . 2019c. ‘Saviour of the South Seas: A Srivijayan Bodhisattva and its Transasiatic Journey’. Public lecture. Singapore: Asian Civilisations Museum, 16 August 2019. . Forthcoming. ‘Gautamaśrī: A Bengali Buddhist pundit in transit through the Himalayas and the Singapore Strait’. Unpublished draft. Skilling, Peter. 2019. ‘Many Lands of Gold’, in Bunchar Pongpanich and Somchet Thinapong (eds.), Suvarnabhumi: The Golden Land, pp. 191–217. Bangkok: Geo-Informatics and Space Technology Development Agency (GISTDA) and Buddhadasa Indapanno Archives (BIA). Soper, Alexander C. and Helen B. Chapin. 1970. ‘A Long Roll of Buddhist Images. I’, Artibus Asiae 32/1: 5–41. http://doi.org/10.2307/3249551. Sørensen, Henrik H. 2020. ‘Donors and Esoteric Buddhism in Dunhuang during the Reign of the Guiyijun’, in Henrik Sørensen and Carmen Meinert (eds.), Buddhism in Central Asia I. Patronage, Legitimation, Sacred Space, and Pilgrimage. Dynamics in the History of Religions 11, pp. 91–122. Leiden: Brill. http://doi. org/10.1163/9789004417731_006. Speyer, J.S. 1904a. ‘Eene Buddhistische inscriptie afkomstig van Java’, Verslagen en mededeelingen der Koninklijke Akademie van Wetenschappen. Afdeeling Letterkunde 4/6: 138–144. . 1904b. ‘Eene Buddhistische inscriptie afkomstig van Java (2de artikel)’, Verslagen en mededeelingen der Koninklijke Akademie van

3/2/22 6:16 PM

From Melayu to Thamel and Back Wetenschappen. Afdeeling Letterkunde 4/6: 253–262. Sri Ambarwati Kusumadewi. 2012. ‘Ādityawarman (1347–1374 Masehi): Kajian Epigrafi’. Master Humaniora, Universitas Indonesia. Srinivasan, Sharada. 2013. ‘Techno-cultural Perspectives on Medieval Southeast Asia and Southern India: Pallava Bronzes and Beyond’, in Marijke J. Klokke and Véronique Degroot (eds.), Materializing Southeast Asia’s Past: Selected Papers from the 12th International Conference of the European Association of Southeast Asian Archaeologists, vol. 2, pp. 167–178. Singapore: NUS Press. Stearns, Cyrus. 1996. ‘The Life and Tibetan Legacy of the Indian Mahāpaṇḍita Vibhūticandra’, Journal of the International Association of Buddhist Studies 19/1: 127–171. http://doi.org/10.2143/ JIABS.19.1.3285951. Suzuki, Koshin (鈴木 晃信), Nobuo Ōtsuka (大塚 伸夫), and Hideaki Kimura (木村 秀明). 2000. ‘Transcribed Sanskrit Text of the Amoghapāśakalparāja, Part III’, Annual of the Institute for Comprehensive Studies of Buddhism, Taisho University 22: 1–64. Takeuchi, Kōzen (武内 孝善). 1976. ‘Sōdai hon’yaku kyōten no tokushoku ni tsuite fu·Sōdai hon’yaku kyōten hennen mokuroku 宋代翻訳経典の 特色について附˙宋代翻訳経典編年目録 (On the Translation of Sutras in “Sung” (宋) period. Appendix: A Chronological Catalogue of the Translations)’, Journal of Esoteric Buddhism 113: 27–53. http://doi.org/10.11168/jeb1947.1976.27. Tanaka, Kimiaki (田中 公明). 2010. Indo ni okeru mandara no seiritsu to hatten (インドにおけ る曼荼羅の成立と発展). Tokyo: Shunjūsha (春秋社). . 2016a. ‘“Indo no butsu” ten ni shuppin sareta ‘Hassenju Hannyakyo’ sōshoku shahon ni tsuite 「インドの仏」展に出品された『八千 頌般若経』装飾写本について (The Sanskrit Manuscript of the Aṣṭasāhasrikā-Prajñāpāramitā Exhibited at the Exhibition “Indian Buddhist Art, from Indian Museum, Kolkata” held at Tokyo National Museum)’, The Memoirs of the Institute for Advanced Studies on Asia 169: viii, 117–130. http://doi.org/10.15083/00026810.

02_CreSouthV1_8P_4Feb22.indd 51

51

. 2016b. ‘Orissa hatsugen no mandara-teki kōzō o motta Chundā (Shundai) zō ni tsuite オ

リッサ発現の曼荼羅的構造をもったチュンダー (准提)像について (The Image of Cundā, which

has a Maṇḍala-like Composition, from Orissa)’, The Memoirs of the Institute for Advanced Studies on Asia 170: iii–iv, 399–410. http://doi. org/10.15083/00026797. . 2018. Bonbun monju kongō kuden kenkyū 梵文「文殊金剛口伝」研究(Mañjuvajramukhyākhyāna: Introduction, Romanized Sanskrit Text and Related Articles). Tokyo: Watanabeshuppan (渡辺出版). . 2019. An Illustrated History of the Mandala: From its Genesis to the Kalacakratantra. Boston: Wisdom Publications. Taniguchi Kōsei (谷口 耕生). 2002. ‘Nara Kokuritsu Hakubutsukan zō kenpon choshoku Fukūkensaku Kannon zō 奈良国立博物館蔵—絹本著色 不空羂索観音像 (On the Nara National Museum’s Buddhist Painting Fukūsensaku Kannon)’, Bulletin of the Nara National Museum 4: 59–70. http://doi.org/10.24737/00000154. Tobi, A.C. 1931. ‘De Buddhistische Bronzen in het Museum te Leiden’, Oudheidkundig Verslag 1930, pp. 158–201. Batavia-Centrum: Albrecht & Co. Tripathi, Ram Shankar and Thakur Sain Negi. 2006. Hevajratantram with Yogaratnamālāpañjikā of Mahāpaṇḍitācārya Kṛṣṇapāda. Bibliotheca Indico-Tibetica series 65. Sarnath: Central Institute of Higher Tibetan Studies. Tuladhar-Douglas, Will. 2006. Remaking Buddhism for Medieval Nepal: The Fifteenth-century Reformation of Newar Buddhism. London: Routledge. Vaidya, P.L. (ed.). 1961. Mahāyānasūtrasaṃgraha. Part I. Darbhanga: The Mithila Institute. . 1963. Saddharmalaṅkāvatārasūtram. Darbhanga: The Mithila Institute. Vajrācārya, Badrīratna. [2005.] Āryāvalokiteśvarayā bratayā utpatti va saṃkṣipta maṇḍala-bhāvana. Yeṃ: Vajrācārya Saṃrakṣaṇa Guṭhī. Vajrācārya, Herākāzī. 1999. Guitaḥyā Bahi vihāra saṃgha (Gustala mahāvihāra, Prathamaśrī mahāvihāra, Vasūccaśīla mahāvihāra) chagu adhyayana. Guitaḥ, Yala: Dīpāvatī Pustakālaya. Vetturini, Gianpaolo. 2007. The bKa’ gdams pa School of Tibetan Buddhism. 2 vols. PhD disser-

3/2/22 6:16 PM

52

Iain Sinclair

tation, University of London, School of Oriental and African Studies. Wade, Geoff. 2014. ‘Beyond the Southern Borders: Southeast Asia in Chinese Texts to the Ninth Century’, in John Guy (ed.), Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia, pp. 25–31. Bangkok and New York: River Books and Metropolitan Museum of Art. Wattanavrangkul, Snong. 1975. Outstanding sculptures of Buddhist and Hindu gods from private collections in Thailand. [Bangkok]: Rianthong Kanpim. Willson, Martin and Martin Brauen. 2000. Deities of Tibetan Buddhism: The Zürich Paintings of the Icons worthwhile to see: bris sku mthoṅ ba don ldan. Boston, MA: Wisdom Publications. Wong, Dorothy (王靜芬). 2015. ‘Bukong juansuo Guanyin xin tan 不空羂索觀音新探’, Journal of the Dunhuang and Turfan Studies 15: 17–52. Woodward, Hiram. 1994–95. ‘The Jayabuddhamahānātha Images of Cambodia’, The

02_CreSouthV1_8P_4Feb22.indd 52

Journal of the Walters Art Gallery 52/53: 105–111. URL: https://www.jstor.org/stable/20169099. . 2003. The Art and Architecture of Thailand: From Prehistoric Times through the Thirteenth Century. Leiden: Brill. . 2007. ‘The Karandavyuha Sutra and Buddhist Art in 10th Century Cambodia’, in Pratapaditya Pal (ed.), Buddhist Art: Form and Meaning, pp. 70–83. Mumbai: Marg Publications. . 2015. ‘Aspects of Buddhism in Tenth-century Cambodia’, in D. Christian Lammerts (ed.), Buddhist Dynamics in Premodern and Early Modern Southeast Asia, pp. 218–260. Singapore: Institute of Southeast Asian Studies. Wyatt, David K. 2001. ‘Relics, Oaths and Politics in Thirteenth-Century Siam’, Journal of Southeast Asian Studies 32/1: 3–65. http://doi.org/10.1017/ S0022463401000017. Yoritomi, Motohiro (頼富本宏). 1990. Mikkyōbutsu no kenkyū 密教仏の研究 (The Evolution of Esoteric Buddhism). Kyoto: Hōzōkan (法蔵館).

3/2/22 6:16 PM

Appendix 2.1: The corpus of images from the Śrīvijayan period The following symbols are used in the inventory of images: ? = unidentified, × = damaged, – = no discernible object. All measurements of images are taken from the selected literature. See also the open dataset Daivatanidhi (doi:10.6084/m9.figshare.17287115).

material

h/w (cm)

date

collection

1

bronze

13.5

9th cent.

2

bronze

13.7×7.8

c. 850– 899

Rijksmuseum, Amsterdam, AKRAK-1992-3

3

bronze?

15×6

8th– 10th cent.

4

silver

unknown

5

bronze

6

bronze

02_CreSouthV1_8P_4Feb22.indd 53

right hands rosary?, abhaya, –, varada

left hands –, ?, –, waterpot

findspot

selected literature

Kota Rentang, Deli Serdang regency, Sumatra

Edwards McKinnon and Azhari forthcoming

rosary, abhaya?, book?, ? (hook or trident?, noose), varada lotus?, waterpot

unknown (presumably Indonesia)

Lunsingh Scheurleer and Klokke 1988: 87, no. 35; Reichle 2007: 109, fig. 4.21

Balai Pelestarian Peninggalan Purbakala Batusangka, Dok. BP3 Batusangkar

rosary?, ×, abhaya, ×

Candi Pulau Sawah 2, Dharmasraya regency, Sumatra

Budi Istiawan 2012: 49, fig. 4.7; Ery Soedewo 2012: 156, Foto 11

9th– 11th cent.

unknown (Museum Nasional, Jakarta?; base marked ‘626.b.’)

rosary?, abhaya, ?, varada

book, ? Yogyakarta, (noose or Java trident), lotus stalk, × (waterpot?)

Leiden University Digital Collections Photographs (Kern Institute) OD-1523, OD1523a

16.5

9th cent.

MMoA, New York, 1987.142.194

rosary, ?, abhaya, varada

book, noose, Java (or lotus stalk, ‘Sivachai’) waterpot

n.d. 1966

20.6

8th– 10th cent.

Rijksmuseum voor Volkenkunde, Leiden, RV-1403-1874

hook?, ×, ×, varada

book?, ×, ×, ×

Tobi 1931: 181, cat. 1874, pl. 52d; Pal 1966: 236, fig. 1; Lunsingh Scheurleer and Klokke 1988: 58, no. 6; Srinivasan 2013: 176, fig. 13.6

–, ×, lotus, ? (waterpot?)

Java

3/2/22 6:16 PM

Iain Sinclair

54 material

h/w (cm)

date

collection

right hands

left hands

findspot

selected literature

7

bronze

21.5

none given (c. 8th– 10th cent.)

Ethnographische Sammlung des Naturhistorischen Staatsmuseums, Vienna

rosary, ?, abhaya?, varada

book, noose?, lotus, waterpot

8

bronze

21.6

9th cent.

Calon Da Collection

rosary?, ?, abhaya?, ×

book?, lotus, ‘Northern waterpot?, ? Thailand’

Bunker and Latchford 2011: 194–195, fig. 6.6

9

bronze

22

9th cent.

Bangkok National Museum (inventory number unknown)

×, ×, ×, ×

book, ?, ?, waterpot

Chaiya, Surat Thani province (Nik Hassan Shuhaimi) or Yala province, Betong

Nik Hassan Shuhaimi 1976: pl. 46; JacqHergoualc’h 2002: pl. 145

10

bronze

22.5

c. 8th– 10th cent.

Ethnographische rosary, ?, book, ?, Sammlung des abhaya, varada lotus stalk?, waterpot Naturhistorischen Staatsmuseums, Vienna

Java (? the collection of origin is in Surakarta)

Heine-Geldern 1925: 20, no. 10, pl. 10

11

bronze

24×12.7

7th– 9th cent.

Vereniging van ?, –, ×, varada Vrienden der Aziatische Kunst, Haastrecht, on loan from Rijksmuseum, AKMAK-1234

12

bronze

28

10th cent.

British Museum, London, 1859, 1228.75

rosary, hook?, book, conch, Prambanan, lotus, varada –, waterpot Yogyakarta, Java

Murphy et al. 2019: 177, no. 147

13

copper alloy

41×23.2

8th– 9th cent.

Private collection

rosary, ×, ?, ×

Guy 2014: 253, cat. 160

14

bronze

40

c. 10th Skanda Trust cent.

rosary, hook, book, lotus abhaya, varada stalk, ×, waterpot

Anlong Veng District, south of Preah Vihear

Bunker and Latchford 2011: 198–199, fig. 6.8

15

bronze

46

9th cent.

rosary, abhaya?, book, ?, –, ? ?, varada

‘Possibly discovered in the Malay peninsula’

Wattanavrangkul 1975: fig. 53; Nik Hassan Shuhaimi 1976: pl. 23; Diskul 1980: 52, no. 28; Jacq-Hergoualc’h 2002: pl. 118

02_CreSouthV1_8P_4Feb22.indd 54

Dr Viroj Karnasut, Bangkok

Java (? the collection of origin is in Surakarta)

book?, –, –, – Bantaeng, Bontonompo district, Gowa regency, South Sulawesi

book, –, ‘reportedly SE lotus stalk?,– Sumatra’ (varada?)

Heine-Geldern 1925: 21, no. 11, pl. 11

Lunsingh Scheurleer and Klokke 1988: 111, no. 59

3/2/22 6:16 PM

Appendix 2.1 material

h/w

date

collection

right hands

55 left hands

findspot

(cm) 16

stone

17

18

selected literature

53

9th– 10th cent.

Museum Nasional, Jakarta Inv. 6024

rosary?, ×, ×, × ×, ×, ×, ×

Muara Sungai Komering, Palembang, Sumatra [1929]

Bernet Kempers 1959: pl. 175; Nik Hassan Shuhaimi 1976: pl. 33; Fontein 1990: 214, cat. 59; Guy 2014: 254, n. 1; Bambang Budi Utomo 2013: 10, Foto 10; ibid. 2016: 132, no. 53; Krishnan 2016: 179, pl. 20

copper alloy

76.4

c. 750– 800

Bangkok National Museum SV24

×, ×, ×, ×

×, ×, ×, ×

Wat Phra Borommathat, Chaiya, Surat Thani

Nik Hassan Shuhaimi 1976: pl. 36; Diskul 1980: 52, pl. 26; Krairiksh [1980]: 142, pl. 32; Jacq-Hergoualc’h 2002: 312, pl. 108; Krairiksh 2012: 248, pl. 2.196; Bambang Budi Utomo 2013: 21, Foto 26; Guy 2014: 259, cat. 166; Krishnan 2016: 187, pl. 28

copper alloy

93×51

8th– 10th cent.

National Museum, Kuala Lumpur, MN.BALAIB.40.2008

rosary, ×, ?, varada

book, lotus stalk, noose?, waterpot

Bidor, Batang Padang, Perak [1936]

Quaritch Wales 1940: 51, pl. 79; Pal 1966: 237; Kwa 1967–1968; Nik Hassan Shuhaimi 1976: pl. 35; Quaritch Wales 1976: pl. 11; Harris and Zainal 1990: 13; JacqHergoualc’h 2002: fig. 160; Bambang Budi Utomo 2013: 20, Foto 24; Guy 2014: 250, cat. 157 et al.; Krishnan 2016: 183, fig. 23; Park 2019: 303, fig. 14

02_CreSouthV1_8P_4Feb22.indd 55

3/2/22 6:16 PM

Iain Sinclair

56 material

h/w (cm)

date

collection

right hands

left hands

findspot

selected literature

19

bronze?

unknown

no date (c. 9th cent.)

Bảo tàng Lịch sử Việt Nam

no discernible objects

×, ×, lotus?, ×

Đang Bình, Ninh Thuận province

Boisselier 1963: fig. 31; Chutiwongs 1984: pl. 177; Béguin 2009: 166

20

bronze?

unknown

no date (9th– 10th cent.)

exhibited at Kunsthal, Rotterdam, 2009.

rosary, ×, ×, varada

book, lotus stalk, ?, waterpot

unknown (Indonesia)

URL: https:// www.flickr.com/ photos/10015390 @ N00/3653707919

21

bronze?

unknown

9th cent.

Museum Ranggawarsita rosary, ?, noose?, varada

book, hook, waterpot, lotus stalk

unknown (Indonesia)

Lokesh Chandra 1999: 304; Park 2019: 301, fig. 11

22

bronze

18.8×7

10th– 11th cent.

Museon, The Hague, Inv. 48659

rosary, ?, ?, ?

hook?, ?, waterpot, lotus stalk

unknown (Java)

URL: https://hdl. handle.net/ 21.12123/16939

23

bronze?

15×5

none given (8th– 10th cent.)

Museon, The Hague, Inv. 48675 (“Mogelijk een copie”)

rosary, ?, ?, ?

?, ?, waterpot, unknown lotus stalk (Java)

URL: https://hdl. handle.net/ 21.12123/16954

02_CreSouthV1_8P_4Feb22.indd 56

3/2/22 6:16 PM

Appendix 2.1

02_CreSouthV1_8P_4Feb22.indd 57

57

3/2/22 6:16 PM

Iain Sinclair

58

Fig. 2.15: Eight-armed Amoghapāśa statues from the Malay Archipelago, drawn to approximate relative scale, where known, 7th–11th centuries. Hand drawings after supplied or published photographs. Numeric labels correspond to Appendix 2.1.

02_CreSouthV1_8P_4Feb22.indd 58

3/2/22 6:16 PM

Appendix 2.2: Amoghapāśamaṇḍalarājāgrī nāma samādhi— Sanskrit text and translation

Sigla A: Amoghapāśapūjā Aṣṭamīvrata, MS Shakya (2010: 21) No. 17, paper, n.d. (c. 20th century), ff. 3v10–8r4. B: Śuklāṣṭamīvratavidhāna, MS Shakya (2010: 21) No. 18, paper, n.d. (c. 20th century), ff. [27]r1–[33]v5. C: Vajrācārya, Badrīratna ed. (2005:[2]–[7]). The word spacing of Vajrācārya’s text is cited verbatim in the apparatus. D: Untitled MS Asha Archives DPN 2159, paper, n.d., ff. 6r5–12r1. tato maṇḍalarājāgrī nāma samādhiḥ188 | tadanu maitrīkaruṇāmuditopekṣācaturbrahmavihārān189 vibhāvya190 om̐ svabhāvaśuddhāḥ sarvadharmāḥ svabhāvaśuddho ’ham | śūnyatājñānavajrasvabhāvātmako191 ’ham ||

tato (ayam̐ kāreṇa niṣpannaṃ vāyumaṇḍalaṃ dhanvākāraṃ nīlapatākāṅkitaṃ192 tasyopari rephodbhavaṃ raktavarṇam agnimaṇḍalaṃ193 trikoṇaṃ raktarephāṅkitaṃ tadupari vam̐ kārapariniṣpannaṃ194 vāruṇamaṇḍalaṃ195 vartulaṃ196 śuklavarṇaṃ śvetaghaṇṭāṅkitaṃ197 tadupari lam̐ kārasambhavaṃ māhendramaṇḍalaṃ caturasraṃ pītavarṇaṃ198 pītatriśucikavajraṃ catuṣkoneṣu suśobhitaṃ199 tadupari sum̐ kāraniṣpannaṃ catūratnamayam aṣṭaśṛṅgopaśobhitaṃ Sumerum200 |a) maṇiratnamayaṃ śrīmat188. samādhiḥ A samādhiṃ vakṣāmi BD samādhiṃ C 189. maitrīkaruṇāmuditopekṣā caturbrahmavihārān BC maitrīkaruṇāmuditopekṣā catubrahmavihārā D trīkaruṇāmuditopekṣāṃ caturbrahmavihāraṃ A 190. vibhāvya D bhāvayet ABC 191. śūnyatā- ABC śūnyatāṃ D 192. dhanvākāraṃ nīla patākāṅkitaṃ C dhanvākāraṃ nīlapatākāṃkita B dhanvākāranilāyatākāṃkite D dhanvābhaṃ nīlapatākāṃkitaṃ A 193. agnimaṇḍalaṃ B≈C agnimaṇḍala D vahnimaṇḍalaṃ A 194. vam̐ kāraniṣpannaṃ A vaṃkārapariniṣpannaṃ B≈C vaṃkārapariniṣpaṃna D 195. vāruṇamaṇḍalam B vārunimaṇḍalaṃ D varuṇamaṇḍalaṃ A vāyu maṇḍalaṃ C 196. vartulaṃ A vaturaṃ B caturaśraṃ CD 197. śvetaghaṇṭāṅkitaṃ A śukla patākāṅkitaṃ C śuklapadāṃkitaṃ B śuklapatākāṃkita D 198. pītavarṇṇaṃ B pitavarṇṇa· D≈C pītaṃ A 199. pitatriśucikavajraṃ catukoneṣu suśobhitaṃ D pītatriśuklavajra catukoṇaṃ suśobhitaṃ B pīta trisūcika vajra catuṣkoṇaṃ suśobhitam C koneṣu triśucikāṅkitavajraṃ A 200. catūratnamayam aṣṭaśṛṃgopaśobhitaṃ sumeruṃ A saptaratnamayaṃ aṣṭāṃgasumeruṃ BC saptaratnamayaṃ aṣṭāṃgaśumyaru D

02_CreSouthV1_8P_4Feb22.indd 59

3/2/22 6:16 PM

60

Iain Sinclair

Potalakācalodaranivāsinaṃ201 kūṭāgāraṃ (bcaturdvāraṃ catustoraṇaṃ bhūṣitaṃ catuṣkoṇeṣu sarveṣu dvāraniryūhasandhiṣu candrārkavajracihnaṃ tu hārārdhahāraracitaṃ catuḥsūtrasamāyutaṃ202 paṭṭasragdāmabhūṣitaṃ203 bakulikramaśīrṣaparyantaṃ204 vajraratnakhacitakumbhastambhāḥ205 | mahāvajrakramaśīrṣas206 tu pakṣiṇī207 kiṅkiṇijālamaṇḍitam | dhvajair aṣṭair alaṅkṛtaṃ cāmarādivibhūṣitaṃb) haribahiḥ susaṃsthitam || abhyantare208 maṇḍale pam̐ kārapariṇataṃ209 viśvadalakamale210 am̐ kārajaṃ211 pūrṇaśaśimaṇḍalam (c | tadupari212 hrīḥkāram atiprabhāsvaraṃ sphuratkiraṇasamūhaṃ213 vyāptanabhastalaṃ214 vibhāvya tatpariṇataṃ svabījagarbhaṃ nālālaṅkṛtaṃ215 vikacakamalaṃ samavalokayet216 | tata etat217 sakalaṃ pariṇatam ātmānaṃ218 bhagavantaṃ dhyāyāt219 | himakarakoṭikiraṇavadātaṃ dehaṃ220 smeramukhaṃ221 dviraṣṭavarṣadeśīyaṃ222 karuṇāsnigdhāvalokaṇaṃ223 śṛṅgārarasaparyupāsitam atiśāntaṃ sarvalakṣaṇasaṃpūrṇaṃ c) sarvālakṣaṇavivarjitam224 ekamukhaṃ tapasviveśadharam ūrdhvajaṭāmakuṭam225 Amitābhakṛtaśekharaṃ divyasuvarṇamāṇikyamuktavaidūryasarvaratnādyābharaṇabhūṣitam | aṣṭaa

Compare Harihariharivāhanodbhavalokeśvarasādhana: yam̐ kāra­niṣpannaṃ vāyu­maṇḍa­laṃ […] sumerum (Sādhanamālā 34, ed. Bhattacharyya 1925: 782–8; Sakuma 2002 No. 9.2). 201. śrīmatpotalakācalodaranivāsinaṃ A śrīmatpotalakāvarodalavirājitaṃ D śrīmatpotarakāvarodananivāsita B śrīmatpotalakāvarodara nivāsita C 202. catuḥsūtrasamāyutaṃ A catusūtrasamāyutaṃ B≈D catuḥsūtra samāyuktaṃ C 203. paṭṭasragdāma- A paṭśragdāma- B paṭṭaśradāma- D patra sragdāma- C 204. vakulikramaśīrṣaparyantaṃ corr. vakulikakarmaśīrṣaparyyantaṃ A≈B vaktarika karmaśīrṣa paryantaṃ C vakulikakarmmasiṣaraparyyaṃtaṃ D 205. vajraratnakhacitakumbhastambhāḥ A vajraratnena ṣacitaṃ kumbhastambha B vajraratnacitaṃ kuṃbhastaṃbhe D vajraratna nakhacitaṃ kumbhastambha C 206. mahāvajrakramaśīrṣas corr. mahāvajrakarmaśīrṣan A mahāvajrakarmmaṇīrṣa B mahāvajrakarmasirṣa CD 207. tu pakṣinī B catuḥ pakṣiṇi C pradakṣini D tu pradakṣiṇaṃ A b Compare Nāgabuddhi’s Viṃśatividhi 4.14–17, 8.12 (ed. Tanaka 2010) et al.: caturasram catu[r]dvāraṃ … ghaṇṭāpatākasaṃśobhaṃ cāmarādivibhūṣitaṃ 208. abhyantare corr. abhyantara- AC atyantare B atentara D 209. paṃkāra pariṇataṃ C≈B paṃkārena parinataṃ D pam̐ kārapariṇata A 210. viśvadalakamale ABD viśvadala kamalaṃ C 211. am̐ kārajaṃ A aṃkārajaṃ B akārajaṃ CD 212. tadupari D tatopari B tatropari AC 213. sphurat- C sphūra- AB spharaṇa- D 214. vyāptanabhastalaṃ A vyāpta nabhastalaṃ C vyāptanabhastaraṃ B vyāptena hastanaṃ D 215. nālāṅkṛtaṃ C nānālaṃkṛtaṃ ABD 216. vikacakamalaṃ samavalokayet C vikacanāla samavalokayet A vikacanālasasavalokayan B vikacanālasamavalokayan · D 217. tata etat A tataḥ etat C tata ye tat B tatra etata D 218. sakalaṃ parinatam ātmānaṃ B sakalapariṇatam ātmānaṃ A sakala paritaṇata sātmāntraṃ C sakalaparinatam ātmānaṃ D 219. dhyāyāt AC dhyāyā B dhyātvā D 220. dehaṃ ABC; D omits 221. smeramukhaṃ C smeramukha A samaramukha B samaramuṣa D 222. -deśīyaṃ corr. -deśiyaṃ BC -darśitaṃ A -deśitaṃ D 223. karuṇāsnigdhāvalokaṇaṃ corr. karaṇāsanigdhavalokaṇaṃ B karuṇasasnigdhāvalokana A karuṇāsgigdha vilokanaṃ C karuṇāśrīśavalokanaṃ D c Compare Padmākaramati’s Khasarpaṇasādhana: tatropari hrīḥkāram […] atiśāntaṃ nānālakṣaṇālaṅkṛtam (Sādhanamālā 14, ed. Bhattacharyya 1925: 3815–408; Sakuma 2002 No. 2.2). 224. sarvālakṣaṇavivarjitaṃ B sarvālakṣaṇavarjitaṃ A≈C sarvalakṣaṇavarjjitaṃ D 225. urdhvajaṭāmakuṭaṃ C urujaṭāmakuṭa B urujaṭāmakuṭaṃ C jaṭāmakuṭa A

02_CreSouthV1_8P_4Feb22.indd 60

3/2/22 6:16 PM

Appendix 2.2

61

bhujaṃ dakṣiṇe prathamabhujenāmṛtaśravadvaradaṃ226 dvitīyabhujenākṣamālādharaṃ tṛtīyabhujena pāśadharaṃ caturthabhujeṇābhayaratnadharam | vāme227 prathamabhujeṇa sanālasahasradalakamaladharaṃ karavigalatpīyūṣadhārāvyavahārarasikaṃ tadadhaḥ228 sūcīmukhaṃ tarpayantaṃ229 dvitīyabhujena Prajñāpustakadharaṃ230 tṛtīyabhujeṇa tridaṇḍadharaṃ231 caturthabhujeṇa kamaṇḍaludharam evaṃbhūtaṃ śrīAmoghapāśalokeśvaram232 ātmānaṃ bhāvayet | bhagavataḥ śirasi hrīm̐ kārotpannaṃ233 raktābhaṃ234 mayūrāsanaṃ Vajradharmam ātmānaṃ235 dhyāyāt236 || 1 || tadbahir237 bhagavataḥ purata238 Amoghāṅkuśo raktavarṇa aṅkuśaṃ savyakare dharaṃ239 vāmakare pāśadharam || 2 || (d bhagavato dakṣiṇe Tārā śyāmā vāmakaravidhṛtaṃ sanālam utpalaṃ240 dakṣiṇakareṇa vikāśayanti241 nānālaṃkāravatī abhinavayauvanodbhinnakucabhārā242 || 3 || Agnau243 Sudhanakumāraś ca kṛtāñjalipuṭaḥ kanakāvabhāsadyutiḥ kumārarūpadhārī244 vāmakakṣe vinyastapustakaḥ245 sakalālaṅkāravān246 || 4 || vāme Bhṛkuṭī caturbhujā hemaprabhā jaṭākalāpinī247 vāme tridaṇḍakamaṇḍaludhārī248 hastā dakṣiṇe vandanābhiṇayākṣasūtradharā karā249 trinetrā || 5 ||

226. dakṣiṇe prathamabhujenāmṛtaśravadvaradaṃ em. dakṣiṇa prathama bhūjonāmṛta suvat varadaṃ C dahinaprathamabhujenāmṛtaśravatvaradaṃ Bac dahinaprathamabhujenāmṛtasravadvaradaṃ A dakṣine prathamabhujenāmṛtaśevatˎ varadaṃ D 227. vāme C vāma- ABD 228. karavigatapīyūṣadhārāvyavahārarasikaṃ tadadho A karavigatapīyuṣadhārāvyavahārarasikatadadha· B kara vigalat pīyuṣa dhārābhya vahāra rasikaṃ C karavigatapiyuṣadhārāvevahāre rasikatadadha D 229. sūcīmukhaṃ tarpayantaṃ C; ABD omit. (All of the expression pertaining to the sūcīmukha here is misplaced and should apply to the first right hand, not the first left hand.) 230. prajñāpustakadharaṃ AD pustakadharaṃ Bac≈C 231. tridaṇḍadharaṃ ABC triśuladharaṃ D 232. śrīamoghapāśalokeśvaram B≈C śrīamoghapāsalokyaśvaram D śrīlokeśvaram A 233. hrīm̐ - BC hrīḥ- AD 234. raktābhaṃ A; BCD omit 235. vajradharmam ātmānaṃ bhagavantaṃ B vajradharmam ātmānaṃ bhagavaṃta D vajradharmaṃ bhagavantaṃ C bhagavantaṃ vajradharmam ātmānaṃ A 236. dhyāyāt BCD dhyāyet A 237. tadbahir corr. tadvahi ACD tadā hi B 238. bhagavataḥ purato corr. bhagavato purato B bhagavatī purataḥ C bhavaṃto D bhagavana A 239. aṅkuśaṃ savyakare dharaṃ corr. aṃkuśasavyakare dharaṃ A aṃkuśa savye kare dharaṃ· Bac kusasavyakaradhara D savyakareṇa aṅkuśadharaṃ C 240. vāmakamaladhṛtaṃ sanālam utpalaṃ B vāmakaradhṛta sarotpaṃna· D vāmakamaladhṛta sanāla mutpālaṃ C vāmakaravidhṛtasanālam utpala A 241. vikāśayanti B vikāsayantīṃ C savikāśayantī A savikāsayaṃti D 242. abhinavayauvanodbhinnakucabhārā em. abhina yaivanodbhannakucabhārā C abhinavayauvanā yauvanodbhinnakucabhārā A abhinavayauvanā yauvanndbhinnakuvabhārā B abhinavayo vanā hinakucabhālā D 243. agnau A agnaye C agne B agni· D 244. kumārarūpadhārī A kūmararūpa dhāri C kumārarūpadhāriḥ D kumārarūpadharo B 245. vāmakakṣe vinyastapustaka A vāmakaṣavinyastaka B vāmakara kṣavinyasta C vāmakakṣaviveśta D 246. sakalālaṃkāravān AC sakalālakṛtatanāt B kamakalālaṃkṛtavān D 247. jaṭākalāpinī A jaṭākalāpanī C jaṭākalaṃpinī B jaṭakalpinā D 248. tridaṇḍakamaṇḍaludhārī BCD tridaṇḍī kamaṇḍaludhārī A 249. vandanābhiṇayākṣasūtradharakarā BC vadanābhinaṃ akṣasutradharā· D varadābhinayākṣasūtradharakarā A

02_CreSouthV1_8P_4Feb22.indd 61

3/2/22 6:16 PM

Iain Sinclair

62

Īśāne Hayagrīvo raktavarṇaḥ kharvalambodara ūrdhvajvalatpiṅgalakeśo bhujaṅgayajñopavītī250 kapilataraśmaśruśrenīparicitamukhamaṇḍalo251 raktavartulatrinetro bhṛkuṭīkuṭilabhrūko vyāghracarmāmbaraḥ252 | daṇḍāyudho dakṣiṇakareṇa253 vandanābhinayī || 6 ||d) * tadbāhyāṣṭadalapadme254 Maitreyādīn vinyāset255 | Maitreyaḥ pītavarṇo256 nāgapuṣpavarapradaḥ || 7 || Kṣitigarbho dakṣiṇe śyāmavarṇaḥ kalaśaṃ cābhayaṃ tathā257 || 8 || paścime Vajrapāṇiḥ śuklābho vajrahastavarapradaḥ258 || 9 || uttare Khagarbhaḥ śyāmābhaś259 cintāmaṇivarapradaḥ260 || 10 || Agnau261 Mañjughoṣaḥ kanakābhaḥ262 khaḍgapustakadharaḥ263 || 11 || Nairṛtye Gaganagañjo raktavarṇo nīlotpalavarapradaḥ || 12 || Vāyavye Viṣkambhī tuṣāravarṇo ratnottamavarapradaḥ || 13 || Īśāne Samantabhadraḥ pītābho raktotpalavarapradaḥ264 || 14 ||e) (f tadbāhye Vahnyādicatuṣkoṇe265 | Agnau266 Puṣpā śuklā puṣpamālācakradhāriṇī267 vāmetarakarā268 || 15 || Nairṛtye269 Dhūpā dhūmravarṇā dhūpakaṭacchuratnagrāhī vāmetarakarā270 || 16 || Vāyavye271 Dīpā kanakavarṇā272 dīpayaṣṭipadmavāmasavyahastā273 || 17 || Aiśāne274 Gandhā raktā gandhaśaṅkhakhaḍgadhāriṇī275 vāmetarabhujā || 18 ||f) (e

250. bhujaṅgayajñopavītī bhujaṃgayajñopavīti A≈B bhūjaṃgayajñopavitī bhujagayajñāpāpati D 251. kapilataraśmaśruśroniparicittamukhamaṇḍalam B kapilataraśmaśruḥ śreṇīparicittamukhamaṇḍalo A kapilatara śmaśnuśreṇī paricittamukhamaṇḍalam C kapilataraśmaśruśrenīparicitamuṣamandula D 252. vyāghracarmāmbaraḥ A≈CD vyāghracarmāvaraḥ B 253. dakṣiṇakareṇa AB dakṣiṇeṇa C dakṣine karṇṇa D d Compare Padmākaramati’s Khasarpaṇasādhana: tasya puratas tārā […] daṇḍāyudhaḥ dakṣiṇa­kareṇa vandanā­ bhinayī (Sādhanamālā 14, ed. Bhattacharyya 1925: 409–41 4; cf. Sakuma 2002 No. 2.2). 254. tadbāhyāṣṭadalapadme AC tadbāhyāṣṭadale padme B tadvāhyāṣṭadme D 255. maitrīyādīn vinyāset C maitrayādivirasat B maitri āvineset· D; A omits 256. maitreyaḥ pītavarṇo corr. maitreyaḥ pītavarṇā A metreya pītavarṇṇābhā B maitreya pītavarṇe C maitriyapitaparṇṇā D 257. cābhayaṃ tathā AB cābhaya tathā D vāmeya tathā C 258. vajrahastapradaḥ em. vajrahastau varaprada C vajrāṃkuśavarapradaḥ A vajrāṃkuśavarapradaṃ B vajrakṣadanapadaṃ D 259. śyāmābhaś corr. śyāmābhā B śyāmābho C syāmavarṇṇa D śyāmaś A 260. cintāmaṇivarapradaḥ A cintāv manivalapradaḥ B cintāmaṇivaraprada C ciṃtāmaṇivalapradāṃ D 261. agnau A agnaya B āgnayai C agni D 262. kanakābhaḥ C kanakābhaṃ B kanakābhaḥ C kanakavarṇṇa D 263. khadgapustakadharaḥ A khaḍga pustaka dhārakaḥ C khaḍgapustakadhāriṇaṃ B ṣaḍgapustakadhāranaṃ D 264. raktotpalavarapradaḥ BC ratnotpaladharapradaḥ A ratnotpalavaradaṃ D e Compare the anonymous Lokanāthasādhana: tadvaraṭakāṣṭadale padme maitreyādiṃ ca vinyaset [...] samantabhadraḥ pītābho ratnotpalavarapradaḥ (Sādhanamālā 18, ed. Bhattacharyya 1925: 4911–502; cf. Sakuma 2002 No. 1.2). 265. tadbāhye vahnyādicatuṣkoṇe corr. tadbāhyavahnyādicatukoṇe BC tadbāhye catuṣkoṇe A tadbāhe catukone D 266. agnau A; BCD omit 267. cakra dhāriṇī C cakradhārī ABD 268. vāmetarakarā BCD vāmetarahastā A 269. naiṛtye AC naiṛtyāṃ B naiṛtyā D 270. vāmetarakarā A vāmatarahastā B vāme tarahastā C vāmakarahastā D 271. vāyavye AC vāyuvyāṃ BD 272. kanakavarṇā ABC; D omits 273. dīpayaṣṭipadmavāmasavyahastā AC dīpayaṣṭipadmavāmasavye haste B dipāvāmatālabhujā· D 274. aiśāne A aiśānyāṃ B īśāne C aiśānyā D 275. gandhaśaṃkhakhaḍgadhāriṇī C gaṃdhaśaṣakharḍgadhāri D gaṃndhaśaṃkhadhārī AB f Compare the Śrīsaṃpuṭatantroktavajrasattvamaṇḍala of Abhayākaragupta’s Niṣpannayogāvalī: tato bāhya-

02_CreSouthV1_8P_4Feb22.indd 62

3/2/22 6:16 PM

Appendix 2.2

63

tadbāhye276 | pūrvadvāre karābhyāṃ tu vajrāṅkuśasamarpitaḥ | jaḥkārabījasaṃjāto277 Vajrāṅkuśaḥ278 sitadyutiḥ || 19 || hūm̐ kārabījasaṃbhūto dakṣiṇadvārasaṃsthitaḥ | nīlavarṇaprabho ’sau279 tu hastābhyāṃ Vajrapāśabhṛt280 || 20 || vam̐ kārabījaniṣpanno Vajrasphoṭārkasamaprabhaḥ281 | hastābhyāṃ śṛṅkhalā gṛhya282 paścimadvārasaṃsthitaḥ || 21 || hoḥkārabījanirjāto Vajrāveśas tu uttare | vajraghaṇṭāṃ susaṃgṛhya283 viśvavarṇo mahādyutiḥ | padmasthacandrabimbe ca sattvaparyaṅkasaṃsthitaḥ || 22 || (g

tadbāhye ’ṣṭāracakre284 | tatra prācyāṃ śubhraAirāvatārūḍho285 vajradhara Indraḥ || 23 || avācyāṃ mahiṣārūḍhaḥ kṛṣṇo daṇḍadharo Yamaḥ286 || 24 || pratīcyāṃ makarārūḍhaḥ śvetaḥ saptaphaṇo287 nāgapāśaśaṅkhabhṛd Varuṇaḥ || 25 || udīcyāṃ naravāhanaḥ pīta aṅkuśagadādharaḥ288 Kuberaḥ || 26 || vidikṣu catvāraḥ | tata289 Agnau290 chāgārūḍho raktaḥ sruvakamaṇḍaludhara291 Agniḥ292 || 27 || Nairṛtye śavārūḍho nīlaḥ khaḍgapheṭakadharo293 rakṣasādhipo Nairṛtiḥ294 || 28 || Vāyavye295 mṛgavāhano nīlo vātapaṭadharo Vāyuḥ296 || 29 || Aiśāne297 vṛṣabhārūḍhaḥ298 sitaḥ triśūlakapālapāṇir Īśānaḥ299 || 30 ||g)

paṭṭikāyām āgneyāṃ puṣpā śuklā puṣpamālācakradhārīvāmetarabhujā [...] aiśānyāṃ gandhā raktā gandhaśaṅkhakhaḍgadhārīvametarakarā (ed. Lee 2004: 15–16). 276. tadbāhye A tadvāhe D; BC omit 277. jaḥkārabījasaṃjāto A jaṃkārabījasaṃ bhūtā B jaṃkāra vīja saṃbhūto C jaṃkāravijamaṃjātā D 278. vajrāṃkuśaḥ A vajrāṃkuśa BCD 279. nīlavarṇa prabho’sau C nīlavarṇaprabhā sau A nīlavarṇṇaprabho sau B nīlavarṇṇaprabhā amo D 280. vajrapāśabhṛt A vajraghaṇṭhabhṛt BCD 281. vajrāsphoṭārkka samprabhaḥ C vajrasphoṭārkasamaprabhaṃ AD vajrasphoṭasamaprabhaḥ B 282. śṛṃkhalā gṛhya AB śṛṃkhalāṃdhṛtvā C śṛkharā gṛhā D 283. vajraghaṇṭāṃ susaṃgṛhya A ghaṇṭhāsaṃgṛhya B ghaṇṭhāṃ saṃgṛhya C ghaṇṭhā saṃgṛhahastāṃbhyā D 284. tadbāhyāṣṭāracakre ABC tadbāhe aṣṭaracakraṃḥ· D 285. śubhra-airāvaṇārūḍho A sūtrair āvaṇārūḍho B airāvatārūḍho C śutairāvanārudhā D 286. kṛṣṇo daṇḍadharo yamaḥ AC kṛṣṇā daṇḍadharā jama B kṛṣṇa daṃdadharā yamaḥ D 287. śvetaḥ saptaphaṇo A śvetasaptaphanā ABD 288. pīta aṃkuśagadādharaḥ A pītāṃkuśagaḍādhara BC pītāṃkusaṃ gadādharaṃ D 289. vidikṣu catvāraḥ || tato B≈D tataḥ C; A omits 290. agnau A ’gneyā B āgnaye A agnayāṃ D 291. raktaḥ sruvakamaṇḍaludhara em. raktasrūvā kamaṇḍarūdharo C raktasūtrakamaṇḍaludhara A raktaśūtrakamaṇḍaludharā B raktasūtrakamaṃḍaludharāṃ D 292. agniḥ A agni B agnauḥ D; A omits 293. nīlaḥ khaḍgapheṭakadharo A≈C nīlakhaḍgapheṭakadharā B nilakhajephatakadharā D 294. rakṣasādhipo naiṛtyaḥ A rakstasādhipo naiṛtya B rākṣasādhipa C rākṣasādhipā·naiṛtyaḥ D 295. vāyavye A vāyuvyāṃ B vāyave C vāyuvyā D 296. nīlo vātapaṭadharo A nīlāvātapatadharā vāyu B haritavarṇa vātapaṭa dharo vāyuḥ C nīlāvātayatadharāḥ vāyu D 297. aiśāne A īśānyāṃ B īśānāṃ C īśānyāṃ D 298. vṛṣabhārūḍhaḥ A vṛṣabhārūḍho B vṛṣamārūḍhaḥ sitaḥ C vṛkhabhārudhā D 299. sitaḥ triśūlakapālapāṇir īśānaḥ A≈B sitaḥ triśūla kapāla pāṇiḥ C śītatriśūlakapālapāsi·īśāne D g Compare the Dharmadhātuvāgīśvaramaṇḍala of Abhayākaragupta’s Niṣpannayogāvalī: …aindryāṃ diśi śubhrairāvatārūḍha indraḥ pīto vajraṃ stanaṃ ca dadhānaḥ […] aiśānyāṃ vṛṣabhārūḍha īśānaḥ sitaḥ triśūlakapālapāṇiḥ […] vāyavyāṃ mṛge vāyur nīlo vātapaṭadharaḥ (ed. Lee 2004: 74–75).

02_CreSouthV1_8P_4Feb22.indd 63

3/2/22 6:16 PM

Iain Sinclair

64

ete sarve ekavaktrāḥ300 svanāyakānanapreritadṛṣṭayo301 yathāśobham avasthitāś302 cintanīyaḥ | evaṃbhūtaṃ samāṇḍaleyaṃ bhāvayet303 || iti Maṇḍalarājāgrī nāma dvitīyo samādhiḥ304 ||

Translation Next is the concentrated state called Maṇḍalarājāgrī. Now, having cultivated the four Brahmic states—friendliness, com­passion, delight, equanimity— [repeat]: om̐ , all dharmas are inherently pure, i am inherently pure. i have the inherent nature of the vajra of knowledge of emptiness.

Then, emanated by means of the syllable yam̐ , is a bow-shaped wind maṇḍala marked with a blue flag; on top of it, arising from a ra[m̐ ], a red fire maṇḍala, triangular, marked with a red ra glyph; above that, transformed from the syllable vam̐ , a water maṇḍala, round, clear in colour, marked with a white bell; above that, arising from the syllable lam̐ , an earthen maṇḍala, four-sided, yellow, blazoned with yellow three-pronged vajras in the four corners; above that, emanated from the syllable sum̐ , Sumeru, made of four jewels, boasting eight peaks underneath. The residence, the crystalline cavern of blessed Potalaka mountain, is a pagoda adorned [with] four doors [and] four tympana, having sun-and-moon emblems at all four corners of the points of contact between the doorway and the pediments (niryūha) and decorated with pearl and half-pearl [strings], enclosed by four cords, beautified by banners and streamers, girded by piping and parapets, the vase-capital columns studded with vajra gems. Also, the sides have a great vajra parapet festooned with a net of tinklebells. [The maṇḍala] is adorned with eight flagstaffs, is well decorated with yaktail fans; the sun outside is in the correct place. In the internal maṇḍala transformed from the syllable paṃ, upon a lotus with all-spreading petals, is the full moon maṇḍala generated from the syllable am̐ . Having visualized upon it the syllable hrīḥ, extremely bright, [its] concentration of shimmering rays pervading the sky, he should look upon the nucleus of his syllable changed into an outspread lotus adorned with a stalk. Then he should imagine all of that transformed into the persona of the Bhagavat. His body the dazzling white of a crore of snow-glints, he has a smiling face, about twice eight years [old], looking compassionate and affectionate, attended by passionate sentiment, most peaceful, complete with all [auspicious] marks, bereft of all non-[auspicious] marks, single-faced, wearing ascetic garb; having a crown of piled-up locks, Amitābha is fashioned in the crest; he is adorned with divine ornaments of gold, ruby, pearl, lapis and jewels of all kinds. He is eight-armed, on the right [side] bestowing a stream of ambrosia with the first hand, holding a rosary with the second hand, holding a noose with the third hand, holding the jewel of unfear with the fourth hand; on the left, holding a thousand-petalled lotus with stalk with the first hand, with the second hand satiating the needlemouthed [preta] below him tasting the to-and-fro of the stream of downpouring nectar, holding a book of the Prajñā with the third hand, holding a waterpot with the fourth hand—in this way should he conceive the persona of the blessed Infallible Noose (Amoghapāśa), the glorious Lokeśvara. 300. ete sarve ekavaktrāḥ A ete sarvva evaṃ B ete sarve eva C te sarva yakavaktrā D 301. svanāyakānanaprerita dṛṣṭayo C svanāyakānanaparittadṛṣṭayo A svanāyakānanapreritadṛṣṭaḥ yo B paritadṛṣṭo yo D 302. yathāśobham avasthitāś B yathāśobhavasthitāś A yathā śobhana vasthitāś C yathāśobham avasthitā D 303. bhāvayet A bhāvayat D bhāvayet || pūrvavat puṣpaṃ nyāśa || ām̐ pām̐ pum̐ pam̐ lām̐ gham̐ stum̐ tam̐ || B bhāvayet || ā· pā· pū· paṃ· lā· ghaṃ· stuti || C 304. maṇḍalarājāgrī nāma dvitīyo samādhiḥ em. maṇḍalarājā dvitīya samādhiḥ A maṇḍalarājagrī nāma dvitīya yoga samādhiḥ C maṇḍalogrī nāma dvitīyayogasamādhiḥ B maṃdalarājaśrīnāmasamādhiḥ D

02_CreSouthV1_8P_4Feb22.indd 64

3/2/22 6:16 PM

Appendix 2.2

65

He should think that the persona of Vajradharma, red in colour, seated on a peacock, is emanated from a hrīm̐ syllable on the Bhagavat’s crest. (1) On his outside, in front of the Bhagavat, is Amoghāṅkuśa, red in colour, a hook held in his right hand, holding a noose in his left. (2) On the Bhagavat’s right is Tārā, dark [green], in the left hand grasping with the stalk a blue waterlily being opened up with the right hand. [She is] replete with a variety of jewelry, bearing breasts of freshest youth. (3) And in the southeast is prince Sudhana, cupping [his hands] in salutation; a book tucked under his left arm, he has the lustrous sheen of gold, maintaining the appearance of a prince laden with all ornaments. (4) On the left is Bhṛkuṭī, four-armed, of golden aura, her locks fanned out, holding in hand a triple staff and waterpot on the right, the arms on the left waving in greeting and holding a seed rosary. [She is] three-eyed. (5) In the northeast is Hayagrīva, red in colour, squat and potbellied, having flamelike orange hair, an anaconda for a sacred thread, tawny facial hair lined around the circle of his mouth, three round red eyes, frown-furrowed brows, and a tiger-pelt garment. Armed with a club in his right hand, he waves in greeting. (6) On the eight-petalled lotus outside of that one should situate Maitreya and so on. Maitreya, yellow in colour, presents a nāga flower and the giving [gesture] (7), and Kṣitigarbha, in the South, dark-coloured, a vase and unfearing (8). In the West, Vajrapāṇi, white-hued, presents a vajra in his hand and the giving [gesture] (9). In the north, Khagarbha, dark-hued, presents a wishing gem and the giving [gesture] (10). In the Southeast, Mañjughoṣa, golden-hued, holds a sword and book (11). In the Southwest, Gaganagañja, red-coloured, presents a blue waterlily and the giving [gesture] (12). In the Northwest, Viṣkambhin, frost-coloured, presents a jewel aloft and the giving [gesture] (13). In the Northeast, Samantabhadra, yellow-hued, presents a lotus and the giving [gesture] (14). In the four corners outside of that, starting with the Southeast, is Puṣpā, white, holding a circular flower wreath in her left and right hands (15). In the Southwest is Dhūpā, smoky-coloured, holding an incense burner in her left and right hands (16). In the Northwest is Dīpā, golden-coloured, holding a lamp wick and lotus in her left and right hands (17). In the Northeast is Gandhā, red, holding a perfume conch and sword in her left and right hands (18). Outside of that, at the eastern doorway, holding a vajra-pommel goad in both hands, arising from the syllable jaḥ, is Vajrāṅkuśa, bright white (19). Staying in the southern doorway, born from the syllable hūm̐ , radiating dark blue, is he who holds the Vajrapāśa in both hands (20). Emanated from the syllable vam̐ is Vajrasphoṭa, the same colour as the sun, holding shackles with both hands, staying at the western doorway (21). Springing forth then from the syllable hoḥ is Vajrāveśa in the north, nicely holding a vajra bell, multicoloured, very bright, sitting in sattvaparyaṅka in a moon disc upon a lotus (22). On the eight-spoked wheel outside of that, in its eastern [sector], is Indra the vajra holder, riding bright white Airāvata (23). In the southern [sector] is Yama the truncheon-holder, black, riding a water buffalo (24). In the western [sector] is Varuṇa, seven-hooded, bearing a serpentine noose and a conch, riding a sea monster (25). In the northern [sector] is Kubera, holding a hook and a mace, yellow, riding a man (26). Then, in the four ordinal directions, is Agni in the Southeast, red, riding a goat, holding a ladle and waterpot (27); Nairṛti in the Southwest, dark blue, riding a corpse, holding a sword and shield, a ruler of ogres (28); Vāyu in the Northwest, dark blue, riding a deer, holding a sail (29); Īśāna in the Northeast, white, riding a bull, trident and skull in hand (30).

02_CreSouthV1_8P_4Feb22.indd 65

3/2/22 6:16 PM

Chapter 3

In the Footsteps of Amoghavajra (705–774): Southern Indian Artistic Mode in Tang China and its Transmission to Tibet1 y u ry k hok hlov

A

introduction

s a starting point, this research analyses the imagery in a distinctive Indian artistic mode2 found in various paintings from Mogao and Yulin caves.3 The material under review belongs to the 8th to 11th century and often features

1. This is a revised version of the article ‘Uncovering Amoghavajra’s Legacy in the Hexi Corridor and Tibet’ that was initially published in Journal of Tibetology 21 (2019). It expands on a talk given at the Seventh International Conference on Tibetan Archaeology & Arts in Chengdu, October 2018. Many people helped me in various stages of preparing it. I would like to thank in particular Tristan Bruck, Analicia Busha, Geoffrey Goble, Shawo Khacham, Rory Lindsay and Darik Thokmay for their assistance with my studies on the topics discussed in this chapter; and Emmanuel Francis, Yuko Fukuroi, Amy Heller, Deepanjana Klein, Arno Klein and Nicolas Revire for sharing with me images from their photographic archives. I am immensely grateful to Andrea Acri, Claudine Bautze-Picron, Rolf Giebel, Yannick Laurent, Robert Mayer, Neil Schmid, Per Sørensen, Jeffrey Sundberg, Joie Szu-Chiao Chen, Charles Orzech, Sam van Schaik, Peter Sharrock, Zhang Changhong, Dorothy Wong and several anonymous reviewers, who read my manuscript, corrected the errors and offered me their extremely valuable feedback. All remaining errors are of course my own. [The editors wish to thank Yannick Laurent for his assistance in editing this chapter, published posthumously] 2. The term ‘artistic mode’ is used as a broader category than artistic style and refers to a general way of representation. Additionally, ‘imagery in an Indian artistic mode’ refers to Indianized images that were made to look as if they were Indian as opposed to images in a traditional local artistic mode, which for the area discussed in this chapter is Chinese. 3. A list of these paintings can be found in Appendix 3.1.

03_CreSouthV1_8P_3Feb22.indd 66

Esoteric Buddhist imagery.4 Since some of the early examples within this group belong to the period of the Tibetan occupation of the Hexi Corridor,5 their artistic mode has often been ascribed—erroneously, in my view—to the cultural influence of Tibet. Several recent publications on this topic provoked me to address the issues that I see in such attributions.6 Contrary to this common opinion, I will argue that this artistic mode is, in fact, influenced by Southern Indian7 art and, in particular, 4. The term ‘Esoteric Buddhism’ is notoriously complex and problematic; in this chapter it is used mostly as a reference to teachings that were introduced to China during the Tang and possessed instructions in the use of mantra, dhāraṇī, and maṇḍala. ‘Esoteric Buddhist imagery’ refers here to representations of deities associated with these teachings. An extensive discussion on the term ‘Esoteric Buddhism’ can be found in Orzech, Sørensen, and Payne 2011: 3–18. 5. In this article, I adopted the dates of the Tibetan conquest of the Hexi Corridor from van Schaik and Galambos 2012: 62–63. Guazhou (Ch. 瓜州, the prefecture where the Yulin caves are located), was captured by the Tibetans in 776, while Shazhou (Ch. 沙州, the prefecture containing Dunhuang) fell in 786. After the collapse of the Tibetan empire, the whole Hexi Corridor was once again under Chinese rule by 851. 6. Recent publications that argue for the Tibetan origin of the discussed imagery include Kapstein 2004, 2009, 2014; Sha 2013; McCoy 2016; Wang 2018; Debreczeny 2019. See fns. 54, 138, 143 for details. 7. Geographically, the Indian subcontinent has been divided into the Northern and Southern parts. While the Northern part includes the northern mountains and the Indo-Gangetic plains, the Southern part is comprised of the Indian Peninsula. Southern India is a geographical designation, which should not be confused with South India, which, in modern administrative terms, includes the

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

67

Map 3.1: Major land and maritime routes between South and East Asia in the first millennium CE and sites mentioned in the article (based on a map from https://en.unesco.org/silkroad/about-silk-roads#pid=1, with the permission of UNESCO).

03_CreSouthV1_8P_3Feb22.indd 67

3/2/22 6:19 PM

68

Yury Khokhlov

certain stylistic and iconographic features can be traced exclusively to the art of the Pallava kingdom in Tamil Nadu and the Pallava influenced art of the Western Deccan, thereby ruling out alleged artistic influences from Tibet. Moreover, the discussed images are closely affiliated with teachings introduced to China by Vajrabodhi (or Vajrabuddhi, Ch. 金剛智, 671–741) in the first half of the 8th century and specifically propagated in Hexi by Amoghavajra (Ch. 不空 Bukong, 705–774) in the mid-8th century. Thus, the presence of such depictions in artwork from the Tibetan period merely signifies the continuation and Tibetan appropriation of the artistic and religious trends that existed in Hexi before the Tibetan conquest. Furthermore, it will be shown that this artistic mode was transmitted to Central Tibet during the Tibetan occupation of the Hexi Corridor and can be recognized in sculptures from a number of early Central Tibetan temples. I argue that this artistic transmission was a result of significant religious transmission, which included, but was likely not limited to, the cult of Vairocana and the eight great bodhisattvas, as well as the teachings of the Yoga tradition (i.e. Yogatantras according to the Tibetan classification). This article is divided into four major parts. Part 1 presents a brief historical overview of Buddhism and artistic trends in Hexi before the Tibetan conquest. Part 2 discusses the most representative area encompassing the Indian states of Andhra Pradesh, Karnataka, Kerala, Tamil Nadu and Telangana as well as the union territories Andaman, Nicobar, Lakshadweep, and Puducherry.    An imaginary division line between the two parts goes along the Vindhya Mountains and the Satpura Range, which separate the Gangetic plains and the Deccan plateau (see Map 3.1, where it is shown as a black dashed line). The two parts of the subcontinent, being interconnected on many levels, are, nevertheless, characterized by different histories, economies and cultures, and often by distinctive artistic styles. See Thapar 2003: 37–68. There is yet another distinction which is particularly important for this research. While the cultural influence of Northern India mostly spread along land-bound trade routes, Southern Indian influence is mostly noticeable in territories with which Southern Indian polities engaged in sea trade. From this perspective, landlocked Tibet belongs to the Northern Indian sphere of influence, while China was naturally predisposed to receive influence from both Northern and Southern parts of India.

03_CreSouthV1_8P_3Feb22.indd 68

depictions within this group of materials and their connection to Southern Indian art and Amoghavajra’s textual heritage. In Part 3, I will provide an overview of Southern Indian influence on Buddhist art in China in general and situate this Hexi group of images within a larger picture. In Part 4, I will argue that the artistic mode under review and its corresponding teachings were transmitted from Hexi to Tibet during the period of Tibetan occupation. Major sites mentioned in the article are shown in Map 3.1, which also illustrates major land and maritime routes between South and East Asia in the first millennium CE.8

part 1.  historical setting By the Han dynasty (206 BC–220 CE), the Hexi Corridor was already a strategic part of the Silk Road and a gateway to China for those travelling from India and Central Asia. This area was an early recipient of Buddhist teachings and artistic styles arriving to China from the West. Frequent contacts with the West led to a highly cosmopolitan Buddhist culture that encouraged a close imitation of foreign artistic models. Imperial Buddhist patronage and expansion of the Tang dynasty to the West in the Tarim Basin created a steady traffic of monks travelling between China and Northern India through Hexi. Indian monks travelling to China and Chinese monks returning to China from India often carried with them Buddhist texts and imagery as tribute to the court. This process resulted in rapid transmission of the newest—and increasingly Esoteric—teachings and in the popularity of Indian icons and Indianized styles in China, and, in particular, in the Hexi Corridor (Dunhuang).9 Whereas the cultural influence of Northern India mostly spread along land-bound trade routes, including those known as the Silk Road, the impact of Southern Indian culture is mostly noticeable in territories with which Southern Indian polities 8. Routes on the Tibetan plateau are not shown in the map. 9. The most recent and comprehensive research on travelling monks as agents of religious and artistic transmission from India to China is in Wong 2018. On Esoteric Buddhism in Dunhuang and depictions of Esoteric deities in Dunhuang art during the Tang, see Schmid 2011.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) engaged in sea trade. Recent scholarship has recognized that maritime relations played an equally important role in the spread of Buddhist ideas and artistic styles.10 Perhaps one of the most important maritime Buddhist imports to Tang China was the Jin’gangding Yoga (Ch. 金剛頂瑜伽 Jin’gangding yuqie; lit. the Yoga of Adamantine-Crown)11 teachings from Southern India. This tantric system includes eighteen texts, out of which only the Sarvatathāgatatattvasaṅgrahatantra (STTS),12 also known as the Vajraśekharasūtra, was transmitted to China during the Tang dynasty. The tradition based on the STTS (hereafter the Yoga tradition) became the foundation for a Buddhism of national protection in China in the second half of the 8th century.13 Propagation of the Yoga tradition in China is attributed to the activities of the Indian monk Vajrabodhi14 and his pupil Amoghavajra.15 It is maintained that Vajrabodhi studied the Jingangding Yoga under Nāgabodhi/Nāgabuddhi in Southern India, most likely at Kāñcīpuram,16 the capital of the Pallava Kingdom and a great centre of religious and Sanskrit learning equal almost to Nālandā.17 It is said that Nāgabodhi in turn was a disciple of Nāgārjuna, who is also believed to be from Kāñcīpuram,18 and who according to legend obtained the Jingangding Yoga teachings from an ‘iron stūpa’, which was probably located in 10. On the spread of Buddhism through maritime networks, see Acri 2016, 2018. 11. These teachings are also known under the name Vajraśekhara. In addition, Rolf Giebel argues that Vajroṣṇīṣa is a correct Sanskrit reconstruction of the Chinese title 金 剛頂瑜伽, see Giebel 1995: 109. 12. On the STTS and its English translation, see Giebel 2001. 13. Yang 2018: 120–158. On the notion of National Protection Buddhism in Tang China, see Orzech 1998: 1–9, 135–209. 14. On Vajrabodhi, see Sundberg and Giebel 2011. 15. On Amoghavajra and his teachings, see Chou 1945: 284–307; Goble 2012; Orzech, Sørensen and Payne 2011: 263–285; Yang 2018. 16. Sundberg and Giebel 2011: 151 17. On Kāñcīpuram, see Thapar 2003: 344. Lokesh Chandra even suggested Kāñcīpuram as the location of Uḍḍiyāna, see Lokesh Chandra 1980. On the other possible locations of Uḍḍiyāna, see Sanderson 2007: 265 and chapter 3 in Vol. 2 of this publication. 18. Sundberg and Giebel 2011: 195, n. 110.

03_CreSouthV1_8P_3Feb22.indd 69

69

Tamil Nadu or Andra Pradesh.19 Vajrabodhi was famed as a great Esoteric master and performer of miraculous deeds, and before travelling to China he was active at the Pallava court in Kāñcīpuram.20 The core territory of the Pallava Kingdom roughly corresponds to the modern-day state of Tamil Nadu. The Pallavas were a major military and cultural power in the Peninsula and Tamil merchants historically played a key role in maritime commerce with Southeast Asia and China.21 In the late 7th and early 8th centuries, the Pallavas’ relations with China in particular strengthened and it is recorded that the two states exchanged several diplomatic missions.22 The Pallava king Narasiṃhavarman II (r. 700–728) is credited with building a Buddhist temple in the Chinese style at the port city of Nāgapaṭṭinam for Chinese Buddhists arriving by sea.23 On the Southern Chinese shore, three temples (Brahmanic or Buddhist) for Tamil merchants were erected in the port city of Guangzhou (Ch. 廣州) in the 8th century or even earlier.24 Xuanzang (Ch. 玄 奘) visited Kāñcīpuram by land routes in the early 7th century and reported the presence of about 100 Buddhist temples and 10,000 Buddhist monks.25 Intent on propagating Buddhism in China, Vajrabodhi joined a Pallava diplomatic mission and travelled to the Tang court as an envoy of the Pallava king. Navigating along established sea routes through Sri Lanka and Sumatra, the mission arrived at Guangzhou, from whence they reached Luoyang (Ch. 洛陽) in the year 720. Vajrabodhi brought to China various Buddhist texts, including a short version of the Jingangding Yoga scripture 19. The story of ‘iron stūpa’ is found in Amoghavajra’s ‘Instructions on the Gate to the Teaching of the Secret Heart of the Great Yoga of the Scripture of the Diamond Summit’ ( 金剛頂經大瑜伽秘密心地法門義訣 Jin’gangding jing da yuqie mimi xindi famen yijue, T 1798 39: 808 a19–b28). See Orzech, Sørensen, and Payne 2011: 152. Amoghavajra’s account is translated integrally in Orzech 1995. See also Tanaka 2018: 309–311. 20. Sundberg and Giebel 2011: 135–137. 21. On the history of maritime relations between these regions, see Guy 2014: 3–13. 22. Sen 2016: 26; Sundberg and Giebel 2011: 144–145. 23. Sundberg and Giebel 2011: 145. 24. Guy 1994: 295. 25. Sundberg and Giebel 2011: 154.

3/2/22 6:19 PM

70

Yury Khokhlov

(a complete one-hundred-thousand-verse text of the scripture was purportedly lost on the way during a storm) and a copy of the Mahāprajñāpāramitā sūtra offered to the Chinese emperor by the Pallava king Narasiṃhavarman II.26 The master spent the remaining twenty-one years of his life in China propagating the Yoga tradition and working on translation projects.27 It is especially relevant for this study that Vajrabodhi was also known in China as a great painter, who was particularly good at painting Indian Buddhist images. He was included in the Record of Famous Painters of Successive Dynasties (Ch. 歷代名畫記, The Lidai minghua ji; 9.16b–17a), compiled by Zhang Yanyuan (Ch. 張彥 遠) in 847.28 The master himself painted numerous maṇḍalas. He also designed wooden sculptures for the Luoyang Guangfu Temple and a stūpa of Vairocana, which was considered an excellent piece of work. None of Vajrabodhi’s works are known to have survived, but it could be surmised that they were most likely executed in the Pallava style. Vajrabodhi’s disciple Amoghavajra, whom he met during his journey to China, is regarded as a towering figure in the history of Chinese Buddhism. Amoghavajra studied under Vajrabodhi for more than twenty years and became a prolific translator, second only to Xuanzang (Ch. 玄奘, c. 602–664), and an imperial preceptor for three successive Tang emperors: Xuanzong (Ch. 玄宗, r. 712–756), Suzong (Ch. 肅宗, r. 756–762), and Daizong (Ch. 代宗, r. 762–779).29 In 742–746, Amoghavajra travelled to Sri Lanka and Southern India, where he received several Esoteric initiations and acquired more than five hundred Buddhist texts, including a complete scripture of the eighteen texts of the Jin’gangding Yoga.30 26. Ibid.: 138, 147. 27. Yang 2018: 27–28. 28. Chou 1945: 276, n. 30, 280; Sundberg and Giebel 2011: 132, n. 12. 29. Goble 2012: 62. 30. Yang 2018: 33–35, 256–257. It is not known what regions Amoghavajra visited while in India; the sources are extremely vague about the Indian part of his trip and give detailed information only on his stay in Sri Lanka. Bearing in mind the shortness of the trip and that Amoghavajra spent most of his time in Sri Lanka, it is unlikely that he travelled beyond Southern India.

03_CreSouthV1_8P_3Feb22.indd 70

Amoghavajra made his first significant translations of the texts obtained during his journey, including the first part of the STTS, in the Hexi Corridor, in the regional capital city of Liangzhou (Ch. 凉州, modern Wuwei) between 754 and 756. He was sent to Hexi upon request of the military governor Geshu Han (Ch. 哥舒翰, 700–757) to assist the troops with ritual support in the battlefield against the Tibetans who constantly harassed the northwestern frontier.31 In Hexi, Amoghavajra was said to have initiated thousands of people into the practices of the Yoga tradition and gained many followers amongst the local nobility, government, and military. The Yoga teachings were propagated in Hexi on an unprecedented scale. This can be illustrated by an excerpt from Amoghavajra’s official biography composed upon his death in 774, which describes the events in 754: In the thirteenth year [of Tianbao], [Bukong] arrived in Wuwei, where he resided at the Kaiyuan Monastery. From the military commissioner to officials of the lowest rank, [Bukong] conducted the service of abhiṣeka for all; as many as several thousand people, both the gentry and commoners, ascended the altar [to receive the service of abhiṣeka]. [Bukong] imparted the teachings of the Five Divisions to his clerical disciple, Hanguang. Next, [he] administered the abhiṣeka ritual of the Five Divisions for Li Yuancong, now the Commissioner of Merit, [with the prestigious rank of] Opening Bureau, and instructed him in the teachings of the Great Maṇḍala of the Diamond Realm.32

It is important to note that Amoghavajra gained his renown as an Esoteric master in Hexi, and from there it spread to the capitals via Hexi officials and military commanders who were his disciples.33 In addition, Hexi came to be regarded as the master’s ancestral place; towards the end of his life, he was awarded the title of Duke of Su and given a nominal fief of 3,000 households in Suzhou (Ch. 肅州), a district in the northwestern part of the Hexi Corridor.34 31. Yang 2018: 41–42, 264–265. 32. Ibid.: 265. ‘The teachings of the five divisions’ refers here to the Yoga tradition, while ‘the Great Maṇḍala of the Diamond Realm’ stands for the Vajradhātu maṇḍala. See ibid.: 30. 33. Ibid.: 41–45. 34. Ibid.: 170–171, 362.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) During the turbulent events of the An Lushan rebellion and the war with the Tibetans, Amoghavajra significantly strengthened his religious and political influence at court. He performed countless Esoteric initiations and protective rituals for the imperial troops, the state, and the emperor. The institutionalization of a close group of his disciples as a form of governmental organization for the purpose of serving the state by performing apotropaic rituals can be regarded as one of Amoghavajra’s major achievements.35 At the same time, and as Robert Sharf has argued, the Esoteric practices propagated by Vajrabodhi, Amoghavajra and monks alike should not be regarded as an independent tradition/sect/school. What these masters promoted was essentially a new ritual technology intended for members of the ruling elite, who could afford spending considerable resources on elaborate rituals conducted for the simultaneous achievement of mundane and supermundane goals. Sharf has emphasized that in the 8th century these new teachings were transmitted to monastics from a variety of backgrounds and beyond sectarian lines imposed by later traditions.36 Amoghavajra and his Buddhist institution reached the peak of their power during the reign of Daizong (762–779). The master had a vast network of supporters and disciples, including the most powerful people of the empire; all the court officials and commanders of the imperial guards were required to be initiated into the Vajradhātu maṇḍala and to participate in rituals orchestrated by Amoghavajra.37 Among the master’s other deeds are the translation of a great number of Buddhist texts and the promotion of the cult of Mañjuśrī and Mount Wutai.38 Liangzhou, where Amoghavajra translated the STTS and initiated thousands of people into the Yoga tradition in 754–756, fell to the Tibetans in 758 and became a major centre of the Tibetan administration over the eastern areas.39 It is worth re35. Ibid.: 194–195. 36. Sharf 2017: 85–86, 114; for a more nuanced point of view, see Orzech 2006. 37. Yang 2018: 188–200. 38. Ibid.: 147–160. 39. van Schaik and Galambos 2012: 46.

03_CreSouthV1_8P_3Feb22.indd 71

71

membering that Tibetan Buddhism and the Tibetan Buddhist order were at this time in their early formative stage. Only in the 760s Buddhism started to be promoted by the Tibetan emperor Trisong Detsen (Tib. Khri srong lde btsan, r. c. 742–797); the first Tibetan monastery of Samye (Tib. Bsam yas) was consecrated only in 779.40 In general, the Tibetan ruling elite was highly suspicious towards Esoteric teachings, however the STTS and closely related texts (largely correspond to Yogatantras according to the Tibetan classification) were not only permitted, but, in parallel with the contemporaneous developments in Chinese state Buddhism, became a foundation for official Buddhist ceremony in the late 8th century (this will be discussed in detail in Part 4).41 It is also clear that Tibetan interest in the Yoga tradition emerged in close temporal proximity to Amoghavajra’s propagation of this tantric system in the Hexi Corridor, which soon after was taken by the Tibetan Empire. Amoghavajra died in 774, two years before the Tibetan conquest of Guazhou (Ch. , the prefecture containing the Yulin caves). With the conquest of Shazhou (Ch. , the prefecture containing Dunhuang) in 786, the whole Hexi Corridor came under Tibetan rule. Tibetans continued the patronage of existing Buddhist institutions in Hexi, which were numerous, and many were large, and sponsored building of several new monasteries and temples.42 However, Tibetan impact on Buddhist thought and Buddhist art in Hexi during this period of occupation remains largely unclear, since major developments with regard to the Tibetan appropriation of Buddhism took place during or after the conquest of the Hexi Corridor. One undisputable sign of the Tibetan input on Dunhuang art, however, is depictions of Tibetans and a Tibetan emperor in several caves at Mogao and Yulin.43 Significantly, a Chinese emperor continued to be depicted as well and his superior status was not diminished. For instance, in a painting in 40. Kapstein 2000: 60. 41. Weinberger 2010: 140–150. 42. Sha 2013: 152–156; van Schaik and Galambos 2012: 62. On Buddhist institutions in Hexi during the Tibet period, see Horlemann 2012. 43. Just a few examples: Yulin Cave 25 and Mogao Caves 158, 231, 359. See Sha 2013: 53–81.

3/2/22 6:19 PM

72

Yury Khokhlov

Mogao Cave 231 dated to circa 839, which shows the debate between Mañjuśrī and Vimalakīrti, a Chinese emperor retains his exalted position as the Buddhist authority and is shown in the entourage of Mañjuśrī, while a Tibetan emperor leads a procession of foreign/barbarian kings on the side of Vimalakīrti.44 The Dunhuang manuscripts mention some mixed Sino-Tibetan Buddhist communities and show no signs of religious opposition between Chinese and Tibetan Buddhists.45 Given the richness of Buddhist culture in Hexi and written records describing Tibetans eagerly seeking to learn the Dharma from the Chinese, the impact of Hexi Buddhist culture on what was to become known as Tibetan Buddhism and Tibetan Buddhist art must have been significant.46 The Tibetan period in Hexi lasted until the mid-9th century, when, following the collapse of the Tibetan Empire, most of the region returned under Tang control. Despite this political upheaval, many Tibetans remained in the region and the Tibetan language continued to be widely used.47 To summarize, when Tibetans arrived in Hexi, they found themselves in a highly sophisticated Buddhist environment, where Esoteric teachings, especially those of the Yoga tradition, enjoyed great prestige. The Buddhist art in Hexi was characterized by an increasing popularity of Indianized styles and Esoteric imagery. 44. For a discussion on the symbolism of representations of the Mañjuśrī–Vimalakīrti debate in Dunhuang, see Schmid 2006: 183–185. For a reproduction of the scene from Cave 231, see Debreczeny 2019: 50, fig. 1.2. 45. Sørensen 2000: 41. 46. For example, according to the Testament of Ba (Tib. Dba’ bzhed / Sba bzhed), a historical account on the origin of Buddhism in Tibet that was probably compiled in the 11th or 12th centuries, two Tibetan delegations were sent to the Chinese capital to learn Buddhist teachings in the second half of the 8th century. It is also stated that one thousand Buddhist texts were granted by the Chinese emperor and brought to Tibet by one of these delegations. See Wangdu and Diemberger 2000: 16–17, 21, 42 n. 91, 49 n. 121–122, 52 n. 136, 138, and 70 n. 245. Moreover, this source shows that Buddhism in 8th-century Tibet was strongly associated with China and sometimes referred to as the doctrine of China. See Wangdu and Diemberger 2000: 37, n. 71. 47. On a wide use of Tibetan language in Hexi and even in correspondence between non-Tibetan parties, see van Schaik and Galambos 2012: 67–68; Takata 2019.

03_CreSouthV1_8P_3Feb22.indd 72

part 2.  definitions and origins A list of paintings executed in the artistic mode discussed in this article is presented in Appendix 3.1. It is by no means exhaustive and includes only the most significant published examples. The majority of the paintings depict maṇḍala-like compositions. Common central figures include Vairocanābhisambodhi (see section 2.2),48 Vajrasattva, a Buddha, four-armed Amoghapāśa (see section 2.3), and two-armed Avalokiteśvara. As will be shown, these images mostly draw from the repertoire of the Yoga tradition. Other subjects include Vairocanābhisambodhi with the eight great bodhisattvas (see section 2.1) as well as Kubera with attendants (see section 2.4). In addition, the list includes portable paintings with solitary figures of bodhisattvas (see section 2.5) and paintings with stylistically mixed imagery (see section 2.6). Before proceeding to a detailed discussion of the paintings under review, it is necessary to mention that for almost all of them potential textual sources can be found in both the Chinese and Tibetan traditions. Since it has been historically maintained that Tibetan Esoteric Buddhism stems directly from the Indian tradition, possible relationships between the parallel Chinese and Tibetan texts have been rarely seriously considered. Moreover, it is commonly believed that developed Esoteric Buddhism did not exist in Dunhuang/Hexi before the Tibetan conquest. Thus, a decisive factor for the attribution of the discussed images to the Tibetan tradition has often been a combination of their Indian artistic mode (as opposed to the Chinese mode), which was understood as a sign of the Tibetan influence, with Esoteric subject matter. In the following, I will challenge the validity of this approach.

48. Following Kimiaki Tanaka, I use Vairocanābhisambodhi as a reference to bejewelled Vairocana displaying dhyāna-mudrā. See Tanaka 2018. This form was initially described in the Vairocanābhisambodhitantra (VAT). See Wayman and Tajima 1998: 8. However, images of Vairo-cana in this form do not necessarily signify the context of the VAT.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

2.1  representations of vairocanābhisambodhi with the eight great bodhisattvas in hexi and their spread to eastern tibet Yulin Cave 25 The scholarship on Yulin Cave 25 perfectly illustrates issues surrounding the attribution of the imagery under review to Tibetan influence and therefore I discuss it first and at greatest length. The cave was excavated during the period of Tibetan rule in Guazhou and is currently dated by the Dunhuang Academy to the early years of the Tibetan period (i.e. 776–781).49 The centre of the cave is occupied by a sculptural representation of Vairocana perhaps dating to the Qing dynasty, which could be a replica of the original sculpture. The wall paintings are regarded as originals. The central (eastern) wall of the main chamber is occupied by Indianized images of Vairocanābhisambodhi and the eight great bodhisattvas (Fig. 3.1).50 The bodhisattvas are identified by Chinese inscriptions as Ākāśagarbha, Kṣitigarbha, Maitreya, Mañjuśrī, Vajrapāṇi, Avalokiteśvara, Sarvanivāraṇaviṣkambhin, and Samantabhadra, who altogether are known as the eight great bodhisattvas.51 This group of nine—Vairocanābhisambodhi and the eight great bodhisattvas—will henceforth be abbreviated as VAEGB. The rest of the paintings in the cave are executed in the Chinese mode. They include the Pure Lands of Maitreya and Amitābha on the northern and southern walls, respectively; the bodhisattvas Mañjuśrī and Samantabhadra with attendants and Sudhana on the western wall; and 49. A review of scholarship on the attribution of the cave to the Tibetan period can be found in Kapstein 2009: 56–57. The date period 776–781 is provided by the Dunhuang Academy in https://www.e-dunhuang.com/ cave/10.0001/0001.0002.0025 (last accessed 23 November 2020). 50. Only four bodhisattvas on the left side have survived to the present day. The group on the right side can be seen in the old photograph taken in the 1940s that are currently preserved in the Lo Archive, Princeton, see Khokhlov 2019: fig. 1. 51. Wang 2018: 108–109. The most comprehensive study on the eight great bodhisattvas as a group is in BautzePicron 1997.

03_CreSouthV1_8P_3Feb22.indd 73

73

six standing bodhisattvas, whose identities are not entirely clear, and who are depicted on the corners of the cave. These Indianized and Chinese images show identical line structures and shading, and the faces of some figures between the two groups bear a striking resemblance.52 From this it is clear that, differences in mode notwithstanding, all the paintings in the cave were likely executed by the same group of artists. A publication produced in 2000 by the Dunhuang Academy for a centennial commemoration of the discovery of the library in Mogao Cave 17 briefly mentions that the central composition (VAEGB) of Yulin Cave 25 ‘indicates close ties to Master Amoghavajra’s spread of Tantrism in the Hexi region’.53 This idea has been ignored in later scholarship. Instead, the coexistence of the two artistic modes as well as the coexistence of Esoteric and traditional Pure Land subjects within the cave has been ascribed to the cultural influence of Tibet. The style of the composition of VAEGB has been understood either as the Tibetan style or as the Pāla style of Northern India being transmitted through Tibet.54 In the following, I will argue against the 52. See Khokhlov 2019: fig. 3. 53. Unfortunately, no arguments supporting this attribution were given. See Zhang 2000: 31. 54. The publications arguing for the Tibetan background of the composition include Kapstein 2004, 2009, 2014; Sha 2013; Wang 2018; Debreczeny 2019. In a series of articles, Matthew Kapstein has developed a theory that Yulin Cave 25 should be identified as a temple built by Tibetans to commemorate a peace treaty between the Tibetan and Tang empires made in 821–822. Although the iconographic programme of Yulin Cave 25 only partially matches the description of the temple found in the Dunhuang manuscripts, the author develops his idea by pointing to the popularity of Vairocana cult in the Tibetan Empire and to the existence of roughly contemporaneous rock carvings of VAEGB in Eastern Tibet (Kapstein 2009: 53, 58). Later on, I will discuss these Tibetan rock carvings in detail and show that they were most likely created under the influence from the Hexi Corridor. As stylistic evidence, the author compared the figure of Vairocana from Cave 25 to an undated sculptural representation of the same deity that was tentatively ascribed to the Imperial period by other scholars. The sculpture, however, is executed in a considerably different style, which makes it difficult to accept the author’s point. Summarizing his argument, Kapstein emphasizes the coexistence of the two styles in the cave and concludes: ‘In both iconography and style, Anxi

3/2/22 6:19 PM

74

Yury Khokhlov

Fig. 3.1: The extant part of the wall painting Vairocana and the eight great bodhisattvas, eastern wall of the main chamber, late 8th century, Tibetan period, Yulin Cave 25, China (with the permission of the Dunhuang Academy).

Yulin 25 thus expresses the coexistence of the Tibetan and Chinese Buddhist worlds. On the surface at least, it seems an especially fitting memorial to a peace-accord between the two powers’ (Kapstein 2009: 54). Recently, the author revised his initial notion and suggested instead that Yulin Cave 25 could be either an earlier Tibetan model for the treaty temple or its copy (Kapstein 2014). As for the coexistence of Indianized and Chinese depictions, as well as the coexistence of Esoteric and Pure Land subjects, such mixed imagery can be already found in Tang dynasty caves predating the Tibetan period (see examples in the next section). Sha Wutian’s research on Cave 25 was published in his monograph devoted to Dunhuang art of the Tibetan period. The author postulates that the figures of VAEGB in Cave 25 represent the Tibetan style, and more specifically, the

03_CreSouthV1_8P_3Feb22.indd 74

northeast Indian Pāla style as transmitted through Tibet (Sha 2013: 360–382, 470–492). These statements seem to ignore the fact that no comparable depictions that can be dated to the period before the Tibetan occupation of Hexi are known from Tibet. Moreover, the discussed material certainly does not belong to the Pāla artistic tradition. Michelle Wang in her recent publication Maṇḍalas in the Making came to the conclusion that the depiction of VAEGB in Cave 25 belongs to the Tibetan Buddhist tradition and is executed in the Tibetan style, which she defines as a ‘Nepalese-derived Tibetan style’. The author compared the images from the cave to Tibetan and Nepalese sculptures and noted various similarities. Significant differences were, however, overlooked. The Eastern Tibetan rock carvings and the coexistence of the two artistic modes

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) current scholarly consensus and demonstrate that neither such thematic and stylistic diversity nor the style of the group of VAEGB can be ascribed to the Tibetan or Pāla-Indian influence.

stylistic considerations Indian and Chinese Modes in Hexi Art before the Tibetan Conquest Indianized imagery started to appear early on in Tang dynasty caves.55 Striking examples can be seen in Mogao Cave 329 that belongs to the early Tang period (618–712). A painting on the east wall depicts two bodhisattvas, which are executed in two different artistic modes.56 The bodhisattva on the right is shown in the Chinese mode, which is characterized by a round halo, long scarves wrapped around the body, and a voluptuous lower garment. The bodhisattva on the left, on the other hand, is shown in the Indian mode. It is comparable to what is found in Yulin Cave 25 (Fig. 3.1), where the figures also have inverted U-shaped halos and wear similar lower garments resembling trousers. Another example of such coexistence of the two modes can be found in Mogao Cave 148 that contains depictions of the Buddha’s Parinirvāṇa and various Pure lands, which are painted in the Chinese mode, and Esoteric imagery (e.g. various multi-armed forms of Avalokiteśvara) painted in the Indian mode. According to the engraved inscription on the stele at the entrance, the cave was completed sometime between 766 and 776—that is within the decade leading up to the Tibetan conin Cave 25, named as ‘stylistic bilingualism’, were regarded as evidence for the Tibetan origin of the cult of VAEGB in Hexi (Wang 2018: 91–121). In the latest contribution to this subject, Karl Debreczeny sees the group of VAEGB in Yulin Cave 25 as a manifestation of the new aesthetic and religious interests introduced under Tibetan rule (Debreczeny 2019: 24–25). The author’s argument is essentially based on the works by Kapstein, Sha and Wang and therefore has the same shortcomings. 55. For example, Mogao Caves 202, 321, 329, and 335. See Sha 2013: 133–134. 56. Khokhlov 2019: fig. 4. For a description of this cave and additional images, see https://www.e-dunhuang.com/ cave/10.0001/0001.0001.0329 (last accessed 22 April 2019).

03_CreSouthV1_8P_3Feb22.indd 75

75

quest of Guazhou in 776.57 Particularly interesting are six rectangular panels painted inside the niches. They contain compositions akin to that of VAEGB in Yulin Cave 25.58 It is thus evident that the coexistence of the Indian and Chinese artistic modes as well as of Pure Land and Esoteric imagery within Yulin Cave 25 is a continuation of the trend that started in Dunhuang earlier. Similarly, the depiction of the group of VAEGB itself seems to be based on a template that was popular in Hexi before the Tibetan conquest.

Southern Indian Artistic Mode Esoteric teachings were transmitted not only through texts, but also through their accompanying images.59 Although none of the original Indian illustrations survive in China, the fact that they existed is evident from Chinese translations made by Indians. For example, Śubhākarasiṃha (Ch. 善 無畏, 637–735), an Indian translator-monk living in Tang China, frequently made references to iconographic illustrations in his texts.60 In addition, the importance of imagery in the transmission of Esoteric teachings was stressed by Kūkai (Ch. 空海, 774–835), who studied in China under Amoghavajra’s successor Huiguo (Ch. 惠果, 746–806). In his inventory (dated 806), Kūkai states: ‘The secret wisdom of the true words is hidden and implicit in scriptural texts; only through imagery can it be 57. Mogao Cave 148 is discussed in Lee 2010: 181–201. Esoteric imagery in the cave is discussed in Liu 2008. 58. See, for example, a panel reproduced in Khokhlov 2019: fig. 5. This panel depicts bodhisattva Bhaiṣajyarāja surrounded by eight bodhisattvas and two heavenly kings. It has a similar layout to the image in Yulin Cave 25 (Fig. 3.1), with a large central figure placed under a canopy and flanked by two groups of smaller figures shown symmetrically on either side. The lotus seat of the central figure and the two chief bodhisattvas are almost identical to the lotus seats in the Yulin composition. In both paintings, the stems of the lotus seats of the bodhisattvas grow from the seat of the central divinity. Small lotus flowers and leaves depicted in the intermediate space are very similar in both compositions as well. 59. This has been recently discussed by Kimiaki Tanaka, see Tanaka 2018: 125–126. 60. Ibid.: 125–126.

3/2/22 6:19 PM

76

Yury Khokhlov

properly transmitted’.61 In China, original Indian illustrations were copied by local artists, and, as in the case of Vajrabodhi, many Indian monks themselves were able to produce such images. In any case, works with Indian background were considered extremely valuable; they were carefully recorded and praised in monastic inventories. Furthermore, accuracy in copying these images was regarded crucial for the authenticity and efficacy of replicas in Esoteric rituals.62 Another important feature of Chinese practice was the standardization of religious imagery. Standardized models and templates for replication were known as yang (Ch. 樣). In her recent book, Hsueh-man Shen (Ch. 沈雪曼) describes many Tang dynasty scrolls with yang of iconic images of various deities, including a group of the eight great bodhisattvas.63 It is through these processes of copying, standardization and replication that elements of Indian costumes and decorations must have been adopted in Chinese art. In this regard, tracing the origins of distinctive decorative elements in early Chinese Esoteric imagery yields an important trace for determining the provenance of the artistic tradition to which Indian prototypes belonged and, by extension, the provenance of the corresponding teachings. The group of VAEGB in Cave 25 was likely based on standardized models. The bodhisattvas are shown either frontally or in a three-quarter profile and all have almost identical decorations. Diversity is achieved through different hand gestures, leg positions, and attributes. The figures are drawn in a distinctive artistic mode, which can be recognized by certain peculiarities (see an enlarged image of Vairocana in Fig. 3.2) such as: – elongated and Y-shaped torsos; – crowns and armbands with spikes of pearls or little spears (henceforth spiked crowns and armbands, see close-ups in Fig. 3.2);

These elements are present in the group of paintings which this article is concerned with. The discussed examples belong to different periods (from the 8th to 11th century) and their styles are different as well. In addition, it is not always the case that all of the features listed above are present in a given image. Of the four characteristics, the spiked decorations are the most distinctive and persistent feature. It is a matter of fact that such decorations are almost exclusively associated with Southern Indian art (see examples in Figs. 3.3, 3.9, 3.18, 3.19). In this regard, they are neither found in Tibetan art prior to the occupation of Hexi, nor in the art of Northern India, Nepal, or Central Asia, artistic traditions which, along with that of China, contributed to the formation of Buddhist art in Tibet. As such, this indeed eliminates the currently assumed possibility of the alleged Tibetan background of the discussed imagery. In contrast, these features correspond to the Southern Indian current in Chinese Esoteric Buddhism represented by Vajrabodhi and Amoghavajra. In the 7th and 8th centuries, there were two regions in Southern India where Esoteric Buddhism was prominent and these exact kinds of spiked jewelry were common in artistic representations. They are Tamil Nadu and the Western Deccan. With regard to Sri Lanka, which was an important destination for both Vajrabodhi and Amoghavajra, it is unlikely that the discussed artistic mode in Hexi art has a Sri Lankan origin. Although Sri Lankan Buddhism and associated art were greatly influenced from the Pallava Kingdom of Tamil Nadu,64 few Sri Lankan images can be compared to the discussed Hexi examples.65

61. Shen 2019: 154. 62. Shen shows that the inventories of objects taken from China to Japan during the Tang specifically comment on images produced by Indian masters or originated from India, see Shen 2019: 154. The earliest extant example of iconographical drawings is an early 9th-century Chinese scroll known as the gobu shinkan (Ch. 五部心觀), see Tanaka 2018: 293–294. 63. Shen 2019: 149–155.

64. On the Pallava Buddhist and artistic influence in Sri Lanka, see Indrapala 2005: 189–192; Dohanian 1983. 65. On Sri Lankan Mahāyāna imagery, see Dohanian 1977. The most comprehensive research on Sri Lankan Buddhist sculptures can be found in von Schroeder 1990. It has to be noted, however, that remaining Esoteric Buddhist material is only lately being studied and the chronology of Sri Lankan Buddhist art of that period is debatable because of a lack of firmly datable material.

03_CreSouthV1_8P_3Feb22.indd 76

– broad necklaces with swags of pendants (henceforth swagged necklaces); and – earrings with large and flat pendants.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

77

Fig. 3.2: Vairocana and close-ups of the crown and armband, detail of Fig. 3.1.

The connections with the art of Tamil Nadu are not surprising as the region was the home of Esoteric teachings propagated by Vajrabodhi and Amoghavajra.66 In addition, Vajrabodhi as a painter most likely belonged to the Pallava artistic tradition that flourished in Tamil Nadu, where the master spent many years studying Esoteric teachings before travelling to China. Having in mind the records about imagery produced by Vajrabodhi in China and the Chinese practices of copying, 66. On Esoteric Buddhism in Tamil Nadu, see Yang 2018: 179–180, n. 370. In particular, on Kāñcīpuram as a centre of Esoteric Buddhism, see fn. 17. Esoteric Buddhism in Tamil Nadu was probably part of the larger tradition encompassing the neighbouring Andhra region, the motherland of the Pallavas. On the development of Esoteric Buddhism in Andhra, see Barber 2008.

03_CreSouthV1_8P_3Feb22.indd 77

standardization, and replication, we can assume that a Chinese artistic tradition based on Pallava art and closely associated with Vajrabodhi emerged in the Tang capitals by the mid-8th century and spread from there to Hexi/Dunhuang. The religious and artistic interexchange between Dunhuang and the Tang capitals is well recorded67 and thus Vajrabodhi’s teachings and related pattern books and iconographic sketches likely reached Hexi already during the master’s lifetime. Nevertheless, the tradition undoubtedly owes its greater popularity in the region to Amoghavajra. As mentioned, the master not only taught in Hexi, but also made his first major translations there. 67. Fraser 2004: 52–53.

3/2/22 6:19 PM

78

Yury Khokhlov

Regarding the Western Deccan, the existence of Esoteric Buddhism in the 8th century in that locale is evident from Ellora cave sculpture. It has been also suggested that the region could have been the origin place of such fundamental scripture as the VAT.68 At the same time, numerous cross-regional military raids of the Pallavas (ruling in Tamil Nadu), and the Cālukyas (ruling in the Deccan) in the 7th and 8th centuries intensified cultural exchange across the peninsula and it is not unlikely that the respective Buddhist communities were in conversation, influencing each other and sharing major scriptures.69 For example, the earliest surviving images displaying iconography that could be related to the STTS are found in the Western Deccan, at Ellora Cave 12 (early 8th century).70 In general, Ellora sculptures from the 8th century onwards reflect increasing influence from the South of the Peninsula, and in particular the Pallava influence.71 Regrettably, we do not know what specific regions Amoghavajra visited while in India; the sources are extremely vague about the Indian part of his trip and give detailed information only on his stay in Sri Lanka.72 It is fair to assume, however, that Tamil Nadu, a place where Vajrabodhi studied the Yoga tradition before going to China, would have been Amoghavalra’s prime destination, though a small possibility that he visited other regions, including the Western Deccan, cannot be entirely excluded. In addition, it is possible that materials collected by Amoghavajra in Sri Lanka were of an Indian origin, as the introduction of Esoteric Buddhism to Sri Lanka is associated with Tamil Buddhists who migrated to the island in the 7th and 8th 68. On the development of Esoteric imagery in the Western Deccan, see Huntington 1981; Malandra 1993; Bautze-Picron 2000. On possible geographical origins of VAT, see Wayman and Tajima 1998: 11–12. 69. On artistic and architectural connections between these regions see Dayalan n.d. 70. Tanaka 2018: 265, fig. 4.9. 71. Malandra 1993: 116–117. The most striking example of Pallava influence is the famous Kailāsanātha monument, which was built in the mid-8th century and partially patterned on Kailāsanātha temple in Kāñcīpuram. 72. Yang 2018: 256–257. Sundberg and Giebel even suggested that Amoghavajra might have limited his travel to Sri Lanka only. See Sundberg and Giebel 2011: 148–151.

03_CreSouthV1_8P_3Feb22.indd 78

centuries fleeing from the antagonistic movements in Śaivism and Vaiṣṇavism that were specifically directed against Buddhists and Jains.73 As a result of the anti-Buddhist campaigns, only a handful of Pallava Buddhist images survived. Most of them have been vandalized and were found in the vicinity of Hindu temples that suggests a violent expropriation and conversion of Buddhist establishments.74 For our purposes, the Pallava stylistic features can be reconstructed from abundant examples of Brahmanic imagery. For instance, a Pallava sculpture in Fig. 3.3 possesses all major characteristics of the discussed artistic mode in

Fig. 3.3: Śiva as Gaṅgādhara (detail), stone relief, 7th century, Pallava period, Lalitankura cave, Tiruchirapalli, Tamil Nadu, India. (Photo courtesy of Emmanuel Francis) 73. Indrapala 2005: 181, 192–193; Yang 2018: 180–181; Davidson 2002: 192. 74. Rao 1915; Verardi 2018: 296ff., fig. 4.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

79

Fig. 3.4: Eight bodhisattvas, early 8th century, first floor of Ellora Cave 12, Maharashtra, India. (Photos courtesy of Deepanjana and Arno Klein)

Hexi art and show remarkable similarities to the aesthetics of the Yulin images (Fig. 3.1): elongated and Y-shaped torsos, tall headdresses, spiked crowns and armbands, and large earrings.75 In the Western Deccan, Buddhist material heritage largely survived. Ellora Caves 11 and 12, which are traditionally dated to the early-8th century, represent the last large-scale Buddhist undertaking in the region.76 The eight great bodhisattvas as a group assume prominent po75. See another Pallava example in Khokhlov 2019: fig. 8. 76. For a detailed analysis of Buddhist art at Ellora in general and these caves in particular, see Malandra 1993.

03_CreSouthV1_8P_3Feb22.indd 79

sitions in both caves. While Cave 11 contains two groups of the standing bodhisattvas, Cave 12 has one group of the seated and two groups of the standing bodhisattvas. Fig. 3.4, from the first floor of Cave 12, shows two groups of the four bodhisattvas flanking a statue of the Buddha with dharmacakra-mudrā. Their identities were reconstructed by Malandra and Bautze-Picron according to their attributes and they appear to be the same bodhisattvas as the bodhisattvas in Yulin Cave 25.77 The Ellora and Yulin bodhisattvas are visually similar as well and convey the Pallava 77. Malandra 1996: 197; Bautze-Picron 1997: 10–11.

3/2/22 6:19 PM

Yury Khokhlov

80

Fig. 3.5: Detail of Fig. 3.1 (left); detail of Fig. 3.4 (right).

aesthetics (see comparison in Fig. 3.5). They have elongated and Y-shaped torsos, crowns with spikes, and swagged necklaces. The armbands display similar leaf-like elements decorated with spikes. Spikes in the armbands of the sculptures are not clearly visible in the pictures as the decorations were finished with plaster, which is largely lost now. It is evident that stylistically the images in Yulin Cave 25 depended on Southern Indian prototypes and that the defining characteristics of the discussed artistic mode have a Southern Indian origin. Therefore, from here onwards this mode in Hexi art will be labelled as the Southern Indian mode.

iconographical considerations Ellora Caves 11 and 12 and their Iconographic Innovations According to Geri Malandra, Ellora Caves 11 and 12 reflect Esoteric teachings, which were new for the

03_CreSouthV1_8P_3Feb22.indd 80

Deccan, and this new iconography was executed in a new style that shows a noticeable influence from the South of the Peninsula.78 She suggested that the teachings reflected in the caves were imported to the Deccan and that they might have had the same origin as the sculptures’ new style. In support for her theory, Malandra noted that contemporaneous or slightly later representations of the eight great bodhisattvas are found in Odisha and Java that points to the Southern (as opposed to the Northern) source of the innovations.79 Indeed, Northern Indian depictions of the eight great bodhisattvas mostly date from the 9th century and later.80 In addition, the popularity of the cult of the eight great bodhisattvas in the 8th or 9th centuries is evident in the Malay Peninsula,81 as well as in 78. Malandra 1993: 116–117; Malandra 1996: 206. 79. Malandra 1996: 205–206. 80. Bautze-Picron 1997: 14–27. 81. Skilling 2011: 375–377; Sharrock and Bunker 2016: 244; Acri 2018.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) China (discussed in this Part and in Part 4), where it was initially introduced by Puṇyodaya (Nati 那提, fl. 650s) in the 7th century probably from Southeast Asia (Cambodia).82 This distribution implies that the cult of the eight great bodhisattvas was popular in the eastern coastal areas of the Indian Peninsula that historically played a key role in maritime relations of Southern Indian polities with Southeast Asia and China. Moreover, the Pallava-influenced style of the Dunhuang depictions points specifically to Tamil Nadu as a source of the cult in China. Similarly, the Pallava-influenced style of the Ellora sculptures should be regarded as evidence for a Tamil Nadu origin of the respective teachings in the Western Deccan as well. As discussed, travelling monks and pilgrims were instrumental in the transmission of new teachings and their art forms. Ellora was an important holy site (Skt. tīrtha), a place of pilgrimage that attracted people from distant lands and different religious communities; its sacred status encouraged commissioning of religious imagery at the site. Lisa Owen has recently shown, for instance, that Ellora Jain caves reflect activities of Jain communities associated with Tamil Nadu and Karnataka.83 Although donors of the discussed Buddhist caves remain unknown, there is a possibility that they hailed from the South as well.84 It is also possible that some Buddhist communities migrated from Tamil Nadu to the Western Deccan during the period of interreligious antagonism in their homeland.

82. Acri 2018. Also see fn. 86. 83. Owen 2012: 16, 20, 160–163. 84. Buddhist monks from Tamil Nadu and Andhra were active even in Northern India. Bautze-Picron has recently discussed a number of stone and bronze images from Kurkihar dating from the 8th to 11th century, which bear inscriptions mentioning the donors who hailed from Kāñcīpuram and Andhra. Notable among these are two circa 9th-century stone slabs that show a set of the eight great bodhisattvas around the Buddha. See Bautze-Picron 2015: figs. 91, 92. In a personal communication on 9 August 2019, Bautze-Picron remarked that these slabs would seem to suggest that the cult of the eight great bodhisattvas might have been introduced to Bihar from Tamil Nadu and Andhra.

03_CreSouthV1_8P_3Feb22.indd 81

81

The Aṣṭamaṇḍalakasūtra In addition to the sculptural groups of the eight great bodhisattvas, Ellora Cave 12 contains five wall reliefs of maṇḍalas of the eight great bodhisattvas centred around a dhyāna-mudrā Buddha, who was identified by scholars as Vairocana (Fig. 3.6). The bodhisattvas in the maṇḍalas were recognized by their attributes and they also happen to be the same bodhisattvas as those depicted in Yulin Cave 25.85 According to Kimiaki Tanaka, the group of Vairocana and the eight great bodhisattvas ultimately originates from the Aṣṭamaṇḍalakasūtra (‘The sūtra of the maṇḍala of the eight great bodhisattvas’), which is also the most probable source of the maṇḍalas in Ellora Cave 12.86 The central figure of the Tathāgata is only briefly described in the text87 and this perhaps explains an iconographic variety of his representations. He can be depicted unadorned, like in the Ellora maṇḍalas (Fig. 3.6), or bejewelled, like in the Yulin composition (Fig. 3.1), and with the bhūmisparśa-, dharmacakra- or

Fig. 3.6: Maṇḍala of the eight great bodhisattvas, early 8th century, Ellora Cave 12, Maharashtra, India. (Photo courtesy of Deepanjana and Arno Klein) 85. On the identification of the figures in these maṇḍalas, see Malandra 1996: 197, 201; Tanaka 2018: 141–144. 86. Tanaka 2018: 92–93. 87. On the description of the eight great bodhisattvas and Buddha in the text, see de Visser 1915: 15–17.

3/2/22 6:19 PM

Yury Khokhlov

82

dhyāna-mudrā. Examples of the group with the central figure displaying the dhyāna-mudrā are relatively rare in India and found only in the southern part, namely in the Western Deccan (Ellora) and Odisha (Ratnagiri).88 Thus, it is evident that the iconography of the group depicted in Yulin Cave 25 was particularly common in Southern India. The Aṣṭamaṇḍalakasūtra was translated by Amoghavajra (T 1167)89 and the central deity of the maṇḍala in his text is described as ‘Tathāgata in His True Golden Form Body’ (如來真金色身).90 This epithet perhaps refers to Vairocanābhisambodhi, a form of Vairocana, which initially appeared in the VAT and is described there as having a golden body (in contrast, for example, with the Vairocana of the STTS, whose colour is white), bejewelled and displaying the dhyāna-mudrā.91 This is precisely how Vairocana is depicted in Yulin Cave 25. To summarize, in both style and iconography the Yulin group of VAEGB is closely related to the Southern Indian examples (Western Deccan, Odisha).

The popularity of the eight great bodhisattvas in Southern Indian Buddhism is reflected in Amoghavajra’s textual heritage. In addition to the translation of the Aṣṭamaṇḍalakasūtra, Amoghavajra integrated the eight great bodhisattvas into practices of such popular prayers as the Bhadracarī and the Uṣṇīṣavijayādhāraṇī.92 These combinations were

probably meant to employ the power of the bodhisattvas to magnify the efficiency of the prayers. According to the Aṣṭamaṇḍalakasūtra, the mantras and maṇḍala of the eight great bodhisattvas are effective in fulfillment of one’s wishes, eradicating one’s sins, setting one on the path to swift enlightenment and protection from harm.93 I believe that, although the Aṣṭamaṇḍalakasūtra is the ultimate source for the VAEGB group, the depiction of VAEGB in Yulin Cave 25 is not a maṇḍala, as the bodhisattvas do not encircle the central figure, and is actually based on another work by Amoghavajra entitled Puxian pusa xing yuan zan (Ch. 普賢菩薩行願贊) (T 297).94 The text was composed between 756 and 774 and consists of four parts, where the first part is a translation of the Samantabhadracaryāpraṇidhānarāja (Ch. 普賢行願品), also known as the Bhadracarī;95 the second part is a eulogy to the eight great bodhisattvas, which is a part of the Aṣṭamaṇḍalakasūtra (T 1167); and the third and final parts are a dhāraṇī and a praise of this dhāraṇī. In short, the work includes the cult of the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra into the practice of the Bhadracarī, an immensely popular aspirational prayer, in which one vows to worship the buddhas and become enlightened for the sake of all beings. The Bhadracarī is known as an independent work and as the final part of the Gaṇḍavyūhasūtra (Ch. 入法界品),96 which itself is a part of the Ava-

88. Northern Indian examples show Buddha/Vairocana either with the bhūmisparśa- or dharmacakra-mudrā. See Bautze-Picron 1997; Tanaka 2018: 81–82. 89. The sūtra was initially translated into Chinese by Puṇyodaya in the 7th century (T 486), but it became particularly popular in the translation by Amoghavajra, who assigned to the maṇḍala of the eight great bodhisattvas much greater importance, see Lin 1935. There are two Tibetan versions of the sūtra; both are believed to be translated during the Imperial period, see Skilling 2011: 376, n. 9, 10, 11. 90. I am grateful to Geoffrey Goble for sharing this information. Personal communication, 25 July 2018. 91. For a description of Vairocana in the VAT, see Wayman and Tajima 1998: 101. The relation between the maṇḍala of the eight great bodhisattvas and the VAT was earlier suggested by Malandra and Tanaka. See Malandra 1996: 76; Tanaka 2018: 77–78. 92. See Amoghavajra’s works T 297 and T 972. On

Amoghavajra’s ritual manual for the Uṣṇīṣavijayādhāraṇī sūtra (T 972), see Wang 2018: 47–50; Lin 2014: 139–143. Amoghavajra produced the manual in 764 and it became the most influential redaction in Tang China (several earlier redactions are also known, including the one by Śubhākarasiṃha). The manual introduces a maṇḍala known in Japanese Buddhism as Sonshō Maṇḍala in the Amoghavajra tradition. This maṇḍala is similar to the maṇḍala of the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra, but, in addition, it includes two wrathful deities, Acala and Trailokyavijaya. See Tanaka 2018: 88, 95. 93. Granoff 1968/1969: 2; Geoffrey Goble, private communication, 25 July 2018. 94. On this text, see Dessein 2003. 95. The Bhadracarī is discussed in Osto 2010. 96. The Bhadracarī is included in the translation of the Gaṇḍavyūhasūtra made by Prajña (Ch. 般若), which was finished in 798 and was based on the Sanskrit copy that

The Bhadracarī

03_CreSouthV1_8P_3Feb22.indd 82

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) taṃsakasūtra (Ch. 華嚴經), a core scripture of the Avataṃsaka (Ch. 華嚴 Huayan) Buddhist tradition.97 According to modern scholarship, the Vairocana of Esoteric Buddhism developed from the Vairocana of the Avataṃsakasūtra, who is featured in the sūtra under the special name of Rocana (Ch. 盧舍那 Locana). The Huayan tradition integrated various Esoteric teachings, and the Huayan Vairocana and the Vairocana of Esoteric Buddism were sometimes ‘fused iconographically’.98 The image of Vairocana in Yulin Cave 25 is an example of such fusion. A cartouche next to it contains the Chinese inscription 清淨法身盧那舍佛 (the Pure Dharmakāya Locana Buddha), which is the name of the Vairocana of the Avataṃsakasūtra.99 Thus, we have Locana Buddha accompanied by the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra. This combination corresponds to the Puxian pusa xing yuan zan (Ch. 普賢菩薩行願贊) (T 297), where the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra are understood as guardians of Locana of the Avataṃsakasūtra.100 Therefore, the Yulin 25 painting (Fig. 3.1) represents the deities of the Aṣṭamaṇḍalakasūtra in the Huayan context. In contrast with the Aṣṭamaṇḍalakasūtra, which prescribes a strictly circular layout for the bodhisattvas, the Puxian pusa xing yuan zan does not provide any guidance in this regard. I believe, this explains a variety of spatial arrangements found in the representations of the group (discussed below). Furthermore, not only the central composition, but also the whole iconographic programme of Yulin Cave 25 can be understood in the Huayan context. First of all, the characteristic Huayan depictions of Mañjuśrī, Samantabhadra, and

had been presented to the Chinese emperor by the King of Odisha. The extant Tibetan translation of the Gaṇḍavyūha was made during the imperial period and also concludes with the Bhadracarī. See Osto 2010: 2–3. 97. See Cleary 1993 for a translation of the sūtra. 98. Orzech, Sørensen, and Payne 2011: 91–92. 99. Wang noted that the characters na 那 and she 舍 have been transposed. See Wang 2018: 109–110. 100. Dessein 2003: 332. As was noted by Wang, Dessein mistakenly replaces bodhisattva Sarvanivāraṇaviṣkambhin with Acalanātha, see Wang 2018: 258, n. 59.

03_CreSouthV1_8P_3Feb22.indd 83

83

Sudhana occupy the western wall.101 The depictions of Amitābha and Maitreya’s Pure Lands as a pair are common in Dunhuang art and their compositions are based on respective sūtras.102 However, their presence in Cave 25 might reflect an additional level of symbolism. For example, aspiration for rebirth in Amitābha’s Pure land is emphasized in the Bhadracarī,103 while Maitreya is one of the most important spiritual friends of Sudhana, who visits Maitreya’s abode in the narrative of the Gaṇḍavyūhasūtra. Therefore, Yulin Cave 25 can be regarded as a manifestation of the donor’s practice of the Bhadracarī and of their aspiration for enlightenment. As a matter of fact, references to the Bhadracarī and/or the Gaṇḍavyūhasūtra can be found in almost all representations of VAEGB in Hexi and Tibet.104

A Wooden Shrine One striking example is a wooden shrine from the Nelson-Atkins Museum of Arts (Fig. 3.7). The shrine has attracted considerable scholarly attention and was variously ascribed to Central Asia, Kashmir, or Tibet.105 On stylistic grounds, I argue that it most likely originates from Hexi, as is evident from Chinese facial features of the figures, the com101. For these images, see Khokhlov 2019: fig. 2. Paired images of Mañjuśrī and Samantabhadra, the major bodhisattvas of the Huayan tradition, were often depicted in Dunhuang caves starting from the early Tang dynasty. According to Wang, this pair reflects the interest in the Gaṇḍavyūhasūtra narrative and the Bhadracarī. See Wang 2018: 247. 102. For example, Mogao Caves 329 and 321. See Wang 2005: 382–386. 103. Osto 2010: 5–6. 104. This also finds parallels at Borobudur (early 9th century), Java. Panel IV-3 from the Gaṇḍavyūha reliefs contains images of Sudhana and the eight great bodhisattvas. See Bautze-Picron 1997: 28, 54 schema 11; Fontein 2012: 173; Long 2009: 165–168. I am very grateful to Claudine Bautze-Picron for drawing my attention to the Borobudur depiction and providing these references. 105. The most notable publications include Granoff 1968/1969; Linrothe 2014; Wang 2018. While Granoff ascribes the shrine to Central Asia, Linrothe suggests a Kashmiri provenance, and Wang argues for a Tibetan origin. See Granoff 1968/1969: 81; Linrothe 2014: 31; Wang 2018: 83.

3/2/22 6:19 PM

84

Yury Khokhlov

Fig. 3.7: Traveling Shrine of Vairocana and Eight Bodhisattvas on a Garbhadhātu maṇḍala, Chinese, 9th or 10th century CE, Tang Dynasty (618–906 CE). Sandalwood (?) with traces of polychromy and brass paint, opened: 12 1/4 × 14 inches (31.1 × 35.6 cm). The Nelson-Atkins Museum of Art, Kansas City, Missouri. Purchase: William Rockhill Nelson Trust, 44–18. Photo: Joshua Ferdinand. Right: close-up of Vairocana’s armband.

bination of the Indian and Chinese artistic modes, and spiked armbands worn by Vairocana, which is a characteristic of the Southern Indian mode in Hexi art (see a close-up of the armband in Fig. 3.7). The central section of the shrine shows Vairocana flanked by two vertical rows of four bodhisattvas, who altogether most likely represent the eight great bodhisattvas.106 Such a layout can be found in Southern Indian art from the 6th century onward and this particular group is depicted in the same manner in stone stele from Odisha, which has been tentatively dated to the 9th or 10th century.107 In the shrine, three more figures are placed below the figure of Vairocana. While Granoff and Wang identify this group as a representation of an abhiṣeka ceremony, Linrothe argues that it depicts a donor of the shrine being blessed by the bodhisattvas.108 I would 106. They were identified as such by Granoff. See Granoff 1968/1969: 90–91. 107. For 6th-century examples in Cave 23 at Nasik, see Bautze-Picron 2000: figs. 9–10. For the Odisha stele, see Bautze-Picron 1997: 27, 53 schema 10; Tanaka 2018: fig. 1.15. 108. Granoff 1968/1969: 90; Linrothe 2014: 31; Wang 2018: 72.

03_CreSouthV1_8P_3Feb22.indd 84

argue that this group likely symbolizes Sudhana’s spiritual path towards enlightenment. The two bodhisattvas thus represent Mañjuśrī (on the left) and Samantabhadra (on the right), who together with a central figure of Vairocana (here the Vairocana of the Avataṃsakasūtra) form the typical trinity of Huayan Buddhism. Therefore, a figure seated on a rug represents Sudhana, who is customarily depicted in Chinese art as a little boy with a shaven head.109 Accordingly, the depiction of Sudhana together with Mañjuśrī and Samantabhadra symbolizes his spiritual journey as described in the Gaṇḍavyūhasūtra; in his search for enlightenment, Sudhana meets fifty-three spiritual fiends (kalyānamitra) starting from Mañjuśrī and ending with Samantabhadra.110 At the same time, these three are part of the final stage of Sudhana’s pilgrimage, which culminates in the pronouncement of the Bhadracarī. Before his final meeting with 109. On representations of Sudhana in Chinese art, see Fontein 1968: 23–110. 110. The representations of Mañjuśrī, Samantabhadra and Sūdhana in Yulin Cave 25 most likely have the same meaning, see fn. 101.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) Samantabhadra, the boy sees Mañjuśrī again and receives from him various kinds of teachings.111 In our composition, Mañjuśrī is shown offering Sudhana a round object. A depiction of this meeting can be also found in a wall painting in Mogao Cave 85 (late 9th century), where it is identified by an accompanying cartouche.112 There Mañjuśrī is also depicted offering Sudhana a round object. This object probably represents the jewel of the teaching, references to which can be found in the sūtra. For example, during their first meeting, Sudhana supplicates to Mañjuśrī: ‘Instruct me, O King, with the jewel of the wheel of teaching’.113 Properly prepared by Mañjuśrī, Sudhana sees Samantabhadra, who is depicted in the shrine with his right hand placed upon Sudhana’s head.114 The crucial moment of their meeting, which takes place at the feet of Vairocana/Locana (exactly as shown in the shrine) and culminates in the pronouncement of the Bhadracarī, is described in the Gaṇḍavyūhasūtra as follows: Just as in this world, in the presence of Vairocana Buddha, the enlightening being Universally Good (bodhisattva Samantabhadra) extended his right hand and laid it on Sudhana’s head … And just as facets of truth entered Sudhana as he was touched by the hand of Universally Good at the feet of Vairocana Buddha, likewise facets of truth entered Sudhana in various ways as he was touched by the multitudes of hands extended from the bodies of all the Universally Goods.115

I believe that an image of the elephant, Samantabhadra’s vehicle, placed in the centre of the Vairocana’s throne is a reference to the enlightenment and the Bhadracarī. Douglas Osto suggested that 111. Cleary 1993: 1502. 112. Wang 2018: 250, fig. 119. 113. Cleary 1993: 1175. 114. A similar representation of the scene where Samantabhadra lays his hand on Sudhana’s head can be found in the Gaṇḍavyūha narrative panels located on the east wall on the fourth level at Borobudur, panel IVB-82. See https://www.photodharma.net/Indonesia/11-GandavyuhaLevel-4/11-Gandavyuha-Level-4.htm, picture 82; Fontein 2012: 147, fig. 33. 115. Cleary 1993: 1508.

03_CreSouthV1_8P_3Feb22.indd 85

85

Samantabhadra and the Bhadracarī were ‘associated with each other in the sense that Samantabhadra is thought to embody the aspirations contained in the verses.’116 In addition, and as was pointed out to me by Claudine Bautze-Picron, lion(s) and elephant(s) form a pair which is often encountered in the ornamentation of the pedestal of throne in India from an early period. The meaning imparted to the two animals is rich, but, in short, an elephant refers to the enlightenment and a lion to the teaching.117 The Nelson-Atkins shrine carries an unobtrusive and mostly illegible Tibetan inscription on the back. It opens with an initial Tibetan mark (yig mgo) followed by the syllables byang chub. According to Rob Lonrothe, these syllables may have originally stood for the words ‘awakening’ (byang chub), ‘bodhicitta’ (byang chub sems), ‘bodhisattva’ (byang chub sems dpa’), which would match the suggested symbolism of the central scene. According to Yannick Laurent, the syllables may have alternatively stood for a personal Tibetan name (i.e. Jangchub).118 It has to be also noted that there is no way to know when and where the inscription was added and its presence per se does not signify a Tibetan provenance of the shrine. It merely suggests that the shrine was at some point used by Tibetan-speaking people.119

116. Osto 2010: 5. 117. Bautze-Picron, personal communication, 9 August 2019. This symbolism is also evident in the Huayan art where a lion is the vehicle of Mañjuśrī, an embodiment of the teaching, while an elephant is the vehicle of Samantabhadra, an embodiment of the enlightenment. On the symbolism of these two animals in Indian art, see BautzePicron 2008. 118. Linrothe 2014: 31–34; Yannick Laurent, personal communication, August 2019. Laurent further remarked that the foreign provenance of this Buddhist shrine and its now effaced epigraph make it difficult to determine the nature of this Tibetan inscription. In comparison with other non-Tibetan portable images bearing Tibetan inscriptions, the epigraph of the shrine may have acted as a laudatory inscription, a descriptive caption, or even as an expression of ownership, as shown by Laurent during his presentation at the 15th Seminar of the International Association for Tibetan Studies in Paris, July 2019. 119. See fn. 47.

3/2/22 6:19 PM

Yury Khokhlov

86

A Portable Painting from Mogao Cave 17 Another version of the group of VAEGB can be found in a portable painting from Mogao Cave 17, which can be dated to the late 9th century on stylistic grounds (Fig. 3.8).120 The bodhisattvas are partially identified by Tibetan inscriptions and partially by iconographic attributes. As with the Yulin painting (Fig. 3.1), the figures display all defining characteristics of the Southern Indian artistic mode. The layout of the composition is the same as that in the central section of the Nelson-Atkins shrine (Fig. 3.7) and that in the aforementioned stele from Odisha.121 The bottom foreground of the painting contains representations of sand hills and birds which are normally associated with depictions of Amitābha’s Pure Land. As mentioned above, the Bhadracarī contains a reference to Amitābha’s Pure Land that perhaps explains why its features are depicted in the painting.122

Eastern Tibetan Rock Carvings There are three rock carvings in Eastern Tibet that depict a group of VAEGB (Fig. 3.9–3.11).123 They have often been offered as evidence for the Tibetan provenance of the composition in Yulin Cave 25.124 The carvings are concentrated in the areas bordering China and date to the early 9th century. At least two, if not all, of them were commissioned by the 120. For an overview of the scholarship and the latest discussion on the painting, see Wang 2018: 111–113. 121. See fn. 104. 122. This was noted by Michelle Wang as well. See Wang 2018: 258, n. 59. 123. For an overview of the carvings, see Huo 2018. 1) Renda (also known as Ldan ma brag) high relief rock carving, dated 804/816, Brag gyab County, Chamdo Prefecture of TAR (Fig. 9). On this carving see Heller 1994; Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute 2014. 2) Sgar thog high relief rock carving, near the town of Sgar thog in Smar khams county, Chamdo Prefecture of TAR (Fig. 10). On this carving, see Yang, Lu and Zhang 2017. 3) ’Bis khog (also known as Bimda) high relief rock carving, dated 806, Yushu Tibetan Autonomous Prefecture, Qinghai province (Fig. 11). On the most recent research on this carving and for an overview of the previous scholarship, see Huo 2017. 124. See fn. 54.

03_CreSouthV1_8P_3Feb22.indd 86

same person—Ye shes dbyangs, who was identified as a high-profile Tibetan monk listed in a Chan lineage in the later Dunhuang texts, and who, according to an inscription accompanying the Renda group, commissioned many other carvings in the area.125 Altogether this strongly suggests that the carvings were created under cultural influence of the Hexi Corridor.126 I argue that these rock carvings are associated with the Bhadracarī and, like the examples from Hexi, reflect the practice established by Amoghavajra in the Puxian pusa xing yuan zan (T 297). This can be supported by the fact that a complete text of the Bhadracarī is carved about 100 metres away from the group of VAEGB at ’Bis khog.127 The inscription is considered part of the monument and covers a large portion of the rock that gives it an exceptional prominence at the site. In addition, a prayer to Vairocana and the eight great bodhisattvas is carved in twenty-three lines to the right of the group of VAEGB at the same cite.128 This prayer corresponds to the eulogy to the eight great bodhisattvas in Amoghavajra’s text. At Renda, a single religious text accompanies the group of VAEGB. It was initially identified as an excerpt from the Bhadracarī, but this was questioned by Amy Heller.129 At present, the text is understood to be a short prayer, probably a local creation, which focuses on the bodhisattva Samantabhadra and is inspired by the Bhadracarī.130 In addition, the historical inscriptions at Renda and ’Bis khog have clear references to the main qualities of the Bhadracarī. They praise the path to enlightenment and dedicate the monuments to the well-being of the Tibetan Emperor, tsenpo (Tib. btsan po).131 Sam van Schaik and Lewis Doney have shown that the Bhadracarī was one of the most 125. Heller 1994; On Ye shes dbyangs, see also van Schaik 2015: 163–174. 126. Huo Wei also argues that these carvings should be understood in the context of the Sino-Tibetan relations, see Huo 2018. 127. Huo 2017: 6–7. 128. Ibid.: 5. 129. Heller 1994: n. 6. 130. Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute 2014: 14. 131. Heller 1994; Huo 2017: 6.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

87

Fig. 3.8: Vairocana and the eight great bodhisattvas, ink and colour on silk, 9th or 10th century, from Mogao Cave 17, 63.5 × 17.95 cm, the British Museum, 1919,0101,0.50 (with the permission of the British Museum).

03_CreSouthV1_8P_3Feb22.indd 87

3/2/22 6:19 PM

88

Yury Khokhlov

Fig. 3.11: Sketch of Sgar thog high relief rock carving, early 9th century, near the town of Sgar thog, Smar khams county, Chamdo prefecture of TAR, China (sketch by Yang Qingfan, with the permission of the author and Journal of Tibetology, courtesy of Zhang Changhong).

Fig. 3.9: Sketch of Renda (Ldan ma brag) high relief rock carving, dated 804/816, Zhangyab county, Chamdo prefecture of TAR, China (sketch by Shage Wangdui, courtesy of Zhang Changhong).

Fig. 3.10: Sketch of Bis khog (Bimda) high relief rock carving, dated 806, Yulshul, Tibetan autonomous prefecture, Qinghai province, China (sketch by Tang Huisheng, courtesy of Zhang Changhong).

03_CreSouthV1_8P_3Feb22.indd 88

popular Buddhist prayers in Imperial Tibet and that it was understood as having the power to give the tsenpo longevity. Since many Tibetan copies of the prayer were found among the Dunhuang manuscripts, the authors have remarked that its popularity among Tibetans might have been due to Chinese influence.132 Indeed, even the link between the Bhadracarī and the tsenpo finds parallels in Chinese practice, as Amoghavajra’s translation of the Bhadracarī was intended as a eulogy to the Chinese emperor as well.133 Furthermore, out of three extant Chinese translations, the one by Amoghavajra most closely corresponds to the Tibetan version.134 Regarding artistic features, the groups of VAEGB in the carvings were depicted either in the Tibetan mode, with the figures clad in Tibetan-style robes (Figs. 3.10 and 3.11), or in the Indian mode (Fig. 3.9).135 The iconography of the group 132. van Schaik and Doney 2007: 185, n. 19. 133. Dessein 2003: 331–332. 134. Osto 2010: 2. 135. It is difficult to go beyond this general observation, as in their current state ’Bis khog and Renda carvings are freshly plastered and painted, while the Sgar thog carving is significantly damaged. Initially, the reliefs were likely covered with plaster, in which smaller details were made.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774) of VAEGB in the carvings corresponds to the Southern Indian and Hexi examples (Vairocana is shown with the dhyāna-mudrā). Compositionally, the Renda group is organized in the same way as the group in a portable painting from Mogao Cave 17 (Fig. 3.8) and in the stele from Odisha.136 It displays typical elements of Chinese design such as the canopy above the central figure, which is almost identical to those in the Hexi examples (Fig. 3.1 and Khokhlov 2019: fig. 5), and two apsarases flying on Chinese-style clouds. The two other carvings follow the layout of the composition in Yulin Cave 25, though the bodhisattvas are depicted standing, as they are depicted at Ellora. Discussion on the transmission of the cult of VAEGB from Hexi to Tibet will continue in Part 4.

Summary of Section 2.1 The depictions of VAEGB in Hexi and Eastern Tibet most likely reflect the popularity of the Bhadracarī. Thus, the central figure in these depictions represents the Vairocana of the Avataṃsakasūtra, whose iconography was fused with that of the Vairocana of the Aṣṭamaṇḍalakasūtra. The connection between the Bhadracarī and the Aṣṭamaṇḍalaka sūtra was established in Amoghavajra’s work the Puxian pusa xing yuan zan (T 297), which integrated the eight great bodhisattvas of the Aṣṭamaṇḍalakasūtra into the practice of the Bhadracarī. The Southern Indian stylistic elements detected in the Hexi depictions correspond to the context of Amoghavajra’s teachings and, by extension, suggest a Southern Indian origin of the cult of VAEGB in Hexi. The Tibetan occupation of the Hexi Corridor in the mid-8th century enabled religious and artistic transmission from this region to Tibet. In this regard, the discussed eastern Tibetan rock carvings bear material witness of that process.

2.2  maṇḍala-like compositions featuring vairocanābhisambodhi At both Mogao and Yulin, Vairocanābhisambodhi is consistently depicted in the Southern Indian Such details could have been lost over time and/or altered during subsequent renovations. 136. See fn. 107.

03_CreSouthV1_8P_3Feb22.indd 89

89

mode. Figure 3.12 shows a maṇḍala-like composition137 from Mogao Cave 14 dating to the 9th century. Similar compositions can be found in Yulin Cave 20, and Yulin Cave 38 dating from the 10th to 11th centuries.138 The compositions feature Vairocanābhisambodhi surrounded by bodhisattvas, Offering Goddesses, gatekeepers and/or guardians of directions. Each of these maṇḍalas is accompanied by a second maṇḍala in the same style, which are depicted on the same or opposite walls and centred around Vajrasattva in Mogao Cave 14 (Fig. 3.13) and a Buddha in the two other caves. Textual sources behind these twin maṇḍalas have not been identified.139 What is clear, however, is that they display iconography similar to that of basic maṇḍalas of the Yoga tradition that, for instance, can be recognized by the presence of such distinctive deities as four or eight Offering Goddesses, who sometimes can be depicted in male forms as Offering Bodhisattvas (marked 137. Although these compositions look like maṇḍalas, it is not clear if they were created and used in a ritual context that established their function as maṇḍalas. For a definition of maṇḍala, see Orzech, Sørensen, and Payne 2011: 81–82. 138. See Khokhlov 2019: figs. 16, 17; for a review of the scholarship and a discussion on these compositions as well as for additional images, see Wang 2018: 156–168, figs. 28, 29, 72, 73, 74, 80, 81. On the ground of their style, Wang ascribes these depictions to the Tibetan tradition. She designates them as ‘Maṇḍala of eight great bodhisattvas’ and groups them together with the previously discussed depictions of VAEGB. In my view, this can be misleading, as the maṇḍala of the eight great bodhisattvas contains only nine deities and is based on the Aṣṭamaṇḍalakasūtra. Compositions with more deities are based on different texts and can feature eight different bodhisattvas. Moreover, the eight bodhisattvas in the maṇḍala of the Aṣṭamaṇḍalakasūtra are not the same eight bodhisattvas as those found in maṇḍalas of the Yoga tradition, to which the discussed examples most likely belong. On the latter group of bodhisattvas, see Tanaka 2018: 194–195. 139. The central deities in the maṇḍalas in Yulin Cave 20 are identified by accompanying inscriptions, which give the same name ‘the pure dharmakāya Vairocana Buddha’ (qingjing fashen Piluzhena fo 清淨法身毘盧遮那佛), see Wang 2018: 165. The name Piluzhena fo corresponds to the name of Vairocana in the eight-fascicle translation of the Avataṃsakasūtra (T 279: 39a 17-18), see Wang 2018: 109. Similarly to the case of Yulin Cave 25, this likely reflects adaptation of Esoteric teachings for the Huayan context.

3/2/22 6:19 PM

Yury Khokhlov

90

Fig. 3.12: Maṇḍala-like composition centred around Vairocana, wall painting, second half of the 9th century, Guiyijun period (848–1036), Mogao Cave 14, China (with the permission of the Dunhuang Academy); circles indicate Offering Goddesses. Right: close-up of the central figure of Vairocana.

Fig. 3.13: Maṇḍala-like composition centred around Vajrasattva, wall painting, second half of the 9th century, Guiyijun period (848–1036), Mogao Cave 14, Gansu Province, China (with the permission of the Dunhuang Academy). Right: close-up of the central figure of Vajrasattva.

03_CreSouthV1_8P_3Feb22.indd 90

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

91

by circles in the images).140 For example, Cave 14 has the paired maṇḍala-like compositions centred around Vairocanābhisambodhi (Fig. 3.12) and Vajrasattva (Fig. 3.13). Somewhat similar maṇḍalas with the same central deities are described in Amoghavajra’s translation of the Adhyardhaśatikā Prajnāpāramitāsūtra (Ch. 般若理趣經) (T 743). According to Kimiaki Tanaka, who uses the alternative title Prajnāpāramitānayasūtra for the text, this sūtra is the oldest in the Yoga cycle and it greatly influenced the development of the other texts, including the STTS.141 Thus, it is clear that there is a strong affiliation between the Southern Indian artistic mode and the Yoga tradition.

2.3  seated amoghapāśa with four arms There are three portable paintings from Mogao Cave 17, which depict maṇḍalas centred on seated Amoghapāśa with four arms.142 The painting in Fig. 3.14 is traditionally dated to the 8th or 9th century.143 While four-armed Amoghapāśa in the centre is identified by a noose held in the upper left hand, the presence of four Offering Goddesses (marked by circles) points to the context of the Yoga teachings.144 Amoghapāśa is an Esoteric variant of Avalokiteśvara; he can manifest in different forms, including peaceful and wrathful, standing and seated, 140. On the Offering Goddesses as features of the basic maṇḍalas of the Yoga cycle, see Tanaka 2018: 279–288. 141. On the maṇḍalas of this sūtra, see Tanaka 2018: 248–280 and Hunter 2018. The first and second maṇḍalas described in the text are centred around Vairocanābhisambodhi (21 deities) and Vajrasattva (17 deities), respectively. There are versions where the second maṇḍala is centred on a Buddha instead of Vajrasattva (Tanaka 2018: 204), which corresponds to examples in Yulin Caves 20 and 38. In Cave 14, the central deities of the maṇḍalas are swapped, with the Vairocanābhisambodhi maṇḍala containing 17 deities and the Vajrasattva maṇḍala containing 21 deities. 142. The paintings were discussed by Kimiaki Tanaka, who relates two of them to the Yoga tradition; see Tanaka 2000: 39–71. 143. The painting was discussed by Karl Debreczeny in Brauen 1997: 80–81, where its Indianized style was ascribed to Tibetan influence. 144. See fn. 140.

03_CreSouthV1_8P_3Feb22.indd 91

Fig. 3.14: Maṇḍala of Amoghapāśa, mid-8th century, from Mogao Cave 17, 204.5 × 107.5 cm with mount, the Guimet museum, MG26466; circles indicate Offering Goddesses. (Photo © RMN-Grand Palais, MNAAG, Paris/Mathieu Ravaux).

with four, six, eight or more arms, and so on. There are nine Chinese translations of several versions of the Amoghapāśadhāraṇīsūtra dating from the 7th to 10th centuries, including one by Amoghavajra.145 Iconographic descriptions are rarely given in the texts and it is often simply mentioned that Amoghapāśa looks like Śiva.146 This probably explains the great variety of his iconographic forms, 145. Wong 2007a: 152. 146. Reis-Habito 1999: 48.

3/2/22 6:19 PM

92

Yury Khokhlov

Fig. 3.15: Viṣāpaharaṇa Śiva, 9th century, Pallava period, copper alloy, H 62 cm, Government Museum, Chennai, Tamil Nadu, India (photo by the American Institute of Indian Studies, negative no. 648.97).

Fig. 3.16: Close-up of the central figure of Amoghapāśa, detail of Fig. 3.14.

since Śiva has many iconographic forms as well. At the same time, at least three translations mention Amoghapāśa with one head and four arms.147 Amoghapāśa (see close-up in Fig. 3.16) and his retinue in our painting are depicted in the Southern Indian mode. Although no comparable Southern Indian images of Amoghapāśa are known, there are many contemporaneous images of four-armed Śiva seated with one leg pendant. These images can represent different forms of Śiva depending on the presence or absence of accompanying figures, hand gestures, and attributes. A late Pallava sculpture of Viṣāpaharaṇa Śiva in Fig. 3.15 is very close

to Amoghapāśa in the painting. Both figures have three eyes and are adorned with almost identical armbands decorated with spikes. The similarities in their proportions and in the position of the second pair of the arms are striking. I discussed the two other maṇḍalas executed in the Sothern Indian mode elsewhere.148 They likely belong to the 10th century and display some iconographic and stylistic variations in comparison with the earlier example in Fig. 3.14. Nevertheless, the deities in the paintings show distinctive elements of the Southern Indian mode (spiked decorations, swagged necklaces) as well.

147. These are the translations by Maṇicintana (Ch. 寶思 惟; 693), Li Wuchan (Ch.李無諂; 670), and Bodhiruci (Ch. 菩提流支; 707–709). See Reis-Habito 1999: 52.

148. Khokhlov 2019: 87–88, figs. 20–21.

03_CreSouthV1_8P_3Feb22.indd 92

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

93

Fig. 3.17: Kubera with attendants, wall painting, Tibetan period, Yulin Cave 15, China (with the permission of the Dunhuang Academy).

03_CreSouthV1_8P_3Feb22.indd 93

3/2/22 6:19 PM

Yury Khokhlov

94

2.4  yulin cave 15 Yulin Cave 15 also contains a composition executed in the Southern Indian mode (Fig. 3.17).149 It features yakṣas (a class of demigods) known under various names, including Kubera and Vaiśravaṇa.150 In India, Kubera is the guardian of the North in a group of the eight dikpālas (guardians of the eight directions) and the king of various demigods.151 In China, he is mostly known as Vaiśravaṇa (Ch. 多 聞天王), the guardian of the North in a group of the four heavenly kings.152 Vaiśravaṇa is traditionally depicted in an armoured form, which probably has a Central Asian origin.153 In Cave 15, the deity is represented as a yakṣa holding a club and a mongoose that corresponds to the iconography of Kubera. Close parallels can be found amongst sculptures of yakṣas at Ellora, two of which are shown in Fig. 3.18. These and many other similar examples belong to Ellora’s Jain caves.154 Lisa Owen has recently discovered that these caves reflect the patronage of Jains from the regions of Tamil Nadu and Karnataka.155 This observation corresponds to the Pallava-influenced style of the sculptures. Owen has remarked that only a few comparable depictions survived in Tamil Nadu and Karnataka,156 and thus the Ellora sculptures are especially valuable in order to reconstruct artistic traditions of these regions. This provides additional evidence for a probable Tamil Nadu background of the discussed Yulin painting. In terms of iconography, images of Brahmanic, Buddhist and Jain yakṣas represent 149. Sha Wutian argues that the painting is executed in the Tibetan style. See Sha 2013: 117–118. 150. While the etymology of the former is debatable, the latter derives from the name of his father Viśravis. On yakṣa cult and iconography, see Misra 1981; Bautze-Picron 2002. In particular, on Kubera/Vaiśravaṇa, see Misra 1981: 59–71. 151. On the iconography and symbolism of the guardians of the directions (dikpālas) in India, see Wessels-Mevissen 2001. 152. On the iconography and symbolism of the four heavenly kings in Central and East Asia, see Shim 2013. 153. Shim 2013: 82–112. 154. On Ellora’s Jain caves, see Owen 2012. In particular, on depictions of yakṣas at Ellora, see ibid.: 81–127. 155. Ibid.: 16, 20, 160–163. 156. Ibid.: 4–5.

03_CreSouthV1_8P_3Feb22.indd 94

a shared concept of wealth and protection and display only minor differences.157 Like the Yulin Kubera, the Ellora yakṣas are placed under an arched canopy formed by a tree with luxuriant foliage and flowers. Such design with a tree as part of a throne back and luxuriant foliage depicted above a divinity is unusual for the art of the Deccan, but is common in the examples from Tamil Nadu.158 The figures are seated with one leg pendant and flanked by two attendants, one of which (on the right) carries a purse made of mongoose skin. In contrast with the Ellora yakṣas, the Yulin Kubera is seated on a throne.159 Nevertheless, both figures are similarly decorated with swagged necklaces as well as with spiked crowns and armbands. Additionally, the right image in Fig. 3.18 and the Yulin Kubera have almost identical square backrests and halos of an inverted U-shape. The appearance of such Kubera images in Hexi could be related to protective rituals centred on Vaiśravaṇa, rituals which Amoghavajra regularly performed to provide imperial troops with supernatural assistance. The earliest recorded instance of Amoghavajra performing the Vaiśravaṇa rituals is specifically linked to the Tibetan siege of Liangzhou.160 There are five texts in the Taishō Canon attributed to him that specifically invoke Vaiśravaṇa for military magic.161 In these texts, he is consistently referred to as a king of yakṣas. Frustratingly, his physical appearance is not always described and where it is given, the description is of his armoured

157. Ibid.: 127–129; Bautze-Picron 2002. 158. I am grateful to Claudine Bautze-Picron for drawing my attention to this feature. Comparable examples can be found in the Jain rock carvings at Kalugumalai, Tamil Nadu, in the 9th or early 10th century, see Owen 2010: figs. 3, 11. Also, see a 9th-century Buddha from Tamil Nadu, now in the Museum of Fine Arts, Boston (accession numbers 1970.3 and 67.4a-b); and a 10th-century Buddha with two bodhisattvas, from Nāgapaṭṭinam, Tamil Nadu, now in the Government Museum Chennai (accession number 81212). 159. Kubera seated on a throne is depicted in the early 8th-century Ellora Cave 25, but the image is significantly damaged. See Owen 2012: fig. 57. 160. Yang 2018: 86–87. 161. These rituals and texts have been discussed in Goble 2013.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

95

Fig. 3.18: Yakṣas with attendants, early 9th century, Ellora Cave 32, Maharashtra, India (photograph by the American Institute of Indian Studies, negative no. 442.45, 441.9).

form.162 However, the Yulin Kubera is shown with the same two attendants as those that are normally depicted with Vaiśravaṇa in Dunhuang art.163 This suggests that, despite the differences in iconography, he was understood to be the same god as Vaiśravaṇa.164 It can be assumed that the aforemen162. Geoffrey Goble, personal communication, 23 July 2018. 163. See, for example, a woodblock print dated to 947 in the British Museum collection (1919,0101,0.245). The attendants are a female deity with a tray of jewels (identified by the museum as Goddess Śrī) and a gandharva (a class of demigods) wearing a tiger-head helmet that covers his head and shoulders. On depictions of gandharvas in Chinese art, see Wong 2018: 73, fig. 2.14, 120, fig. 3.8. Another example of Vaiśravaṇa attended by a gandharva is a 9th-century painting in the British Museum collection (1919,0101,0.38). 164. The conflation between the dikpālas and the heavenly kings is also evident in the compositions in Figs. 3.12 and 3.13. While the North, South, and West directions are guarded

03_CreSouthV1_8P_3Feb22.indd 95

tioned protective rituals were of an Indian origin and therefore they must have been initially centred on Kubera, whose images likely accompanied the original Indian texts, but whose name was replaced with its popular local equivalent, Vaiśravaṇa, in the Chinese translations.

2.5  portable paintings with solitary figures of bodhisattvas Fig. 3.19 shows a solitary standing bodhisattva painted in the Southern Indian mode.165 Its vertical format implies that the painting was displayed by the dikpālas (Kubera is depicted in the right top corner in Fig. 3.12 and in the left top corner in Fig. 3.13; Yama and Varuṇa are in the bottom corners), the East is guarded by the heavenly king of the East, Dhṛtarāṣṭra, who is clearly recognized by his attribute, a lute. 165. Another example of such is in Khokhlov 2019: fig. 26.

3/2/22 6:19 PM

96

Yury Khokhlov on a banner and was possibly produced as part of a set representing a group of the eight great bodhisattvas (the images of this type are listed in Appendix 3.1 under numbers 10–13). Such banners were probably used during Esoteric rituals.166 For example, they could have been hung to form a circle as it is prescribed for images of the eight great bodhisattvas in the already mentioned Amoghavajra’s ritual manual for recitation of the Uṣṇīṣavijayādhāraṇīsūtra (T 972).167 Parallels to these images of standing bodhisattvas can be found at Ellora as well. As mentioned, caves 11 and 12 contain several groups of the eight great bodhisattvas depicted standing.168 An outstanding painting in Fig. 3.20, however, could be a direct copy from a Pallava or Sri Lankan original as it literally radiates the Pallava aesthetics and, in contrast with the image in Fig. 3.19, shows almost no Chinese constitute. The bodhisattva sports a tubular and extremely tall hairstyle (jaṭāmukuṭa), a typical feature of Pallava imagery and Sri Lankan imagery in the Pallava style. A figure attributed as Tārā from the British Museum in Fig. 3.21 is one striking example of this fashion. The sculpture was discovered in Sri Lanka, but it might have been exported to Sri Lanka from the Pallava domain.169 The bodhisattva and Tārā are clad in lower garments of a similar design and their headdresses are secured by nearly identical and leaf-like crown elements. Unlike the other discussed images, the crown and armbands of the bodhisattva do not have spiked elements. It has to be noted that various decorations without spikes were also common in Pallava art.

2.6  examples of mixed imagery As discussed in Part 2.1, the bodhisattvas in Yulin Cave 25 (Fig. 3.1) were likely painted in accordance with standardized models known in Chinese art as yang. Several portable paintings discovered in Mogao Cave 17 suggest that yang models of boFig. 3.19: Bodhisattva Samantabhadra, ink and colour on silk, from Mogao Cave 17, 9th or 10th century, 59.8 × 17.8 cm, the Guimet Museum, MG17770. (Photo © MNAAG, Paris, Dist. RMN-Grand Palais/image musée Guimet)

03_CreSouthV1_8P_3Feb22.indd 96

166. On the use of banners during Buddhist rituals, see Wang 2018: 46. 167. See fn. 92 and Wang 2018: 47. 168. Malandra 1993: figs. 222, 223, 251, 252. 169. Srinivasan 1999: 110.

3/2/22 6:19 PM

In the Footsteps of Amoghavajra (705–774)

97

Fig. 3.20: Bodhisattva, ink and colour on silk, 8th century, from Mogao Cave 17, 57.5 × 38.1 cm, the British Museum, 1919,0101,0.7 (with the permission of the British Museum).

03_CreSouthV1_8P_3Feb22.indd 97

3/2/22 6:20 PM

Yury Khokhlov

98

Fig. 3.21: Tārā?, copper alloy, 8th century, Sri Lanka, H 143 cm, the British Museum, 1830,0612.4 (with the permission of the British Museum).

dhisattvas in the Southern Indian mode were possibly also used for depicting attending bodhisattvas in heavenly assemblies where the rest of the figures are in the Chinese mode. Perhaps the earliest example of such imagery is a portable painting shown in Fig. 3.22, which belongs to the Tibetan period and is dated to 836. The painting contains Chinese and Tibetan inscriptions and displays a mixture of the Chinese and Indian artistic modes. It has been suggested that the Indianized images in the painting represent the Tibetan style

03_CreSouthV1_8P_3Feb22.indd 98

and their appearance along with Chinese images is a sign of Tibetan influence, which allegedly materialized in Dunhuang art during the period of the Tibetan occupation.170 I would argue that neither in terms of iconography nor in terms of artistic mode can this painting be ascribed to Tibetan influence. Although the painting is titled by the British museum as ‘Paradise of Bhaiṣajyaguru’, it was likely centred on Cintāmaṇicakra, Thousand-armed Avalokiteśvara, and Amoghapāśa, which was a popular grouping in Dunhuang. Their images are largely lost, but the deities are mentioned in the accompanying Chinese inscription. Cults of these Esoteric forms of Avalokiteśvara gained popularity during the reign of Empress Wu (Ch. 武則天, r. 684–705) and were promoted by the court along with the Huayan doctrine.171 As a result, images of these deities are often found along with those belonging to the Huayan tradition.172 In this painting too, characteristic figures of two main bodhisattvas of Huayan Buddhism, Samantabhadra riding an elephant and Mañjuśrī riding a lion, are depicted right above the Esoteric versions of Avalokiteśvara. Finally, an assembly of Buddha Bhaiṣajyaguru occupies the top of the painting. All these subjects were popular in Dunhuang before the Tibetan period and altogether they can be found at already discussed Mogao Cave 148, which was excavated between 766 and 776 (i.e. in the decade before the Tibetan conquest).173 As with Mogao Cave 148 and Yulin Cave 25, the painting shows the coexistence of the two artistic modes. The bodhisattvas flanking the figure of Bhaiṣajyaguru are shown in the Indian mode, while the rest of the figures are depicted in the Chinese mode. These bodhisattvas represent the same type as that in the VAEGB composition 170. Karmay 1975: 8–14; Wang 2018: 98–107. 171. On cults of Esoteric deities during the Tang, see Wong 2018: 184–200. 172. On Esoteric imagery in Huayan context, see Wong 2012: 223–260. 173. The iconographic programme of Mogao Cave 148 includes sculptural representations of Cintāmaṇicakra and Amoghapāśa (no longer extant) and a painted image of Thousand-armed Avalokiteśvara along with distinctive Huayan images of Samantabhadra riding an elephant and Mañjuśrī riding a lion as well as several heavenly assemblies including that of Buddha Bhaiṣajyaguru. See Lee 2010: 91.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

99

Fig. 3.22: Paradise of Bhaiṣajyaguru, ink and colour on silk, from Mogao Cave 17, dated 836, Tibetan period, 152.3 × 177.8 cm, the British Museum, 1919,0101,0.32 (with the permission of the British Museum). Right: close-up of the bodhisattva.

of Yulin Cave 25 (Fig. 3.1) and display the characteristics of Southern Indian art such as spiked decorations and swagged necklaces (see a close-up of the bodhisattva in Fig. 3.22). A large rectangular cartouche in the centre of the painting contains Chinese and Tibetan inscriptions. The Chinese inscription is almost completely faded and illegible to the naked eye, while the Tibetan inscription is relatively clear.174 The Chinese inscription was read with the help of infrared photography and it includes the date. Although the inscriptions are similar in their content, it is clear that they are not translations of each other. The Tibetan one mentions an image of Pariṇatacakra instead of Amoghapāśa in the Chinese inscription, and furthermore dedicates the painting to all living beings, whereas the Chinese inscription dedicates it to the donor’s deceased parents. Bearing in mind that the Chinese inscription is faded, and the majority of the painting is lost, it could be that the painting was restored at some point and then rededicated with the Tibetan inscription. It is also worth remem174. For the photograph of the inscriptions, see Wang 2018: 102, fig. 44.

03_CreSouthV1_8P_3Feb22.indd 99

bering that during the Tibetan period the Tibetan language was the language of the ruler and was used in Hexi in parallel with the Chinese and not only by Tibetan people.175 In this regard, a Tibetan inscription per se cannot be regarded as evidence of the Tibetan religious and/or artistic influence. Another example of such synthesis is a 10th century painting from the Musée Guimet (Fig. 3.23) that depicts the ten stages of bodhisattvahood according to the Daśabhūmikasūtra, which is itself a part of the Avataṃsakasūtra. The bodhisattvas in the Southern Indian mode are included in one of the assemblies whereas the other figures are depicted in the Chinese mode.176 The examples discussed in this section likely reflect the popularity of Indian icons in China as well as the creativity of Dunhuang artists.177 175. See fn. 47. 176. The iconography of the painting is discussed in Wong 2007b: 343. According to Wong, the painting is the only known illustration of this subject. 177. In addition, in a British Museum painting from Mogao Cave 17 (Ch.XXVI. a. 007), which represents the Huayan Mañjuśrī riding a lion, the bodhisattva is depicted

3/2/22 6:20 PM

Yury Khokhlov

100

Fig 3.23: Ten stages of the bodhisattvahood according to the Avataṃsakasūtra, ink and colour on silk, from Mogao Cave 17, early 10th century, 286 × 189 cm, the Guimet museum, MG26465. (Photo © RMN-Grand Palais, MNAAG, Paris/Richard Lambert). Right: close-up of one of the assemblies.

conclusion for part 2 The images from Mogao and Yulin discussed above share a great deal of iconographic and stylistic similarities with Pallava and Pallava-influenced examples from Southern India. Most of them can be related to the Yoga and other teachings propagated in China by Vajrabodhi and Amoghavajra. Their stylistic consistency, which is already evident in the earliest examples (late 8th century), suggests that this artistic tradition emerged earlier and thus it must have been based on the sketch books brought to China by Vajrabodhi, who was active in the Tang capitals from 720 to 740, and who was a prominent painter himself. His own works must have proin the Southern Indian mode. Another example from the British Museum collection, where attending bodhisattvas are shown in the Southern Indian mode, is a painting Ch.ivi.0034.

03_CreSouthV1_8P_3Feb22.indd 100

moted the Pallava aesthetics as well. However, the popularity of such imagery in Dunhuang art in the second half of the 8th century is in all likelihood related to Amoghavajra’s activities in the Hexi Corridor. Since the dates of the discussed imagery span from the 8th to the 11th century, it can be surmised that Amoghavajra’s teachings were adopted by the Tibetans in Hexi and remained popular in the region even after the collapse of both the Tibetan and Tang empires. Nevertheless, they were likely suppressed by a new wave of Esoteric transmission, which started arriving in Hexi from northeastern India in the late 10th century.178 These new teach178. In the late 10th century, a large number of Chinese monks were sent to northeastern India through the Hexi Corridor in search of new Esoteric teachings. At the same time, many Indian monks arrived in China attracted by the lavish Buddhist patronage of the early Song emperors. It can be assumed that travelling through the Hexi

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) ings brought along new aesthetics characterized by the Pāla style.

part 3.  a brief overview of the southern indian influence on buddhist art in china This part aims to demonstrate that the existence of a Southern Indian mode in Hexi art should not be perceived as disturbing or exceptional. There are other examples in Tang Buddhist art that complement the Hexi material and reflect the centuries-long interactions between the polities of Southern India, Sri Lanka, Southeast Asia, and China. Historical records from the 3rd century CE onwards tell us about a great number of Buddhist monks travelling between India and China by sea.179 Sri Lanka, whose Buddhist culture was intimately linked to that of Southern India, was an important transit stop for many of them. Early examples of these travelling monks include Faxian (Ch. 法顯, 337/342–422), who spent two years in Sri Lanka on his way from Northern India to China; and Bodhidharma (Ch. 菩提達摩), the founder of Chinese Chan Buddhism, who travelled from the Pallava Kingdom to China in the late 5th or early 6th century. Indianized polities of Southeast Asia played an important intermediate role in Buddhist and artistic transmission from Southern India to China. The Indianization of the region started in the early historical period and was facilitated by Tamil seafaring merchants. The process included adoption of Southern Indian religious cults, models of statehood, languages, scripts, and artistic styles. Early Buddhist and Brahmanic icons found in Southeast Asia display a dominating Southern Indian

Corridor, these monks shared teachings and imagery with local Buddhist communities. On these movements, see van Schaik and Galambos 2012: 35–61; Sen 2016: 102, 110–125. The earliest examples of Pāla style imagery in Hexi are found in the early 11th-century Mogao Cave 76, see Toyka-Fuong 1998. On Pāla-style imagery in Hexi art, see Khokhlov 2018. 179. On Buddhist monks travelling between China and India by sea, see Acri 2018.

03_CreSouthV1_8P_3Feb22.indd 101

101

influence.180 According to Chinese sources, the Southeast Asian polities of Funan (Mekong delta, territory of Cambodia), Lynyi (territory of Central Vietnam) and Panpan (Malay peninsula, territory of Thailand) often sent Buddhist monks, texts, and images to the dynasties of Southern China.181 Although Buddhist art of the Southern Chinese dynasties before the Tang is almost completely lost, in his pioneering research on textual sources Alexander Soper argued that in the late 5th and 6th centuries it was greatly influenced by the Southern Indian artistic tradition transmitted by maritime routes as well as through Indianized polities of Southeast Asia.182 Soper states that artists from Southern China were famed as great specialists in rendering ‘Indian Buddhist icons’ and their legacy can be traced in Buddhist art from the Northern Qi (550–557) to the Northern Song dynasties (960– 1127).183 Soper supported his argument by pointing to Southern Indian features found in some Northern Qi examples.184 His theory can be further confirmed thanks to sculptures recently excavated in the eastern coastal area of Shandong province.185 As was discussed in Part 1, interactions between Southern India and China increased dramatically during the Tang dynasty. However, the resulting artistic current in Tang art has rarely been highlighted in art historical literature. The problem partially lies in the lack of surviving material from Chang’an and Luoyang where foreign influences on art were especially strong. The capitals were prime destinations and places of residence for the majority of foreign monks and this was also true for Vajrabodhi and Amoghavajra. These metropolises suffered badly from the calamities of Chinese history and none of their hundreds of Buddhist temples have survived.186 Nevertheless, in addition to the discussed Dunhuang images, the Southern Indian facet of Tang Buddhist art can be further illustrated by a few more examples. 180. Guy 2014: 3–21. 181. Wade 2014: 25–31. 182. Soper 1960. 183. Ibid.: 88–89. 184. Ibid.: 95–96, fig. 20. 185. For a discussion on this group of sculptures, see Okada 2004. See also Khokhlov 2019: 98–99, figs. 31–32. 186. Wong and Field 2008: 132–135.

3/2/22 6:20 PM

Yury Khokhlov

102

3.1  longmen buddha sculptures About a hundred of sculptures depicting the Buddha seated in bhadrāsana are distributed in several caves at Longmen (Ch. 龍門) and Gongxian (Ch. 鞏縣) (Fig. 3.24). Each one of them is about one metre high and is dated by inscriptions to the period between 655 and 680. These inscriptions name the sculptures as ‘Buddha of the King of Udayana’.187 Amy McNair argues for a Northern Indian stylistic origin of these images and compares them with a Buddha figure from Sarnath. This attribution, however, ignores clear features of Southern Indian art present in the sculptures such as low uṣṇīṣa, bare right shoulder, and ankle-sweeping robe, a vestige of the Amaravati tradition. The closest comparisons are metal sculptures excavated in Andhra Pradesh and Tamil Nadu, which are datable to circa the 6th century (Fig. 3.25).188 Several sculptures with similar stylistic characteristics are also known from Southeast Asian polities, thus suggesting the possibility of transmission through Southeast Asia as well.189

Fig. 3.25: Buddha, copper alloy, 5th–6th century, from Nāgapaṭṭinam, Tamil Nadu, India, H 40.6 cm, the Metropolitan Museum, 1998.414 (Public Domain).

Fig. 3.24: ‘King Udayana Buddha’ images, stone, ca. 655–680, Tang dynasty, H ca. 100 cm, Longmen, Henan province, China (photograph by and with the permission of N. Revire). 187. On these sculptures, see McNair 2007: 102–103. 188. In addition to the image in Fig. 3.25, see Buddha figures from the Buddhapad hoard, now in the British Museum collection (1905.1218.2; 1905.1218.3): https://www.britishmuseum.org/collection/object/A_1905-1218-2, https://www. britishmuseum.org/collection/object/A_1905-1218-3. 189. See Guy 2014: cat. 20–23, 50, 51.

03_CreSouthV1_8P_3Feb22.indd 102

3.2  two metal sculptures The Metropolitan Museum of Art has a metal figure of a bodhisattva (Fig. 3.27), which was reportedly found in Vietnam and dated by the museum to the 7th century. Another similar figure, which was attributed to the Sui Dynasty (581–618) of China, has recently been published (Fig. 3.26).190 These figures show a synthesis of Chinese and Southern Indian aesthetics, similar to what is found 190. von de Schulenburg et al. 2015: 150–151.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

Fig. 3.26: Bodhisattva, copper alloy, 8th century, Tang dynasty, China, 43 cm, private collection (courtesy of the owner and Dr Stephan von der Schulenburg).

Fig. 3.27: Bodhisattva Avalokitesvara, copper alloy, 8th century, Tang dynasty, China, 36.5 cm, the Metropolitan Museum, accession no. 1993.387.14 (Public Domain).

in the painting from Mogao Cave 17 in Fig. 3.19. Both display distinctive swagged necklaces, while the crown of the bodhisattva in Fig. 3.26 is decorated with spikes. The date of their production is unlikely to be earlier than the 8th century, as the figures show closer connections with late Pallava material, such as the sculpture in Fig. 3.28. The fact that the territory of Northern Vietnam (Đại Việt, Red River delta) belonged to Tang China suggests a probable southern Chinese origin for these sculptures. The occurrence of such imagery in southern China is not surprising, as Guangzhou

03_CreSouthV1_8P_3Feb22.indd 103

103

Fig. 3.28: Bodhisattva Maitreya, copper alloy, 8th or 9th century, Pallava period, 39.5 cm, from Tanjor, Tamil Nadu, formerly in the Government Museum of Chennai, stolen, present whereabouts unknown (with the permission of Serindia Publications, courtesy of Shane Suvikapakornkul).

was a major entry point to China for seafaring merchants and monks arriving from South Asia. As was previously mentioned, three Tamil temples existed in Guangzhou in the 8th century that clearly demonstrates significant religious transmission.191 Moreover, Amoghavajra himself was popular in the area as well. The master preached in Guangzhou right before his trip to South Asia. He attracted a great number of local followers, converted thou191. See fn. 24.

3/2/22 6:20 PM

104

Yury Khokhlov

sands of people to Buddhism, and conferred Esoteric initiations on members of the ruling elite.192 In addition, upon his return to China, Amoghavajra was temporarily banished from the court and spent four years (749–753) in the Guangzhou area.193 It is worth noting that the sculptures under discussion are visually close to a famous icon of the Nanzhao and Dali kingdoms in Yunnan, the Acuoye Guanyin, which can also ultimately be aligned with a Pallava model.194 Taken together, these examples confirm considerable Southern Indian influence in the region.

3.3  nara embroidery and hōryūji murals The embroidery in Fig. 3.29 belongs to the Nara National Museum in Japan. It has traditionally been dated to the early 8th century and is believed to have been produced in China, likely in one of the Tang capitals.195 The work depicts a heavenly assembly with the Buddha attended by a retinue of bodhisattvas, monks, and lay people. I regard images of the Buddha with two leading bodhisattvas as heavily influenced by Southern Indian and Sri Lankan prototypes. The Buddha exhibits the same iconography as that of the Longmen sculptures (Fig. 3.24). At the same time, his bulky proportions and a robe drape in the Amarāvatī style are strikingly similar to that in a circa 7th century Pallava sculpture in Fig. 3.30.196 The folds of the robes in both figures are arranged in a similar and distinctive fashion. They are concentrated along the upper edge of the robe and below the waist, creating a flat area on the stomach. The two chief bodhisattvas with their wide shoulders and almost naked torsos stand out visually from the group (see a close-up in Fig. 3.31). They represent a particular type of bodhisattva, which was popular in Pallava and Sri Lankan art 192. Yang 2018: 33; Chou 1945: 288–289. 193. Yang 2018: 41. 194. On Southeast Asian stylistic connections of this icon, see Guy 1995: 86. 195. Wong 2018: 153–154. 196. On this sculpture, see Rao 1915; Aiyappan and Srinivasan 2000: 70–72.

03_CreSouthV1_8P_3Feb22.indd 104

(see sculptures in Fig. 3.31).197 Bodhisattvas of this type are characterized by an upright posture, and by the abhaya- or vitarka-mudrā performed by right hand and kaṭaka-mudrā performed by left hand. The mudrās are held distinctively at elbow-level. These figures can be bejewelled or unadorned, but always wear a distinctive flat and wide sash, which probably represents the skin of a black deer (Skt. kṛṣṇājina) worn in the manner of the sacred thread (Skt. yajñopavīta).198 It is loosely wrapped across the torso, forming a loop on the left shoulder and drapes down to the right hip. The bodhisattvas in the embroidery follow this model closely. As has been recently shown by Osmund Bopearachchi, this model likely represents Bodhisattva Avalokiteśvara as a protector of mariners.199 Such images were often found along river and sea trade routes that points to a maritime way of transmission of this iconographic form to China. It has long been noted that these two bodhisattvas in the embroidery resemble the bodhisattvas in the murals that are used to adorn the main hall of Hōryūji Temple in Japan (Fig. 3.32).200 These murals depicted heavenly assemblies of four buddhas as well as eight bodhisattvas, whose identities are not entirely certain. It is believed that the murals date to the early 8th century and represent the metropolitan Tang style that has largely been lost in China. Moreover, the figures in the murals were most likely executed with the help of full-sized stencils, which are believed to have been brought to Japan from one of the Tang capitals.201 Fig. 3.33 compares the bodhisattva images from Yulin Cave 25, the embroidery, and the Hōryūji murals. Other than certain variations in decorations, the figures look strikingly similar in their 197. A number of metal sculptures representing bodhisattvas of this type were recovered in Indonesia and Thailand and it is assumed that they were imported from the Pallava Kingdom or Sri Lanka. See Guy 2014: cat. 6; a Pallava example is in cat. 7. 198. It was attributed as such by Dohanian. See Dohanian 1977: 41. 199. Bopearachchi 2014: 161–187. 200. On the Hōryūji and its murals, see Wong and Field 2008. The most recent research on the dating of the Hōryūji murals can be found in Wong 2008. 201. Wong and Field 2008: 147–152.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

105

Fig. 3.29: Shaka-nyorai (Śākya-muni) preaching at Ryōju-sen (Gṛdhra-kūṭa) mountain, silk embroidery, probably Chinese work, 8th century, Tang Dynasty, China, 207 × 157 cm, Nara National Museum, 647-0, Japan (the photograph is provided by Nara National Museum).

03_CreSouthV1_8P_3Feb22.indd 105

3/2/22 6:20 PM

Yury Khokhlov

106

Fig. 3.30: Buddha, stone, 7th century, Pallava period, 240 cm, excavated on the territory of Kamakshidevi Temple, Kancipuram, Tamil Nadu, Government Museum Chennai (photograph by and courtesy of Yuko Fukuroi). Right: Buddha, detail of Fig. 3.29.

Fig. 3.31: (left) Bodhisattva, copper alloy, 8th–9th century, H 17.9 cm, Sri Lanka, Los Angeles County Museum of Art (Public Domain). Centre: Bodhisattva, detail of Fig. 3.29. Right: Bodhisattva, copper alloy, 8th century, Sri Lanka, Colombo National Museum (photograph by and with the permission of Osmund Bopearachchi).

03_CreSouthV1_8P_3Feb22.indd 106

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

107

Fig. 3.32: Buddha Amida Pure Land, wall painting, panel 6, west wall, Kondō, Hōryūji, 8th century, Ikaruga, Nara Prefecture, Japan, not extant anymore (courtesy of Dorothy Wong).

Fig. 3.33: (left) Bodhisattva, detail of Fig. 3.32. Centre: Bodhisattva, detail of Fig. 3.1. Right: Bodhisattva, detail of Fig. 3.29.

03_CreSouthV1_8P_3Feb22.indd 107

3/2/22 6:20 PM

108

Yury Khokhlov

linear structures, proportions, postures, gestures, and facial features. What unite them most is their elongated and almost naked torsos, remarkably broad shoulders, and contrasting narrow waists. Though located thousands of miles apart, the Hōryūji and Yulin bodhisattvas, with their almost identical swagged necklaces of the Southern Indian type, and similarly drawn sash and kṛṣṇājina, look as if they are from the same painting. Such similitude suggests that the images were based on standardized models. As mentioned, many of these yang, standardized models and templates for the replication of images, were recorded in official inventories of objects transferred from Tang China to Japan.202 The Hōryūji example also shows that artistic models were transmitted from China to Japan even with the help of full-sized stencils. As was also already mentioned, significant artistic interexchange existed between Dunhuang and the Tang capitals as well.203 In my view, taken as a group, the imagery from Yulin Cave 25, the Nara embroidery, and the Hōryūji murals confirm the existence of a distinctive tradition in Tang Buddhist art, which was most likely associated with Vajrabodhi as a painter.204 It should be mentioned that the South202. See fn. 63. 203. See fn. 67. 204. Remarkable similarities between the images of bodhisattvas from Yulin Cave 25, the Nara embroidery, and the Hōryūji murals would also seem to suggest that they were created in a relatively close temporal proximity. Out of the three bodhisattvas in Fig. 3.33, only the Yulin image is securely datable. It was created between 776 and 851, but most likely in the late 8th century. In contrast, the Hōryūji murals and the Nara embroidery have been variously dated by scholars on stylistic grounds from the late 7th to the late 8th century (Wong 2008: 143). In her recent research, Wong compares the Hōryūji figures with bodhisattvas from Dunhuang caves of the early Tang dynasty (Wong 2008: 158–168). Although some of her comparative examples (especially from Cave 329) are, indeed, very close in clothing and decorations, their body proportions are strikingly different. The Dunhuang figures are slim and elongated, which is in sharp contrast to mighty torsos and wide shoulders of the Hōryūji and Nara bodhisattvas. In my view, such fullness of figures points to a later date. In fact, the painting from Yulin Cave 25 (Fig. 3.1) is the closest comparison from Dunhuang. In the light of the presented analysis, the dating of the Hōryūji murals and the Nara embroidery should perhaps be reconsidered.

03_CreSouthV1_8P_3Feb22.indd 108

ern Indian influence on the Hōryūji murals has previously been suggested by at least two other scholars. Jeannine Auboyer noted similarities in clothing between the Hōryūji figures and Ellora statues, while Marilyn Rhie argued that the facial features of the bodhisattvas reflect the influence of Southern Indian art, an influence she saw as resulting from the maritime exchange of Tang China with Southern India.205

part 4.  from the hexi corridor to central tibet Tibetan rule in the Hexi Corridor lasted roughly from the mid-8th till the mid-9th century and coincided with a period of the most intense promotion of Buddhism in the Tibetan Empire. As discussed, the region was one of the major centres for the teaching and practice associated with the Southern Indian Yoga tradition, which became dominant in the official Buddhist ceremony and doctrine of both China and Tibet in the latter half of the 8th century. The Tibetan religious historiography ascribes the introduction of the Yoga teachings in Tibet to Northern Indian master Buddhaguhya, who was supposedly active in the 8th century. However, his identity and life story cannot be historically verified, and some scholars cast doubt on his influence in Tibet.206 On the other hand, the Testament of Ba (Tib. Dba’ bzhed/Sba bzhed), a historical account on the origin of Buddhism in Tibet, which was probably compiled in the 11th or 12th centuries, attests to significant activities of Chinese monks from Hexi in Central Tibet in the second half of the 8th century. It says, for example, that the Buddhist master Moheyan from Dunhuang was patronized by the Tibetan court and even had a majority of followers in Tibet in the late 8th century.207 According to the Testament of Ba, tensions developed between the Chinese and newly arrived Indian teachers, who allegedly taught a different path towards enlightenment. The Testament of Ba further states that these tensions were resolved in a formal debate, where 205. Auboyer 1941: 71; Rhie 1988: 27–28. 206. Weinberger 2003: 30, 83–90; Nagasawa 2017. 207. Wangdu and Diemberger 2000: 76, n. 287; van Schaik 2015: 14–17.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

109

Altogether this would seem to suggest that Budthe Chinese side was defeated and expelled from Tibet. As discussed by van Schaik, this episode has dhist masters from Hexi, who have been perhaps no documentary value and was likely designed to misleadingly labelled as Chan monks in modern validate new Indian lineages, which started arriving scholarship,214 were instrumental in the transmisin Tibet in the 11th century, as the only genuine sion of the Yoga tradition and likely many other source of Buddhism.208 It has been traditionally ‘not Chan’ teachings to the Tibetans in the 8th and understood that the Chinese party represented the 9th centuries. Chan tradition, which emphasizes a concept-free This can be also supported by the discussed meditation as opposed to scholastic Indian tradi- example of Ye shes dbyangs, the Tibetan monk, tion advocating a gradual path and a multitude who has been identified as a member of the Chan of methods including Esoteric teachings. In the lineage in the later Dunhuang texts, and who proplight of recent scholarship, however, this picture agated the cult of VAEGB as part of Amoghavajra’s looks not only simplified, but also anachronistic. practice of the Bhadracarī. As argued by Sharf, the very notion that Chan and According to Weinberger and Dalton, it is Esoteric teachings constituted independent tradi- nearly certain that the STTS, the root Tantra of tions in China in the 8th century can be mislead- the Yoga tradition, was transmitted to Tibet and ing. Moreover, these teachings had a few areas of probably even translated into Tibetan during the congruence and were likely practised in the same Imperial period, though no details of the translaBuddhist circles.209 With regard to Hexi, van Schaik tion are known and the text is missing from the has demonstrated that in the 8th and 9th centu- Imperial catalogues.215 In contrast, the Chinese ries monks in Dunhuang, who were identified as translations are much better documented and members of Chan lineages in the later texts, prac- historically attested, and it is not impossible that tised various Esoteric rituals, including those asso- the early Tibetan renditions could have been based ciated with the Yoga tradition in general and the on the Chinese texts. The earliest work in Chinese Vajradhātu maṇḍala in particular.210 Furthermore, summarizing the principles of the STTS dates to several Chinese and Tibetan sources suggest that 723 and is authored by Vajrabodhi. It was followed Moheyan himself was proficient in Esoteric rituals by Amoghavajra’s translation of the first part of the as well.211 These traits find striking parallels in the STTS produced in Liangzhou around 755.216 teachings of the Nyigmapa (Tib. rNying ma pa), the As a matter of fact, the Chinese legacy in oldest school of Tibetan Buddhism, which traces Tibetan Buddhism was deliberately minimized its origin back to the Imperial period. According to by later Tibetan historiographies and by a pracSnellgrove, there was a considerable vogue of Yoga- tice of editing the colophons and texts of Tibetan tantras in early Tibet and many Mahāyogatantras translations from non-Sanskrit sources in order to of the Nyigmapa, which were largely rejected by connect them directly to the Indian tradition. This the later Tibetan schools ‘as unorthodox and prob- practice started in as early as the 9th century.217 As ably Chinese’, conform to the content and style James Gentry put it, of ‘orthodox Yogatantras’, such as the STTS.212 In the general unreliability of translation coloaddition, the Nyigmapa tradition of Dzogchen (Tib. phons and centuries of scribal and editorial acrDzogs chen, the Great Perfection), which includes tivity, which often included efforts during the elements of the Mahāyoga and Chan teachings, was Imperial period and thereafter to standardize viewed by some 13th and 14th century authoritaTibetan scriptural translations according to tive Tibetan scholars as the Chinese doctrine of an imperially-decreed common lexicon, have Moheyan.213 208. 209. 210. 211. 212. 213.

van Schaik 2015: 15. Sharf 2017: 53. van Schaik 2015: 168. Ibid.: 132–133. Snellgrove 1987: 457–460. Karmay 2007: 11, 140–142.

03_CreSouthV1_8P_3Feb22.indd 109

214. For example, see Kapstein 2000: 43. 215. Weinberger 2010: 146–150; Dalton 2017: 329–330. 216. Weinberger 2003: 9; Yang 2018: 41. 217. Issues concerning the Chinese legacy in Tibetan Buddhism were discussed in Li 2016: 95. Specifically on doubtful colophons in tantric material, see van der Kuijp 1992; Almogi 2008.

3/2/22 6:20 PM

Yury Khokhlov

110

obscured the process by which these collections and their translations came into being and were transmitted. As a result, the role of Chinese and other language translations of Indic texts in Tibetan translation activity and the formation of Tibetan scriptural language remains little understood’.218

Regarding the Tibetan lineages of transmission, it is clear that for such important religious traditions like that of the Yoga, they were of great political consequence, and therefore there is no a priori reason to assume that they are much more reliable than the colophons.

4.1  the main temple of samye As was noted by Weinberger, one of the strongest pieces of evidence for the importance of the Yoga tradition in early Tibetan Buddhism is the first Tibetan monastery of Samye.219 The iconographic programme of its main temple, which was seriously damaged during the Cultural Revolution, can be reconstructed from various historical sources including the Testament of Ba and a Tibetan Chronicle The Mirror Illuminating the Royal Genealogies (Tib. Rgyal rabs gsal ba’i me long).220 In the following, I will try to make sense of it with the help of Kimiaki Tanaka’s recent interpretation.221 The main temple had three floors and a Chinese-type roof.222 On the first floor, the central icon was Buddha Śākyamuni, which was surrounded by sculptures of the eight great bodhisattvas, and two additional bodhisattvas, Acala and Trailokyavijaya. As Tanaka remarked, the combination of the Buddha/Vairocana with the eight great bodhisattvas, Acala, and Trailokyavijaya corresponds to a maṇḍala introduced by Amoghavajra in a ritual manual for recitation of the Uṣṇīṣavijayādhāraṇī (T 972).223 Amoghavajra was the most enthusiastic propagator of this dhāraṇī; he elevated its status from being a scripture with salvific and protective properties for individuals, to a scripture of the 218. Gentry 2019. 219. Weinberger 2010: 140. 220. English translations of these sources can be found in Wangdu and Diemberger 2000 and Sørensen 1994. 221. Tanaka 2018: 83–85. 222. Sørensen 1994: 376–381. 223. See fn. 92.

03_CreSouthV1_8P_3Feb22.indd 110

national defence and protection.224 Bearing in mind the nationwide significance of Samye, the emphasis on apotropaic qualities of Buddhist teachings in its iconographic programme would have made perfect sense. It is not unlikely that the enormous power ascribed to the Uṣṇīṣavijayādhāraṇī in Tang China led to the appropriation of the related rituals in Tibet. A textual fragment in the Tibetan language found in Dunhuang, which describes a ritual involving Vairocana, the eight great bodhisattvas, Acala, and Trailokyavijaya, could be regarded as possible evidence for such appropriation.225 The second floor was centred on Vairocana of an unknown affiliation flanked by the eight great bodhisattvas and two additional bodhisattvas, the same as those on the first floor. Buddhas of the past, present and future; Medicine Buddha; and Buddha Amitābha were installed in front of the Vairocana and bodhisattvas. A pair of Chinese gatekeepers holding a vajra, known as King and Kang, was also included in the composition. It is tempting to suggest that Vairocana and the eight great bodhisattvas on this floor might have represented the practice of the Bhadracarī, which, as discussed above, was an extremely popular aspirational prayer in Imperial Tibet and was believed to prolong the tsenpo’s life. In this regard, it is worth mentioning that other depictions related to the Avataṃsakasūtra subjects were present at Samye as well. While Sudhana’s spiritual journey from the Gaṇḍavyūhasūtra was illustrated in a wall frieze in the great courtyard, the Daśabhūmikasūtra, which describes the ten stages of bodhisattvahood, was illustrated in murals of the main temple’s third floor.226 As mentioned, the only known illustration for the Daśabhūmikasūtra comes from Dunhuang227 (Fig. 3.23) and a similar composition could 224. Lin 2014: 139–140. For example, Emperor Daizong, under the threat from advancing Uighurs and Tibetans in 765, ordered Amoghavajra to install a group of monks at court to chant the dhāraṇī for the sake of the nation. In 776, Daizong ordered all nuns and monks of the country to chant the dhāraṇī. See Yang 2018: 159. 225. The text title is Rnam par snang mdzad ’khor dang bcas pa la bstod pa. It is found in P.T. 7a and in IOL TIB J 366/3. See Tanaka 2018: 81; Tanaka 2000; Heller 1997; Dalton and van Schaik 2006: 97–98. 226. van Schaik and Doney 2007: 187–188; Sørensen 1994: 3bh79. 227. See fn. 176.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) have been depicted at Samye. A principal image of the third floor represented Vairocana with four heads (facing four directions) surrounded by eight unidentified bodhisattvas, two for each head, and many other deities. This group has traditionally been understood as representing the Sarvavid-Vairocana maṇḍala of the Sarvadurgatipariśodhanatantra (hereafter SDP).228 However, as Kimiaki Tanaka has pointed out, Vairocana with four heads could just as well stand for the Vajradhātu-Vairocana, the central deity of the STTS.229 Tanaka further notes that images of the four Buddhas of the Vajradhātu maṇḍala depicted back-to-back were installed at Samye during the modern reconstruction.230 Moreover, Tibetan historical source Ide’u Chos ’byung explicitly states that all images at the third floor were erected as a depiction of the god-assembly of the Vajradhātu.231 The SDP was specifically associated with mortuary rites and thus the alleged installation of the Sarvavid-Vairocana maṇḍala as the central composition of the principal temple requires further explanation. In contrast, a version of the Vajradhātu maṇḍala fits perfectly with the official status of Samye. As discussed, this maṇḍala was the cornerstone of the Yoga tradition and a central element of state protection rituals in contemporaneous China.232 228. Weinberger 2010: 140. The SDP is regarded as an explanatory Tantra on part two of the STTS, see Tanaka 2018: 321. It is centred on Buddha’s Uṣṇīṣa deities and mirrors the narrative found in various versions of the Sūtra of the Revered and Victorious Dhāraṇī of the Buddha’s Uṣṇīṣa (Uṣṇīṣavijayādhāraṇī). Both the SDP and the Uṣṇīṣavijayādhāraṇī were specifically employed for mortuary rituals, see Wang 2018: 25, 42–50; Weinberger 2010: 146. The history of the early translation of the SDP into Tibetan is far from clear. The text is listed in the Imperial catalogue and the colophon, which dates to the 8th century, gives the names of Indian and Tibetan translators, see Weinberger 2010: 148–149. However, the colophon is proven to be unreliable, and the identity of the translators has been questioned: see van der Kuijp 1992. 229.  Vajradhātu-Vairocana can be depicted with four heads as well, but in contrast with Sarvavid-Vairocana, who performs the dhyana-mudrā, Vajradhātu-Vairocana performs the bodhyaṅgī-mudrā. See Tanaka 2018: 83; on four-faced Vairocana in the STTS, see Weinberger 2003: 187. 230. Tanaka 2018: 84. 231. Sørensen 1994: 379, n. 1260. 232. There is strong evidence showing that the Vajradhātu maṇḍala was known in Tibet in the 8th century. See Weinberger 2010: 144.

03_CreSouthV1_8P_3Feb22.indd 111

111

In addition, the main temple at Samye can be compared to the Golden Pavilion of Jin’ge monastery (Ch. 金閣寺) constructed in 767 on Mount Wutai.233 The construction of the monastery was initiated by Amoghavajra who started promoting the cult of Mount Wutai towards the end of his life. Built with lavish imperial sponsorship, it became one of the principal monasteries responsible to carry out apotropaic rituals for the sake of the nation. Similarly with Samye, the Golden Pavilion, a main temple of the monastery, had a three-storey design with a vertically oriented iconographic programme centred around Vairocana. As was suggested by Wei-Cheng Lin, the pavilion was likely conceived to represent a three-dimensional version of the Vajradhātu maṇḍala whose five-fold system correlated with the five monasteries and five peaks of Mount Wutai.234 The Tibetans must have been aware of these developments in China, as the popularity of Wutai Shan in the Tibetan Empire is well attested in historical sources and, according to the Old Dynastic History of the Tang, the Tibetan tsenpo once even requested a picture of Mount Wutai from the Tang court.235 Significantly, the first representations of Mount Wutai appeared in Dunhuag art during the Tibetan period.236 The discussed parallels are striking and suggest that the decision to build Samye might have been motivated by the apotropaic qualities of certain Buddhist teachings. I have later come across a 1983 article by Christopher Beckwith, where he put forward a theory suggesting that the Tibetan adoption of Buddhism as a state religion in general and the building of Samye in particular were inspired by the Chinese model of National Protection Buddhism maintained by Amoghavajra, and were, at least partially, dictated by a need to resolve the crisis of legitimacy faced by the Tibetan emperor Trisong Detsen.237 I was delighted to learn that I 233. On the Golden Pavilion and Jin’ge monastery, see Lin 2014: 131–154. 234. Lin 2014: 149. 235. Debreczeny 2011: 8–9; Wong 1993: 39. 236. Wong 1993: 41. 237. Beckwith 1983. On the notion of National Protection Buddhism in Tang China, see Orzech 1998. In particular, on the term ‘National Protection Buddhism’, see Orzech 1998: 3, n. 9.

3/2/22 6:20 PM

Yury Khokhlov

112

came to the same conclusion independently, but from a different angle. Regrettably, no original images survived from Samye which could provide art historical evidence for sources of their corresponding teachings. Nevertheless, I argue that Samye’s aesthetics can be reconstructed from the artwork created at the earliest stage of Buddhist revival in Tibet in the 11th century.

11th century clay sculptures from central tibetan temples After the collapse of the empire in the mid-9th century, Tibet lost its colonies and went to a socalled period of fragmentation. The monastic tradition perished to be reintroduced only in the late 10th century with a religious movement known in Tibetan history as the ‘later spread of the Dharma’ (Tib. phyi dar). Although most of Hexi was recaptured by the Tang forces, its capital and a former Amoghavajra’s seat, Liangzhou, remained under local Tibetan rule up to the 11th century, when it was absorbed into the Xi Xia Empire.238 The region undoubtedly continued to hold a position of Buddhist authority in Central Tibet, as in the late 10th century a group of men from there were sent to the areas of Liangzhou and its neighbour, the Tibetan kingdom of Tsongkha, in order to bring the monastic tradition back to Central Tibet.239 Being ordained in these areas, the new monks, sometimes labelled as the Eastern Vinaya monks, returned to Central Tibet and initiated the formation of Buddhist communities, the reconstruction of old temples, and the building of new ones. Although Samye and other royal temples were not in use and in a poor state, they still contained sculptures and paintings.240 In addition, a network of small temples continued to function under the patronage of local rulers. According to Tibetan historical sources, hundreds of Buddhist establishments were renovated and built in Central Tibet during the 11th century. Renovation of old temples, symbols of Tibet’s glorious past, was a top priority during this period241 and it can 238. 239. 240. 241.

van Schaik and Galambos 2012: 64–65. Ibid.: 68–74. Davidson 2005: 94–95. On the activities of the eastern Vinaya monks and

03_CreSouthV1_8P_3Feb22.indd 112

therefore be surmised that survived specimens of Imperial art were renovated and replicated in the new temples. For example, Samye monastery was not only renovated, but was also taken as a model for Grwa thang monastery (founded in 1081).242 Thus it is the artistic styles of the late Imperial period that would have been prevalent in the 11th century Buddhist art in Central Tibet. In this regard, I would like to draw attention to a group of clay sculptures from the Central Tibetan temples/monasteries of G.ye dmar, Rtis gnas gsar, Zhwa lu Gser khang, Rkyang bu, Grwa thang, Zho nang, and Snye thang.243 Although most of them were lost during the Cultural Revolution, their images can still be seen in pre-1950 photographs taken by Li Gotami Govinda and Giuseppe Tucci’s expedition.244 Luckily, the sculptures of Snye thang monastery have survived.245 These temples were founded or renovated at the earliest stage of the Buddhist revival in the 11th century. Despite the fact that some of these temples are located at a considerable distance from one another, their sculptures demonstrate remarkable stylistic consistency. This, I believe, confirms that the sculptures reflect Buddhist aesthetics of the Imperial period, which must have been a common source of inspiration for Central Tibetan artists at the time.246 With regard to the iconographic prothe general situation in Central Tibet at the time, see ibid.: 84–115. 242. Vitali 1990: 49; Heller 2002: 40–41; Stoddard 2018: 96–97. 243. These temples were discussed in Vitali 1990: 37–69; Lo Bue 2000; von Schroeder 2001: 836–851. 244. Many photographs are reproduced in publications listed in the previous footnote. Also see Govinda 1979: 40–50. 245. For description and images of the sculptures, see von Schroeder 2001: 860–869. Although von Schroeder dates the sculptures to circa 1200, I believe they belong to the 11th century. 246. Along the same lines, Heather Stoddard suggested that surviving murals at Grwa thang reflect the aesthetics employed at Samye in the late 8th century. Stoddard 2018: 96–97. However, in addition to the imagery in the early style, the murals also include some depictions in the Pāla style of northeastern India, which was introduced to Tibet in the 11th century. On the history and paintings of Grwa thang, see Heller 2002.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

113

Fig. 3.34: Vairocana and Amoghasiddhi as parts of the sculptural Vajradhātu Maṇḍala, clay sculpture, 11th century, Rkyang bu temple, Central Tibet, not extant anymore (photograph by Fosco Maraini, 1937; courtesy of Ulrich von Schroeder).

grammes, much of the subject matter depicted in these temples had already been popular during the Imperial period. For example, four-faced Vairocana with retinue and Vairocana grouped with eight bodhisattvas can be found in the iconographic programmes of both Samye and the temples under discussion. Most remarkably, the sculptures, which depict these groupings, show the same distinctive elements of the Southern Indian artistic mode as they are found in the imagery from Mogao and Yulin. For example, Rkyang bu temple, which was established shortly before 1037,247 contained a sculptural representation of the Vajradhātu maṇḍala with four-faced Vairocana at its centre (Fig. 3.34).248 As suggested above, Vairocana in this form had 247. Vitali 1990: 56–59. 248. This figure has been often misidentified as Sarvavid-Vairocana in the earlier publications.

03_CreSouthV1_8P_3Feb22.indd 113

likely been installed on the third floor of the main temple at Samye. Another four-faced Vairocana is recorded in a temple with a Chinese roof (Tib. rgya phibs) at Rtis gnas gsar, which was founded during the Imperial period and renovated in the 11th century.249 The Vajradhātu sculptures in Rkyang bu temple are decorated with spiked armbands and swagged necklaces, which are characteristics of the Southern Indian mode in Hexi art (Fig. 3.35). The sculptures’ style and splendor show striking parallels with the 8th and 9th century Hexi imagery (see examples in Figs. 3.1, 3.12, 3.13). Their square and tiered thrones are similar to those in the Hexi depictions as well. It is important to stress that these sculptures are more closely related to the earlier Hexi examples than to contemporaneous 10th and 11th century ones.250 The contempora249. Vitali 1990: 51; von Schroeder 2001: 840–841. 250. See Khokhlov 2019: figs. 17, 20, 21.

3/2/22 6:20 PM

114

Yury Khokhlov

Fig. 3.35: Ratnasambhava and attending deities as parts of the sculptural Vajradhātu maṇḍala, clay sculptures, 11th century, Rkyang bu temple, Central Tibet, not extant anymore (photograph by Fosco Maraini, 1937; courtesy of Ulrich von Schroeder).

neous imagery from the Hexi Corridor shows a considerable decline in artistic quality that does not match the refinement and monumentality of these Fig. 3.36: One of the eight bodhisattvas surrounding Tibetan sculptures. This would seem to confirm the figure of Vairocanābhisambodhi, clay sculpture, that the sculptures represent the revival of the Im- 11th century, G.ye dmar temple, Central Tibet perial aesthetics, which in turn were influenced by (photograph by Li Gotami Govinda, 1979; courtesy of Ulrich von Schroeder). the 8th and 9th century art from the Hexi Corridor. Groups of the eight bodhisattvas are recorded in all the temples (Figs. 3.37 and 3.38). In addition, at G.ye dmar there was a group of sixteen bo- bodhisattvas accompany crowned figures displaydhisattvas (Fig. 3.36), which probably represented ing the dhyāna-mudrā, which have been identified sixteen bodhisattvas of the Bhadrakalpa, a promi- either as Vairocanābhisambodhi or as Amitābha/ nent group in the Yoga tradition and members of Amitāyus.252 These deities are differentiated only by the Vajradhātu maṇḍala.251 In all these groups, the the presence/absence of a bowl/vase in their hands. It is possible that initially all of them represented 251. On the sixteen bodhisattvas of the Bhadrakalpa, see Tanaka 2018: 169–170.

03_CreSouthV1_8P_3Feb22.indd 114

252. von Schroeder 2001: fig. XIII-8, 9, 11, 13, 34B.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774)

115

Fig. 3.37: Vairocanābhisambodhi (worshipped as Amitāyus) and one of the eight bodhisattvas, clay sculptures, 11th century, G.ye dmar temple, Central Tibet (photograph by Li Gotami Govinda, 1979; courtesy of Ulrich von Schroeder).

Vairocanābhisambodhi. The bowls/vases could have been added later when the cult of Vairocana declined in popularity.253 As discussed, examples of a group of bodhisattvas and Vairocana displaying the dhyana-mudrā are relatively rare and, apart from Tibet, they can be found in Southern India and in the Hexi Corridor. Regarding artistic features, the Tibetan sculptures are executed either in the Tibetan mode (clad in Tibetan-style robes) 253. The same conversion is evident in the early 9th-century rock carving at ’Bis khog in Eastern Tibet, where Vairocana (identified by inscription) holds a bowl in his hands (see Fig. 3.10). Another example of Vairocana who was probably converted into Amitābha at some point can be seen in the main temple of Tabo. See Klimburg-Salter and Luczanits 1998: 101, 143–146, fig. 146.

03_CreSouthV1_8P_3Feb22.indd 115

(Figs. 3.36 and 3.37), or the Indian mode (Fig. 3.38). These two modes of representation can already be found in the early 9th century rock carvings in Eastern Tibet (Figs. 3.9–3.11). Significantly, the sculptures in both the Tibetan and Indian modes are adorned with spiked armbands that is a clear indication of the influence from the Hexi Corridor. In Figs. 3.36 and 3.37, I have highlighted the armbands with blue arrows; in Fig. 3.38, some remaining spikes are clearly visible in the armbands of the bodhisattva. Despite substantial renovation, the figures of Vairocana and the bodhisattva in Fig. 3.38 still convey the same splendid aesthetics as that we see in the late 8th century images from Yulin Cave 25 (in particular compare Vairocana in Fig. 3.38 with Vairocana in Fig. 3.2).

3/2/22 6:20 PM

Yury Khokhlov

116

Fig. 3.38: Vairocanābhisambodhi (worshipped as Amitāyus) and one of the eight bodhisattvas, clay sculptures, 11th century, H 310 cm and 330 cm, Snye Thang monastery, Central Tibet (photograph by Alain Bordier, 1998; courtesy of Ulrich von Schroeder).

As it was already stated, since no direct contacts between the Tibetan Empire and Southern Indian polities are known to date, the source of the Southern Indian features in Tibetan art can only be Hexi/China. Also, it is very important to note that not all discussed sculptures display these elements. They are mostly found in sculptures which represent subjects of the Yoga tradition and Vairocanābhisambodhi with bodhisattvas.254 The same correlation is evident in Hexi art. As highlighted, imagery played a pivotal role in the transmission of Esoteric teachings, hence it can be deduced that these parallels between the Tibetan and Hexi depictions reflect not only the artistic transmission, 254. For depictions which do not display the element of the Southern Indian mode, see sculptures at G.ye dmar in von Schroeder 2001: fig. XIII-10; and murals at Grwa thang in Vitali 1990: plates 29–34.

03_CreSouthV1_8P_3Feb22.indd 116

but also the transmission of the Yoga tradition, as well as the cult of VAEGB from the Hexi Corridor to Tibet. Although the Chinese textual antecedents and visual evidence from Hexi provide solid grounds for this conclusion, much research needs to be done in the area of Tibetan textual scholarship. Such research could potentially uncover additional textual evidence as well as possible traces of Chinese lexical influence in the relevant Tibetan texts. From a larger perspective, these stylistic and iconographic traits from Hexi, which are detected in the early Tibetan imagery, provide a strong support of the Chinese background of the Tibetan adoption of Buddhism as a state religion. Finally, in parallel with the Southern Indian mode in Hexi art, the artistic mode represented by these Tibetan sculptures gradually vanished in Tibet after the 11th century. Its disappearance likely reflects a decline of interest in the older teachings

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) and their imagery caused by the arrival of new teachings and corresponding imagery in the Pāla style from northeastern India.255

conclusion

117

priation of Chinese religious and artistic trends, which were popular in the region in the 8th century. Furthermore, the artistic and iconographical considerations presented in this research support the theory that the Tibetan adoption of Buddhism as a state religion was most likely based on the Chinese model of National Protection Buddhism established by Amoghavajra.

In tracing the artistic background and doctrinal affiliations of a group of images from Mogao and Yulin, this research elucidates the intricate web of politics, religion and art that connected the Pallava appendix 3.1 Kingdom of Southern India, Tang China and Imperial Tibet. Addressing a period of particularly A list of paintings from Mogao and Yulin that intense Buddhist interexchange between South display the Southern Indian artistic mode disand East Asia, this article contributes to reshaping cussed in this chapter our understanding of several key issues. Firstly, it This list is not exhaustive and contains only the highlights considerable Southern Indian traits in most important published examples. Buddhist art of Tang China and identifies a distinctive artistic tradition, which was most likely 1) A wall painting of Kubera with two attendants, Yulin Cave 15, Tibetan period, Fig. 3.17. established in China by Vajrabodhi, a renowned Buddhist master as well as a painter from the 2) A wall painting of Vairocanābhisambodhi and the eight great bodhisattvas (the deities are Pallava Kingdom of Southern India. Specific styidentified by accompanying Chinese inscriplistic and iconographic features found in the Duntions), Yulin Cave 25, Tibetan period, late 8th huang imagery can be confidently traced to the century, Fig. 3.1. art of the Pallava kingdom in Tamil Nadu and to 3) A portable painting of the Amoghapāśa the Pallava influenced art in the Western Deccan. from Mogao Cave 17, 8th or 9th maṇḍala, The subjects of these representations belong to century, Musée Guimet, MG 26466, Fig. 3.14. the Yoga tradition and the cult of the eight great bodhisattvas, which were propagated in China by 4) A portable painting labelled as ‘the Pure Land of Bhaiṣajyaguru’, from Mogao Cave 17, dated Vajrabodhi and Amoghavajra. 836, British Museum, Ch.xxxvii.004, Fig. 3.22. On the other hand, this chapter helps in recon5) Two wall paintings depicting maṇḍala-like structing artistic features of Buddhist art of Tamil compositions centred around VairocanābhiNadu that was largely lost in India. In addition, the sambodhi and Vajrasattva, Mogao Cave 14, Pallava influenced mode of the discussed Chinese second half of the 9th century, Figs. 3.12 and and Tibetan images constitutes material evidence 3.13. for the popularity of the Yoga tradition and the 6) A portable painting depicting a heavenly cult of the eight great bodhisattvas in Tamil Nadu assembly, from Mogao Cave 17, 9th century, in the 8th century. British Museum, Ch.ivi.0034. Last, but not least, this research defies the idea 7) A portable painting depicting Mañjuśrī riding of ‘early Tibetan Buddhist art’ as represented by the a lion, from Mogao Cave 17, 9th century, artwork created in the Hexi Corridor during and British Museum, Ch.XXVI.a.007. after the period of Tibetan rule. Moreover, it shows 8) Four large images of seated Avalokiteśvara quite the opposite and suggests that the conquest with four Offering Goddesses, painted on four of the Hexi Corridor enabled the Tibetan approslopes of the ceiling, Mogao Cave 161, 9th or 10th century, discussed in Wang 2018: 145–150, Figs. 63–66. 255. In a forthcoming article, I argue that in this process 9) Two wall paintings depicting maṇḍala-like the Hexi Corridor played a crucial role as well, and these northeastern Indian teachings and Pāla aesthetics were compositions centred around Vairocanābhiinitially introduced to Tibet through Hexi. See fn. 178 and sambodhi and a buddha, Yulin Cave 20, 9th Khokhlov 2018. century.

03_CreSouthV1_8P_3Feb22.indd 117

3/2/22 6:20 PM

118

Yury Khokhlov

10) A portable painting depicting a standing figure of Samantabhadra, from Mogao Cave 17, 9th or 10th century, Musée Guimet, MG. 17770, Fig. 3.19. 11) A portable painting depicting a standing bodhisattva, from Mogao Cave 17, 9th or 10th century, British Museum, Ch.Iv. 007. 12) A portable painting depicting a standing bodhisattva, from Mogao Cave 17, 9th or 10th century, British Museum, Ch.Iv.0031. 13) A portable painting depicting a standing bodhisattva, from Mogao Cave 17, 9th or 10th century, British Museum, Ch.Iv.004. 14) A portable painting depicting a seated figure of Avalokiteśvara, from Mogao Cave 17, dated by the museum to the 9th century but likely belonging to the 10th century, British Museum, Ch.00401. 15) A portable painting depicting seated bodhisattva, from Mogao Cave 17, dated by the museum to the 9th century but likely belonging to the 10th century, British Museum, Ch.00377. 16) A portable painting depicting a maṇḍala of Avalokiteśvara, from Mogao Cave 17, dated by the museum to the 8th century but likely belonging to the 10th century, British Museum, Ch.xxii.0017. 17) A portable painting depicting a maṇḍala of Avalokiteśvara, from Mogao Cave 17, dated by the museum to the 8th or 9th century but likely belonging to the 10th century, British Museum, Ch.Iv.0024. 18) A portable painting of Vairocanābhisambodhi and the eight great bodhisattvas (the deities are partially identified by Tibetan inscriptions), from Mogao Cave 17, late 9th or early 10th century, British Museum, Ch.0074, Fig. 3.8. 19) A portable painting of the Amoghapāśa maṇḍala, from Mogao Cave 17, 10th century, Musée Guimet, EO.3579. 20) A portable painting of Amoghapāśa maṇḍala, from Mogao Cave 17, dated by the museum to the 8th century but likely belonging to the 10th century, Musée Guimet, EO.1131. 21) A portable painting of the ten stages of bodhisattvahood, from Mogao Cave 17, 10th century, Musée Guimet, MG. 26465, Fig. 3.23.

03_CreSouthV1_8P_3Feb22.indd 118

22) Two wall paintings depicting maṇḍala-like compositions centred around Vairocanābhisambodhi and a buddha, Yulin Cave 38, 10th or 11th century. 23) Wall paintings inside the Heavenly King stūpa (Ch. 天王堂) at Mogao, 10th century. The content was only partially published and includes: – six multi-armed deities, each surrounded by four Offering Goddesses/Bodhisattvas, painted on the ceiling. – eight standing bodhisattvas and four heavenly kings painted on the walls. Discussed in Dunhuang Academy 2003: 197– 212. 24) A portable painting (fragment) of a maṇḍala, from Mogao Cave 17, 10th century, Indian National Museum, Ch.00383 c, discussed in Klimburg-Salter 1982: 137, plates 62, 63. 25) A wooden stupa reliquary, found at Yulin, 10th century, Gansu Provincial Museum, discussed in Shen and Gardner Gates 2016: 28, fig. 8.

References

Acri, Andrea. 2016. Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons. Singapore: ISEAS – Yusof Ishak Institute. . 2018. ‘Maritime Buddhism’, in Oxford Research Encyclopedia of Religion. Oxford: Oxford University Press. [Last accessed 14 November 2019 at: http://oxfordre.com/religion/view/ 10.1093/acrefore/9780199340378.001.0001/acrefore-9780199340378-e-638] Aiyappan, Ayinapalli and P.R. Srinivasan. 2000. Story of Buddhism with Special Reference to South India. Chennai: Commissioner of Museums, Govt. Museum. Almogi, Orna. 2008. ‘How Authentic are Titles and Colophons of Tantric Works in the Tibetan Canon? The Case of Three Works and Their Authors and Translators’, Beitrage Zur Zentralasienforschung 14: 87–124. Asha Rani, Mathur. 1988. The Great Traditions: Indian Bronze Masterpieces. New Delhi: Festival of India.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) Auboyer, Jeannine. 1941. Les Influences et les Réminiscences Étangères au Kondō Du Hōryūji, Dessins de l’Auteur. Paris: P. Geuthner. Barber, A.W. 2008. ‘Two Mahāyāna Developments Along the Krishna River’, in Sree Padma (Holt) and A.W. Barber (eds.), Buddhism in the Krishna River Valley of Andhra, pp. 151–167. Albany: State University of New York Press. Bautze-Picron, Claudine. 1997. ‘Le groupe des huits grandes bodhisatva en Inde: genèse et développement’, in Natasha Eilenberg (ed.), Living a Life in Accord with the Dhamma: Papers in Honor of Professor Jean Boisselier on his Eightieth Birthday, pp. 1–55. Bangkok: Silpakorn University. . 2000. ‘Nasik: The Late Mahâyâna Caves 2, 15, 20 & 23-24’, in European Association of South Asian Archaeologists, Maurizio Taddei and Giuseppe De Marco (eds.), South Asian Archaeology 1997: Proceedings of the Fourteenth International Conference of the European Association of South Asian Archaeologists, held in the Istituto italiano per l’Africa e l’Oriente, Palazzo Brancaccio, Rome, 7–14 July 1997, pp. 1201–1227. Rome: Istituto Italiano per l’Africa e l’Oriente. . 2002. ‘Nidhis and other images of richness and fertility in Ajanta’, East and West 52: 225– 284. . 2008. ‘Antagonistes et complémentaires, le lion et l’éléphantdans la personnalité du Buddha’, in Nalini Balbir and Georges-Jean Pinault (eds.), Penser, dire et représenter l’animal dans le monde indien, pp. 523–572. Paris: Librairie Honoré Champion, Éditeur. . 2015. The Forgotten Place: Stone Images from Kurkihar, Bihar. New Delhi: Archaeological Survey of India. Beckwith, Christopher. 1983. ‘The Revolt of 755 in Tibet’, in Ernst Steinkener and Helmut Tauscher (eds.), Contributions on Tibetan Language, History, and Culture: Proceedings of the Csoma de Koros Symposium Held at Velm—Vienna, Austria, Vol. I, pp. 1–16. Vienna: Arbeitskreis fur Tibetische und Buddhistische Studien Universität Wien. Bopearachchi, Osmund. 2014. ‘Sri Lanka and the Maritime Trade: Bodhisattva Avalokiteśvara

03_CreSouthV1_8P_3Feb22.indd 119

119

as the Protector of Mariners’, in Parul Pandya Dhar and Upinder Singh (eds.), Asian Encounters: Exploring Connected Histories, pp. 161–187. New Delhi: Oxford University Press. Brauen, Martin. 1997. The Mandala: Sacred Circle in Tibetan Buddhism. 1st ed. Boston: Shambhala. Chou, Yi-liang. 1945. ‘Tantrism in China’, Harvard Journal of Asiatic Studies 8/3–4: 241–332. Cleary, Thomas F. 1993. The Flower Ornament Scripture. Boston: Shambhala. Dalton, Jacob P. 2011. The Taming of the Demons: Violence and Liberation in Tibetan Buddhism. New Haven and London: Yale University Press. . 2017. ‘On the Significance of the Āryatattvasaṃgraha-sādhanopāyikā and Its Commentary’, in Yael Bentor and Meir Shahar (eds.), Chinese and Tibetan Esoteric Buddhism, pp. 321–337. Leiden: Brill. Dalton, Jacob P. and Sam van Schaik. 2006. Tibetan Tantric Manuscripts from Dunhuang: A Descriptive Catalogue of the Stein Collection at the British Library. Leiden and Boston: Brill. Davidson, Ronald M. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New York: Columbia University Press. . 2005. Tibetan Renaissance: Tantric Buddhism in the Rebirth of Tibetan Culture. New York: Columbia University Press. Dayalan, Duraiswamy. n.d. Art, Architecture and Cultural Link Between Maharastra and South India (Andhra Pradesh & Tamil Nadu). [Last accessed 24 November 2020 at: https://www.academia.edu/25250507/ART_ ARCHITECTURE_AND_CULTURAL_ LINK_BETWEEN_MAHARASHTRA_AND_ SOUTH_INDIA_ANDHRA_PRADESH_and_ TAMIL_NADU] de Visser, Marinus Willem. 1915. The Bodhisttva Ti-tsang (Jizō) in China and Japan. Berlin: Oesterheld & Co. Verlag (Erste Sonderveröffentlichung der Ostasiatischen Zeitschrift). Debreczeny, Karl. 2011. ‘Wutai Shan: Pilgrimage to Five-Peak Mountain’, Journal of the International Association of Tibetan Studies 6: 1–133. . 2019. ‘Faith and Empire: An Overview’, in Karl Debreczeny (ed.), Faith and Empire: Art

3/2/22 6:20 PM

120

Yury Khokhlov

and Politics in Tibetan Buddhism, pp. 17–53. New York: Rubin Museum of Art. Dessein, Bart. 2003. ‘The Glow of the Vow of the Teacher Samantabhadra “Puxian Pusa Xingyuan Zan” (T.297) Samantabhadracaryāpraṇidhānarāja’, Acta Orientalia Academiae Scientiarum Hungaricae 56/2–4: 317–338. Dohanian, Diran K. 1977. The Mahāyāna Buddhist Sculpture of Ceylon. New York: Garland House. . 1983. ‘Sinhalese Sculptures in the Pallava Style’, Archives of Asian Art 36: 6–21. Dunhuang Academy (Dunhuang yanjiuyuan) (eds.). 2003. The Complete Collection of Dunhuang Grottoes (Dunhuang Shiku Quanji). Hong Kong: The Commercial Press (Shangwu yinshuguan). . 2013. Chinese Grottoes: Dunhuang Mogao Grottoes, Volume 4. (Zhongguo shiku: Dunhuang Mogaoku (si)). 2nd ed. Beijing: Cultural Relics Publisher (Wenwu chubanshe). . 2014. The Yulin Grottoes (Yulin Ku). Nanjing: Jiangsu Arts Publisher (Jiangsu meishu chubanshe). Fontein, Jan. 1968. The Pilgrimage of Sudhana: A Study of Gaṇḍavyūha Illustrations in China, Japan and Java. The Hague: Mouton. . 2012. Entering the Dharmadhātu: A Study of the Gandavyūha Reliefs of Borobudur (Studies in Asian Art and Archaeology, vol. 26). Leiden and Boston: Brill. Fraser, Sarah E. 2004. Performing the Visual: The Practice of Buddhist Wall Painting in China and Central Asia, 618–960. Stanford, CA: Stanford University Press. Gentry, James. 2019. ‘An Inquiry into Source-Text Multilingualism in Tibetan Scriptural Translations’. Abstract of a talk delivered at the 15th Seminar of the International Association for Tibetan Studies in Paris, July 2019. [Last accessed 24 November 2020 at: https://www.academia.edu/39792552/An_Inquiry_into_Source Text_Multilingualism_in_Tibetan_Scriptural_Translations] Giebel, Rolf W. 1995. ‘The Chin-kang-ting ching yü-ch’ieh shih-pa-hui chih-kuei: An Annotated Translation’, Journal of Naritasan Institute for Buddhist Studies 18: 107–201.

03_CreSouthV1_8P_3Feb22.indd 120

. 2001. Two Esoteric Sutras: The Adamantine Pinnacle Sutra, the Susiddhikara Sutra. Berkeley, California: Numata Center for Buddhist Translation and Research. Goble, Geoffrey. 2012. Chinese Esoteric Buddhism: Amoghavajra and the Ruling Elite. PhD thesis, Indiana University. . 2013. ‘The Legendary Siege of Anxi: Myth, History, and Truth in Chinese Buddhism’, Pacific World: Journal of the Institute of Buddhist Studies 3/15: 1–32. Govinda, Li Gotami. 1979. Tibet in Pictures: Text and Photos. Berkeley, CA: Dharma Pub. Granoff, Phyllis. 1968/1969. ‘A Portable Buddhist Shrine from Central Asia’, Archives of Asian Art 22: 80–95. Guy, John. 1994. ‘The Lost Temples of Nagapattinam and Quanzhou: A Study in Sino-Indian Relations’, in Silk Road Art and Archeology 3: 291–310. . 1995. ‘The Avalokitesvara of Yunnan and Some South East Asian Connections’, in Rosemary Scott and John Guy (eds.), South East Asia and China: Art, Commerce and Interaction, Percival David Foundation Colloquies on Art and Archeology of Asia, pp. 64–83. London: University of London. . 2014. ‘Introducing Early Southeast Asia’, in John Guy (ed.), Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia. New York: Metropolitan Museum of Art. Heller, Amy. 1994. ‘Ninth Century Buddhist Images Carved at Ldan Ma Brag to Commemorate Tibeto-Chinese Negotiations’, in Per Kvaerne (ed.), Tibetan Studies: Proceedings of the 6th Seminar of the International Association for Tibetan Studies, Fagernes, 1992, pp. 335–349. Oslo: Institute for Comparative Research in Human Culture. . 1997. ‘P.T. 7A, P.T. 108, P.T. 240 and Beijing Bstan ‘gyur 3489: Ancient Tibetan Rituals Dedicated to Vairocana’, in Gene Smith (ed.), The Pandita and the Siddha: Tibetan Studies in Honor of E. Gene Smith, pp. 85–91. Dharamsala: Amnye Machen Institute. . 2002. ‘The Paintings of Gra thang: History and Iconography of an 11th Century Tibetan Temple’, Tibet Journal 27/1–2: 39–72.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) Horlemann, Bianca. 2012. ‘Buddhist Sites in Amdo and former Longyou from the 8th to the 13th Century’, in Cristina Scherrer-Schaub (ed.), PIATS Oxford 2003, Old Tibetan Studies Dedicated to the Memory of Ronald E. Emmerick, pp. 119–157. Leiden: Brill. Hunter, Harriet. 2018. A Transmission and Its Transformation: The Liqujing Shibahui Mantuluo in Daigoji Research. PhD dissertation, Leiden University. Huntington, John C. 1981. ‘Cave Six at Aurangabad: A Tantrayana Monument?’ in J.G. Williams (ed.), Kalādarśana: American Studies in the Art of India, pp. 47–55. Leiden: E.J. Brill. Huo, Wei. 2017. ‘Archaeological Investigation on Parinirvana in Yushu of Qinghai and New Findings (Qinghai Yushu darirulai fotang de kaogu diaocha yu xinfaxian)’, Qinghai Ethnic Research (Qinghai minzu yanjiu) 28/1: 4–9. . 2018. ‘Rock Reliefs in East Tibet and Yizhou in the Context of Cultural Exchanges between Tang and Tubo’, Journal of Tibetology 16: 301–313. Indrapala, Karthigesu. 2005. The Evolution of an Ethnic Identity: The Tamils in Sri Lanka c.300 BCE to c.1200 CE. Sydney: MV Publications for the South Asian Study Centre. Kapstein, Matthew T. 2000. The Tibetan Assimilation of Buddhism: Conversion, Contestation, and Memory. Oxford: Oxford University Press. . 2004. ‘The Treaty Temple of De-Ga G.YuTshal: Identification and Iconography’, in Huo Wei and Y. Li (eds.), Tibetan Archaeology and Art (Xizang Kaogu Yu Yishu), pp. 98–127. Chengdu: Sichuan renmin chubanshe. . 2009. ‘The Treaty Temple of the Turquoise Grove’, in Matthew T. Kapstein (ed.), Buddhism Between Tibet and China, pp. 21–72. Boston: Wisdom Publications. . 2014. ‘The Treaty Temple of De ga g.yu tshal: Reconsiderations’, Journal of Tibetology 10: 32–34. Karmay, Heather. 1975. Early Sino-Tibetan Art. Warminster: Aris and Philips lts. Karmay, Samten Gyaltsen. 2007. The Great Perfection (rDzogs chen): A Philosophical and Meditative Teaching of Tibetan Buddhism. Leiden and Boston: Brill.

03_CreSouthV1_8P_3Feb22.indd 121

121

Khokhlov, Yury. 2018. ‘From India to the Doorstep of China: Indian Pāla Style in the Hexi Corridor’, unpublished paper based on a talk given at The Silk Road and the Art of the Tibetan Plateau Forum (Pāla Art and its Influence in China, 青藏高原絲路與藝術論壇—波羅藝術與中國), Asia Society, Hong Kong, 26–27 May 2018. [Abstract last accessed 14 November 2019 at: https://www.academia.edu/36748689/From_ India_to_the_Doorstep_of_China_Indian_ Pala_Style_in_the_Hexi_Corridor] . 2019. ‘Uncovering Amoghavajra’s Legacy in the Hexi Corridor and Tibet’, Journal of Tibetology 21: 48–133. Klimburg-Salter, Deborah E. 1982. The Silk Route and the Diamond Path: Esoteric Buddhist Art on the Trans-Himalayan Trade Routes. Los Angeles: UCLA Art Council. Klimburg-Salter, Deborah E. and Christian Luczanits. 1998. Tabo: A Lamp for the Kingdom: Early-Indo-Tibetan Buddhist Art in the Western Himalaya. New York: Thames and Hudson. Lee, Sonya S. 2010. Surviving Nirvana: Death of the Buddha in Chinese Visual Culture. Hong Kong: Hong Kong University Press. Li, Channa. 2016. ‘Translationship Lost in Transmission: Elusive Attributions of Two Tibetan Sūtra Translations’, Revue d’Études Tibétaines 37: 207–230. Lin Li-Kuang. 1935. ‘Puṇyodaya (Na-t’i), un propagateur du tantrisme en Chine et au Cambodge à l’époque de Hiuan-tsang’, Journal Asiatique 227: 83–100. Lin Wei-Cheng. 2014. Building a Sacred Mountain: The Buddhist Architecture of China’s Mount Wutai. Seattle: University of Washington Press. Linrothe, Robert N. 2014. ‘What They Saw: Kaśmīrā in the Eyes of Western Himalayas’, in Collecting Paradise: Buddhist Art of Kashmir and its Legacies, exh. cat., pp. 29–107. New York: Rubin Museum of Art. Liu Yongzhen. 2008. ‘Explanation of the Ceiling Images in the Southern and Northern Alcoves of Mogao Cave 148 (Mogaoku di 148 ku nanbeikan tianjing tuxiang jieshuo)’, in Duhuang Research Academy (ed.), The Inheritance and

3/2/22 6:20 PM

122

Yury Khokhlov

Innovation of Dunhuang Wall Art: Proceedings of the International Conference (Duhunag Bihua Yishu Jicheng yu Chuangxin: Guoyi Xueshu Yantaohui Lunwenji), pp. 518–536. Shanghai: Shanghai cishu chubanshe. Lo Bue, Erberto. 2000. Tibet: Templi scomparsi fotografati da Fosco Maraini. Torino: Ananke. Lokesh Chandra. 1980. ‘Oḍḍiyāna: A New Interpretation’, in Michael Aris and Aung San Suu Kyi (eds.), Tibetan Studies in Honour of Hugh Richardson, pp. 73–78. Warminster and Oxford: Aris and Phillips. Long, Mark. 2009. Caṇḍi Mendut: Womb of the Tathāgata, written, compiled, translated and illustrated. New Delhi: Aditya Prakashan. Malandra, Gery Hockfield. 1993. Unfolding a Mandala: The Buddhist Cave Temples at Ellora. Suny Series in Buddhist Studies. Albany, N.Y.: State University of New York Press. . 1996. ‘The Mandala at Ellora/Ellora in the Mandala’, Journal of the International Association of Buddhist Studies 19: 181–207. McCoy, Michelle. 2016. ‘Maṇḍala of Avalokiteśvara of the Unfailing Lasso (Amoghapāśa)’, in Neville Agnew, Marcia Reed and Tevvy Ball (eds.), Cave Temples of Dunhuang: Buddhist Art of China’s Silk Road, pp. 236–239. Los Angeles: J. Paul Getty Trust. McNair, Amy. 2007. Donors of Longmen: Faith, Politics, and Patronage in Medieval Chinese Buddhist Sculpture. Honolulu: University of Hawai’i Press. Misra, R.N. 1981. Yaksha Cult and Iconography. New Delhi: Munshiram Manoharlal. Nagasawa, Jake. 2017. [Last accessed 14 November 2019 at: https://treasuryoflives.org/biographies/ view/Buddhaguhya/10546] Okada, Ken. 2004. ‘Observations on Buddhist Stone Sculptures from the Longxing Monastery’, in Xiaoneng Yang (ed.), New Perspectives on China’s Past: Chinese Archaeology in the Twentieth Century, pp. 381–391. New Haven, CT: Yale University Press. Orzech, Charles D. 1995. ‘The Legend of the Iron Stūpa’, in Donald S. Lopez, Jr. (ed.), Buddhism in Practice, pp. 314–317. NJ: Princeton University Press.

03_CreSouthV1_8P_3Feb22.indd 122

. 1996. ‘Maṇḍalas on the Move: Reflections from Chinese Esoteric Buddhism Circa 800’, Journal of the International Association of Buddhist Studies 19: 209–245. . 1998. Politics and Transcendent Wisdom: The Scripture for Humane Kings in the Creation of Chinese Buddhism. University Park: The Pensilvania State University Press. . 2006. ‘The “Great Teaching of Yoga,” the Chinese Appropriation of the Tantras, and the Question of Esoteric Buddhism’, Journal of Chinese Religions 34: 29–78. Orzech, Charles D., Henrik H. Sørensen, and Richard K. Payne (eds.). 2011. Esoteric Buddhism and the Tantras in East Asia. Leiden and Boston: Brill. Osto, Douglas. 2004. The Gaṇḍavyūha-sūtra: A Study of Wealth, Gender and Power in an Indian Buddhist Narrative. PhD dissertation, School of Oriental and African Studies, University of London. . 2010. ‘A New Translation of the Sanskrit Bhadracarī with Introduction and Notes’, New Zealand Journal of Asian Studies 12/2: 1–21. Owen, Lisa N. 2010. ‘Demarcating Sacred Space: The Jina Images at Kalugumalai’, International Journal of Jaina Studies 6/4: 1–28. . 2012. Carving Devotion in the Jain Caves at Ellora. Leiden and Boston: Brill. Padma, Sree and A.W. Barber. 2008. Buddhism in the Krishna River Valley of Andhra. Albany: State University of New York Press. Rao, Gopinatha T.A. 1915. ‘Bauddha Vestiges in Kanchipura’, Indian Antiquary 44: 127–129. Reis Habito, Maria. 1999. ‘The Amoghapāśa Kalparāja Sūtra: A Historical and Analytical Study’, Studies in Central and East Asian Religions 11: 39–67. Revire, Nicolas. 2012. ‘New Perspectives on the Origin and Spread of Bhadrāsana Buddhas throughout Southeast Asia (7th–8th centuries)’, in Mai Lin Tjoa-Bonatz, Andreas Reinecke and Dominik Bonatz (eds.), Connecting Empires and States: Selected Papers from the 13th Conference of the European Association of Southeast Asian Archaeologists, Vol. 2, pp. 127–143. Singapore: NUS Press.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) Rhie, Marilyn M. 1988. Interrelationships between the Buddhist Art of China and the Art of India and Central Asia from 618-755 A.D. Napoli: Istituto Universitario Orientale. Sanderson, Alexis. 2007. ‘The Śaiva Exegesis of Kashmir’, in Dominic Goodall and André Padoux (eds.), Mélanges tantriques à la mémoire d’Hélène Brunner: Tantric Studies in Memory of Hélène Brunner, pp. 231–442. Pondicherry: Institut français de Pondichéry/École française d’Extrême-Orient. Schmid, Neil. 2006. ‘The Material Culture of Exegesis and Liturgy and a Change in the Artistic Representations in Dunhuang Caves, ca.7001000’, Asia Major, 3rd series, Vol. 19, No. 1/2 (China at the Crossroads: A Festschrift in Honor of Victor H. Mair), pp. 171–210. . 2011. ‘Dunhuang and Central Asia’, in Charles Orzech, Henrik Sørensen and Richard Payne (eds.), Esoteric Buddhism and the Tantras in East Asia, pp. 365–374. Leiden and Boston: Brill. Sen, Tansen. 2016. Buddhism, Diplomacy, and Trade: The Realignment of India-China Relations, 600–1400. Lanham: Rowman & Littlefield. Sha Wutian. 2013. Studies on Dunhuang Grottoes during the Tubo Period (Tubo Tongzhi Shiqi Dunhuang Shiku Yanjiu). Beijing: Zhongguo shehui kexue chubanshe. Shanxi Province Academy of Archaeology and TAR Relics Preservation Institute (Shanxisheng kaogu yanjiuyuan, Xizang zizhiqu wenwu baohu yanjiusuo). 2014. ‘Archaeological Report on Brag gYab County lDan Ma Brag Cliff Carvings (Xizang chayaxian danzha moya zaoxiang kaogu diaocha jianbao)’, Archaeology and Relics (Kaogu yu wenwu) 6: 7–14. Sharf, Robert. 2017. ‘Buddhist Veda and the Rise of Chan’, in Yael Bentor and Meir Shahar (eds.), Chinese and Tibetan Esoteric Buddhism, pp. 85–120. Leiden and Boston: Brill. Sharrock, Peter D. and Emma C. Bunker. 2016. ‘Seeds of Vajrabodhi: Buddhist Ritual Bronzes from Java and Khorat’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons, pp. 237–252. Singapore: ISEAS – Yusof Ishak Institute.

03_CreSouthV1_8P_3Feb22.indd 123

123

Shen, Hsueh-man. 2019. Authentic Replicas: Buddhist Art in Medieval China. Honolulu: University of Hawai‘i Press. Shen, Hsueh-man and Mimi Gardner Gates. 2016. ‘The Buddhist Cave Temples in Dunhuang Art, Spirituality, Cultural Heritage’, in Neville Agnew, Marcia Reed, and Tevvy Ball (eds.), Cave Temples of Dunhuang: Buddhist Art of China’s Silk Road, pp. 139–165. Los Angeles: J. Paul Getty Trust. Shim, Yeoung Shin. 2013. Four Heavenly Kings: Iconography and Symbolism seen Through Literary Evidence and Imagery. PhD thesis, The University of Pennsylvania. Skilling, Peter. 2011. ‘Buddhism and the Circulation of Ritual in Early Peninsular Southeast Asia’, in Pierre-Yves Manguin, A. Mani, and Geoff Wade (eds.), Early Interactions between South and Southeast Asia: Reflections on Cross-cultural Exchange, pp. 371–384. Singapore: Institute of Southeast Asian Studies. Snellgrove, David L. 1978. The Image of the Buddha. London: Serindia Publications/UNESCO. . 1987. Indo-Tibetan Buddhism: Indian Buddhists and their Tibetan Successors. London: Serindia Publications. Soper, Alexander C. 1960. ‘South Chinese Influence on the Buddhist Art of the Six Dynasties Period’, Bulletin of the Museum of Far Eastern Antiquities 32: 47–112. Sørensen, Henrik H. 2000. ‘Perspectives on Buddhism in Dunhuang during the Tang and Five Dynasties Period’, in Vadime Elisseeff (ed.), The Silk Roads: Highways of Culture and Commerce, pp. 27–48. New York and Oxford: Berghahn Books. . 2011. ‘Central Divinities in Esoteric Buddhist Pantheon in China’, in Charles Orzech, Henrik Sørensen, and Richard Payne (eds.), Esoteric Buddhism and the Tantras in East Asia, pp. 90–132. Leiden and Boston: Brill. Sørensen, Per. 1994. The Mirror Illuminating the Royal Genealogies: Tibetan Buddhist Historiography: An Annotated Translation of the XIVth century Tibetan Chronicle: rGyal-rabs gsal- ba’i me-long. Weisbaden: Harrassowitz. Srinivasan, Sharada. 1999. ‘Lead Isotope and Trace Element Analysis in the Study of over One

3/2/22 6:20 PM

124

Yury Khokhlov

Hundred South Indian Metal Icons’, Archaeometry 41: 91–116. Stoddard, Heather. 2018. ‘Drathang Monastery’, in Thomas Liard, Robert Thurman, Heather Stoddard, Cameron Bailey, and Jakob Winkler (eds.), Murals of Tibet, pp. 96–105. Milan, New York, Berlin, London: Taschen. Sundberg, Jeffrey R. and Rolf Giebel. 2011. ‘The Life of the Tang Court Monk Vajrabodhi as Chronicled by Lü Xiang (呂向): South Indian and Śrī Laṅkān Antecedents to the Arrival of the Buddhist Vajrayāna in Eighth-Century Java and China’, Pacific World: Journal of the Institute of Buddhist Studies 13/3: 129–222. Takata, Tokio. 2019. ‘Tibetan Dominion over Dunhuang and the Formation of a Tibeto-Chinese Community’, in Erika Forte (ed.), Buddhist Road Paper 6.1. Special Issue: Central Asian Networks. Rethinking the Interplay of Religions, Art and Politics Across the Tarim Basin (5th–10th c.), pp. 85–106. Tanaka, Kimiaki. 2000. Tonkō: Mikkyō to bijutsu 敦煌: 密教と美術. Kyoto: Hōzōkan. . 2018. An Illustrated History of the Maṇḍala: From its Genesis to the Kālacakratantra. Boston: Wisdom Publications. Thapar, Romila. 2003. Early India: From the Origins to AD 1300. Berkeley, CA: University of California Press. Toyka-Fuong, Ursula. 1998. ‘The Influence of Pāla Art on 11th-Century Wall-Painting of Grotto 76 in Dunhuang’, in Deborah Klimburg-Salter and Eva Allinger (eds.), The Inner Asian International Style 12th-14th Centuries, pp. 67–96. Vienna: Verlag der Österreichischen Akademie der Wissenschaften. van der Kuijp, Leonard W.J. 1992. ‘Notes Apropos of the Transmission of the Sarvadurgatipariśodhanatantra in Tibet’, Studien zur Indologie und Iranistik 16: 109–125. van Schaik, Sam. 2015. Tibetan Zen: Discovering a Lost Tradition. Boston and London: Snow Lion. van Schaik, Sam and Lewis Doney. 2007. ‘The Prayer, the Priest, and the Tsenpo: An Early Buddhist Narrative from Dunhuang’, Journal of the International Association of Buddhist Studies 30/1–2: 175–217.

03_CreSouthV1_8P_3Feb22.indd 124

van Schaik, Sam and Imre Galambos. 2012. Manuscripts and Travellers: The Sino-Tibetan Documents of a Tenth-Century Buddhist Pilgrim. Berlin: De Gruyter. Verardi, Giovanni. 2018. The Gods and the Heretics: Crisis and Ruin of Indian Buddhism. New Delhi and Buenos Aires: Aditya Prakashan and Fundación Bodhiyāna. Vitali, Roberto. 1990. Early Temples of Central Tibet. London: Serindia Publications. von der Schulenburg, Stephan, Elke Hessel, Karsten Schmidt, and Matthias Wagner (eds.). 2015. Buddha: 108 Begegnungen/Encounters. Köln: Wienand Verlag. von Schroeder, Ulrich. 1990. Buddhist Sculptures of Sri Lanka. Hong Kong: Visual Dharma Publications. . 2001. Buddhist Sculptures in Tibet. Hong Kong: Visual Dharma Publications. Wade, Geoff. 2014. ‘Beyond the Southern Borders: Southeast Asia in Chinese Texts to the Ninth Century’, in John Guy (ed.), Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia. New York: Metropolitan Museum of Art. Wang, Eugene Y. 2005. Shaping the Lotus Sutra Buddhist Visual Culture in Medieval China. Seattle and London: University of Washington Press. Wang, Michelle C. 2018. Maṇḍalas in the Making: The Visual Culture of Esoteric Buddhism at Dunhuang. Leiden and Boston: Brill. Wangdu, Pasang and Hildegard Diemberger. 2000. dBa’ bzhed: The Royal Narrative Concerning the Bringing of the Buddha’s Doctrine to Tibet. Wien: Verlag der Österreichischen Akademie der Wissenschaften. Wayman, Alex and Ryujun Tajima. 1998. The Enlightenment of Vairocana. Delhi: Motilal Banarsidass Publishers. Weinberger, Steven N. 2003. The Significance of Yoga Tantra and the Compendium of Principles (Tattvasaṃgraha Tantra) within Tantric Buddhism in India and Tibet. PhD dissertation, University of Virginia. . 2010. ‘The Yoga Tantras and the Social Context of Their Transmission to Tibet’, ChungHwa Buddhist Journal 23: 131–166.

3/2/22 6:20 PM

In the Footsteps of Amoghavajra (705–774) Wessels-Mevissen, Corinna. 2001. The Gods of the Directions in Ancient India: Origin and Early Development in Art and Literature (Until c. 1000 A.D.). Monographien zur indischen Archäologie, Kunst und Philologie, Band 14. Berlin: Dietrich Reimer Verlag. Whitfield, Roderik. 1982–1986. Art of Central Asia: The Stein Collection in the British Museum. 3 vols. Tokyo: Kodansha International, in cooperation with the Trustees of the British Museum. Wong, Dorothy C. 1993. ‘A Reassessment of the Representation of Mt Wutai from Dunhuang Cave 61’, Archives of Asian Art 46: 27–52. . 2007a. ‘The Case of Amoghapāśa’, Journal of Inner Asian Art and Archeology 2: 151–158. . 2007b. ‘The Huayan/Kegon/Hwaŏm Paintings in East Asia’, in Imre Hamar (ed.), Reflecting Mirrors: Perspectives on Huayan Buddhism, pp. 337–384. Wiesbaden: Harrassowitz. . 2008. ‘Reassessing the Wall Paintings of Hōryūji’, in Dorothy C. Wong and Eric M. Field (eds.), Hōryūji Reconsidered, pp. 131–198. Newcastle: Cambridge Scholars. . 2012. ‘The Art of Avataṃsaka Buddhism at the Courts of Empress Wu and Emperor Shōmu/Empress Kōmyō’, in Imre Hamar and

03_CreSouthV1_8P_3Feb22.indd 125

125

Robert M. Gimello (eds.), Avataṃsaka Buddhism in East Asia: Huayan, Kegon, Flower Ornament Buddhism: Origins and Adaptation of a Visual Culture, pp. 223–260. Wiesbaden: Harrassowitz Verlag. . 2018. Buddhist Pilgrim-Monks as Agents of Cultural and Artistic Transmission: The International Buddhist Art Style in East Asia, ca. 645–770. Singapore: NUS Press. Wong, Dorothy C. and Eric M. Field. 2008. Hōryūji Reconsidered. Newcastle: Cambridge Scholars. Yang Qingfan, Lu Suwen, and Zhang Yanqing. 2017. ‘Archaeological Report on the Newly Discovered Tubo Dynasty Rock Sculptures in the Town of Sgar thog in Smar khams County, Tibet’, Journal of Tibetology 16: 233–251. Yang, Zeng. 2018. A Biographical Study on Bukong 不空 (aka. Amoghavajra, 705–774): Networks, Institutions, and Identities. PhD thesis, The University of British Columbia. Zhang Wenbin (ed.). 2000. A Centennial Commemoration of the Discovery of the Cave Library (Dunhuang: jinian Dunhuang zangjingdong faxian yibai zhounian). Beijing: Chaohua chubanshe.

3/2/22 6:20 PM

Chapter 4

Heruka-Maṇḍalas across Maritime Asia1 peter shar rock

H

introduction

erukas,2 fierce icons of the supreme Buddhas of the esoteric Vajrayāna (‘thunderbolt’) vehicle, emerged as key soteriological aids in the monasteries of northern India under the Pāla dynasty (8th–11th centuries). Recent archaeological research suggests that they achieved a second, quite different florescence as instruments of royal power in the 11th–15th centuries across Maritime Asia. Even as the Herukas headed towards extinction under armed Islam in the great monasteries of Bengal and Bihar in the

1. My thanks for incisive comments on the chapter go to Andrea Acri, Iain Sinclair, Swati Chemburkar, Olivier Cunin, Shivani Kapoor, and two peer reviewers. 2. Herukas are a class of fierce deities emitted into the saṃsāra by the Ādi-Buddha Vajrasattva or Akṣobhya, head of the vajra clan and Jina of the eastern universe. Their manifestations, appearing first in the Sarvabuddhasamāyogaḍākinījālaśaṃvara, include Cakrasaṃvara and the eight-headed, sixteen-armed Hevajra, who is the principal Heruka of Khmer Esoteric Buddhism found in the material record. See de Mallmann 1986: 182, and Dasgupta 1950: 98: ‘This Vajra-sattva, the Lord Supreme of the Tantric Buddhists, is found in the Buddhist Tantras bearing many other names of which the most important are Hevajra and Heruka.’ Alexis Sanderson (2009: 148, n. 340) adds: ‘The origin of the name Heruka has not been explained in a satisfactory manner. Indigenous sources explain it only through artificial semantic analyses based on superficial similarities of sound ... We may note that the name Hevajra, that of the second major deity of the Yoginītantras, appears to have a similar origin, having been conjured up from the Mantra HE VAJRA PAŚYA “O Vajra[-being], behold!” that is uttered when the blindfold is removed from the candidate’s eyes in the presence of the Maṇḍala (Sarvatathāgatatattvasaṃgraha, section 230).’

04_CreSouthV1_4P_22Jan22.indd 126

late 12th century, the flight of the Indian Buddhist masters fuelled a sudden growth and refinement of Buddhism in Nepal and Tibet, while their monastic vehicle entered the courts of the kings and emperors defending their domains in Maritime Asia. Indian Buddhist strategists had apparently long aspired to overlay a Heruka maṇḍala on Śaiva holy sites3 to implement forms of ‘state protection Buddhism’ that would rival the hold of Śaivism over the courts of the subcontinent. This aspiration was to materialize only outside India. This chapter brings together the patches of evidence scattered over the wide and varied geography of the southern seas to trace this royal rebirth of the Herukas in Maritime Asia. In so doing, it attempts to carry the baton forward for the final seaward leg of the history of fierce Buddhist deities that Rob Linrothe (1999) brilliantly traced across the subcontinent. He concluded with these notable 3. The Hevajratantra was widely disseminated and is dated to the late 8th century by Snellgrove (1959: vol. I: 14), to the late 9th century by Samuel (2008: 285), and to the 9th–10th centuries by Davidson (2005: 51). See below the list of the target holy sites in Hevajratantra I.vii 13–16. Griffiths (2014) recently found textual evidence for the presence of this text in Sumatra. In his close reading of mantras inscribed on gold foils found at Padang Lawas, he acknowledges a tip from Harunaga Isaacson about the insertion of the Sanskrit conjunction tadanu (‘after that’, Monier-Williams 1899: 434), which is not part of the mantra. Isaacson thought the insertion ‘... proves unequivocally that the mantra is being taken directly from this chapter of the [Hevajra-] Tantra, thus furnishing strong evidence that the text itself was transmitted and read in Padang Lawas’ (ibid.: 230). This appears likely from the fact that Atiśa and Dharmakīrti, both experts in this text, studied together for twelve years in Sumatra in the early 11th century; see below.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia remarks on the eight-headed, sixteen-armed Heruka ‘Hevajra’: If there was a Hevajra cult, it seems to have thrived in Southeast Asia. Hevajra imagery may have come overland to Thailand from eastern India, direct from Bengal via Burma. The transmission of Hevajra teachings was probably reinforced through contacts along the maritime routes from island Southeast Asia and eastern India. Whatever the exact route, it clearly had considerable influence in the highest levels of society. Seemingly removed from his yogic and monastic origins, Hevajra was utilized in the royal cult, not as in the Ming court, to improve relations with Tibet, but as part of an attempt at local legitimization ... It appears that despite the earlier origin of the texts and ideas behind the Hevajra imagery, they were not influential enough to generate a lasting impact until the 11th or 12th century. By that time, however, the ideas and images quickly flowed in eastern, southern and northern directions. Islam alone proved an impenetrable barrier. (ibid.: 274)

127

state definitively to Buddhism and related Buddhist politico-religious strategies, as Acri observes, thrived in their localized adaptations in Java and Sumatra until the 15th century.5 Such royal Heruka cults in the states of the southern seas have been identified for a century, but their importance to rulers and the ways they were deployed is only now emerging more clearly. Few texts have survived in the tropics so the historical data is much thinner than in China, Korea, and Japan,6 but some recent icon recoveries and close studies of temple ruins have, for example, brought some clarity to a 12th century dispute between the Khmer empire and neighbouring Campā, as well as to inter-state confrontations some decades later on the islands of Java and Sumatra. They furnish indications that the Buddhists in the south attempted to replicate the system of Vidyāpīṭha Śaivism, where Śaiva masters drew authority and supernatural powers to support kings from potent, networked holy pilgrimage sites. This is sketched as an aspirational strategy in two Yoginītantras, where Buddhist Bhairavas or Herukas are superimposed on the

I will propose that we can discern a long-term ec- holding cylinders, had previously been identified as the lectic strategy devised by Khmer Buddhists to bring supreme sixth Esoteric Buddha Vajradhara in an unconventional version of the crossed-arms vajrahuṃkāra-mudrā. the established state Śaivism under a dominant The hospitals, according to their inscribed dedication stelas parasol of Esoteric Buddhism—a model defined (Sharrock and Jacques 2017: 231) were supplied three times a in the Tantras devoted to the Herukas Hevajra year with precise amounts of herbs, camphor, resins, honey and Cakrasaṃvara who were to figure in Khmer and with small amounts of precious minerals from the iconography. For the people, the king’s Buddhist royal storehouse in the capital. In the Bayon state temple, drive was underpinned with a tantric Buddhist in face-tower sanctuary BY21 on the upper level beside the central sanctuary, the full medical triad is presented thus: triad headed by Bhaiṣajyaguru, master of remedies, ‘Lord Master of Healing, the [King of] Lapis Lazuli Radiwho presided over a remarkable and unprecedented ance; Lord Sunlight [or Solar Radiance]; Lord Moonlight empire-wide network of 102 hospitals that were (Lunar Radiance)’ (Maxwell 2007: 124). open to all.4 This strategy succeeded in turning the 5. Acri 2016: 8. 4. The Buddha presiding over the triad erected in the hospitals was only recently identified as the Khmer tantric Bhaiṣajyaguru by Hiram Woodward in 2009 at a Banteay Chhmar conference in Sisophon, Cambodia. Woodward had long sought a Khmer triad of icons that would have been inserted into pedestals pierced with three tenon holes in numerous ancient sites of Khorat hospitals, but only solved the mystery when he saw at a New York dealer’s a bronze Buddha and two Bodhisattvas in Bayon-style attire seated on thrones welded together (Woodward 2011: 497). The crowned Buddha, sitting with the ritual vajra and ghaṇṭā bell crossed at his chest and flanked by bodhisattvas

04_CreSouthV1_4P_22Jan22.indd 127

6. Here are Chinese and Japanese examples of the kinds of records that may have been lost in the disintegration of mediaeval temple libraries in topical zones: ‘In 767, Amoghavajra initiated the ordination of 37 monks for repeated performances of rituals on Mount Wutai to “establish the state as a field of merit” (T 2120.52 835 b17-c9). Significantly, this number refers to the 37 central deities of the vajradhātumaṇḍala’ (Lehnert 2012: 263). Conlan (2011: 107–114) describes how during the 1330s civil war in Japan the southern court deployed monks and relics to holy mountain sites considered as defining the Shingon Diamond and Womb maṇḍalas over the landscape of an ensuing battle. They attributed the outcome to the efficacy of the monk’s prayer rituals.

28/1/22 5:38 PM

Peter Sharrock

128

Śaiva Vidyāpīṭhas in maṇḍalic semi-circles descending from Mount Kailāsa (Fig. 4.3) in order to create supramundane fields of merit over Buddhist royal domains. The compilers of the Hevajratantra7 and the Saṃvarodayatantra include the maritime polity Suvarṇadvīpa (probably Sumatra) in the outer maṇḍalic circle (see below).

the hevajra-system and its royal dimension in maritime asia J.J. Boeles (1966: 26) analysed two Khmer Yoginī bronzes from Thailand and, noting Grünwedel’s (1900: 63) work on Kublai Khan’s Hevajra abhiṣeka, joined earlier scholars who pointed to signs of a significant royal role for the Hevajratantra beyond the subcontinent: The cult of Hevajra, though its most important rites were secret, was by no means confined to an obscure sect. It must at one time have been of great importance throughout the Far East. In 1261 the great Khublai Khan, the ruler of most of Asia and part of Europe, was consecrated as Hevajra by means of abhiṣeka rites exactly like those expounded in the Hevajrasekaprakriyā [ceremony manual; see Finot 1934].

Heruka icons played a role on both sides of a long dispute over Vijaya, a new Cam deep water port and new capital that the Cams and Khmers had collaborated in building so their economies would benefit from the burgeoning Song dynasty maritime trade. Cam prince Vidyānandana, who first led in a Khmer army to recover control of the port, soon broke with his patron in Angkor and proclaimed himself king of Campā. After defeating another Khmer force sent against him, he erected a Buddhist temple to Heruka in 1194 CE at the major Cam Śaiva mountain temple complex of Mỹ Sơn, before being eventually vanquished and forced into exile by a third Khmer army.8 Shortly after 7. Tanaka’s position underlines the significance of this text: ‘The Sakya school of Tibetan Buddhism ... is based on the Hevajratantra ... The Hevajratantra was one of the most popular tantras in late tantric Buddhism, and consequently many commentaries were written’ (Tanaka 2018: 224). 8. The Heruka was erected by prince Vidyānandana of Tumprauk-vijaya under his new regnal name of King Sūryavarman along with a large, dated, and unusually clear

04_CreSouthV1_4P_22Jan22.indd 128

his defeat, a large (3.6 m) stone icon of Hevajra, according to recent research, was erected in the Bayon, the first Buddhist state temple in Angkor, capital of the Khmer empire and, according to recent LiDAR remote sensing archaeology,9 probably the largest city on earth outside China. The icon is now thought to have been erected in the main eastern entrance to the state temple, facing towards Campā. Such findings lend support to an earlier bold though unsubstantiated claim by Bautze-Picron that ‘... Hevajra is the protector of the State in Cambodia and perhaps also Campā whereas it is Mahākāla who assumes this role in the Yuan Empire’ (Bautze-Picron 2014: 107). Some decades later the emperor Kublai Khan of China10 (then planning his eventually unsuccessful seaborne invasion of the southern seas) and East Javanese King Kṛtanagara underwent Heruka consecrations either ahead of campaigns of conquest or in response to the threat of invasion. A pattern thus emerges in inter-state confrontations in which supramundane Esoteric Buddhist maṇḍalas were invoked to underpin imperial strategies in the mundane sphere. In China, a century before Kublai, an 11th-century chronicle entitled Comprehensive Mirror for Aiding Government shows how ‘fundamental elements of Esoteric Buddhism were incorporated into the most rarefied strata of the Chinese imperial government’ (Goble 2016: 124): His Highness set up a maṇḍala (daochang 道 場) in the Triple Hall (sandian 三殿) taking the historical inscription (C. 92 B), now kept in the museum at the site. Schweyer summarized his story thus: ‘When he reneged on his relationship with Jayavarman VII, king Sūryavarman turned to emperor Long Cán of the neighbouring Đai Việt for acknowledgement of his legitimacy, which was granted in 1199. But the extraordinary military career of the Cam king was to end in disaster. Vietnamese records show that after Jayavarman sent yet another army to defeat him in 1203, Vidyānandana fled and requested asylum in the Đai Việt. He was rejected and disappeared without trace. His paternal uncle, called in the inscriptions Yuvaraja Oṅ Dhanapati-grama, became a governor under the Khmer authority from 1203 to ca 1220, when the Khmer army withdrew after Jayavarman’s death’ (Schweyer 2009: 327). See also Finot 1904: 168–169, Sharrock 2006: 183–186, Jacques 2015: 142–144. 9. Evans 2016. 10. Willemen 1983: 16.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia Palace Women (gongren 宮人) to act as Buddhas and Bodhisattvas and the North Gate military elite (beimen wushi 北門武士) to act as Vajra[pāṇis? -sattvas?] and divine kings. He summoned the Grand Ministers (dachen 大 臣) to prostrate to and circumambulate them. (Zizhi tongjian 資治通鑒 222)

This did not amount to a restructuring of the Chinese state but rather to the casting of a mantle of Esoteric Buddhist cosmology and potent ritual technology over the long-established state apparatus. I will propose that resurgent Khmer Buddhists of the 11th and 12th centuries cast a similar ritual mantle over the temples that had established the Śaiva state over four centuries and that this would contribute to a definitive Khmer turn to Buddhism. The steady absorption of Śaiva and Śākta texts into the Esoteric Buddhist Yoginītantras composed in the great Ganges valley monasteries under the Pāla dynasty is traced in Sanderson’s seminal 2009 study of ‘The Śaiva Age’. In often deliberately obscure language, the Buddhists taught not only accelerated attainment of Buddhahood but made ritual technologies for invoking supernatural powers available to kings in their affairs of state. He detects the transfers primarily from Śaiva texts already embedded in Brahmanical court rituals: Now, this co-existence of Buddhism and Śaivism under royal patronage was surely facilitated by the fact that the form of Buddhism adopted and developed was one that had equipped itself […] with a repertoire of Tantric ceremonies that paralleled that of the Śaivas and indeed had modelled itself upon it, offering initiation by introduction before a Maṇḍala in which the central deity of the cult and its retinue of divine emanations […] through the use of Mantras, Mudrās, visualization, and fire-sacrifice (homaḥ); and this was presented not only as a new and more powerful means of attaining Buddha-hood but also, as in the Śaiva case, as enabling the production of supernatural effects (siddhiḥ) such as the averting of danger (śāntiḥ), the harming of enemies (abhicāraḥ), and the control of the rain (varṣāpaṇam and ativṛṣṭidhāraṇam)… The latter is particularly important from the point

04_CreSouthV1_4P_22Jan22.indd 129

129

of view of Buddhism’s relations with its royal patrons, since such rituals enabled it to match the Śaivas by promising kings more tangible benefits than the mere accumulation of merit through the support of the Buddha, his teaching, and the Saṅgha. (Sanderson 2009: 124)

The supernatural powers embedded in maṇḍalas were deployed for leveraging intra-state legitimization after regime change as well as inter-state confrontations. The supramundane hierarchy captured in the maṇḍala proffered weaponized defence mechanisms for inter-state struggles, which found a more responsive audience outside the subcontinent and notably in the conflicts between the neighbouring Khmer and Cam domains in the 12th century, between China and Java in the 13th century, as well as between Java and Śrīvijaya in the 13th and 14th centuries. Although these royal invocations of supernatural power were based on metaphysics rather than physics, a modern reader may find some value in comparing them with the 20th century cold war nuclear standoff between the US and the Soviet Union. Bade (2016: 155) glosses the Kublai-Kṛtanagara confrontation as follows: As a technique for obtaining power and control over the world, Tantric Buddhism—and especially its rituals—came to occupy the same place in 13th and 14th century Asian societies that science and technology occupies in our own time. The advocates of scientific method, like the advocates of Tantric Buddhism, have claimed that this method alone leads to knowledge and that knowledge is power.

The power of new knowledge is touched on in the fifth opening line in the Chinese version of the Hevajratantra: ‘There is great power in this learning by which one is capable of various performances, viz. subjugating, petrifying, or driving away the army of others’.11 Though a far cry from what is known of early Buddhism and Śākyamuni’s 5th century BCE call in Magadha for withdrawal from normal family life into a saṅgha focused on meditation and study, both contexts shared a claim to breakthrough mechanisms to attain a new form of knowledge. The later Esoteric Buddhist state pro11. Willemen 1983: 35.

28/1/22 5:38 PM

130

Peter Sharrock

tection mechanisms may also seem to run counter to the spirit of emperor Aśoka’s standing down his army after the brutal 261 BCE battle for Kaliṅga, its earth ploughed with salt, and then governing the Mauryan state through personalized Buddhist edicts administered by civilian envoys.12 But Aśoka’s remarkable accomplishment barely survived his reign and the saṅgha leaders were soon faced with the return of conquering Buddhist kṣatriyas. The early propagation of the new maṇḍala knowledge beyond the subcontinent begins to be recorded in the 7th century in the journeys of learned and enterprising monks over the land Silk Routes. It expands in the 8th century over the maritime routes, from the monasteries and courts of south India and Sri Lanka, to Java, Śrīvijaya and then China (with disciples following on to Korea and Japan). The texts of the Yogatantras and the Womb-realm and Vajra-realm maṇḍalas they defined would soon be impacting the design of impressive ceremonial centres being built by a series of expanding rice empires in the south,13 and were captured in Chinese translations, conserved in state archives, and reflected in state rituals. Sanderson (2009: 125) points out that the ‘monarch is not mentioned in the treatments of initiation given in the Mahāvairocanābhisambodhi [MVS] and Sarvatathāgatatattvasaṃgraha [STTS], the two great Tantras that were translated 12. See Thapar 1961: 62. Kenneth Hall (1996: 101, n. 10) draws a parallel between the reign of Aśoka and that of East Javanese King Hayam Wuruk (r. 1350–1389 CE): ‘Asoka, who completed the Mauryan conquest of northern India in what Indian texts describe to have been extremely brutal asuryavijaya, “demonic victories”. The highlight of his aggressive actions was the dismemberment of Kalinga, which concluded with Asoka’s troops salting over that region’s soil. Thereafter Asoka became devoted to religion as the basis of his consolidation as the chakravartin, “universal monarch”. As Asoka asserts in his Thirteenth Rock Edict, he was now apologetic for his past actions of brutality, and now would pursue dharmavijaya, “righteous victory/moral victory”, and employed dharma ministers to integrate his newly conquered realm.’ 13. See in the second volume of this edited collection Hudaya Kandahjaya’s analysis of how the architects of Borobudur, when elaborating geometrical and arithmetical computations for the construction of the monument, were inspired by key elements within the Vajradhātu maṇḍala and the Garbhadhātu maṇḍala.

04_CreSouthV1_4P_22Jan22.indd 130

into Chinese in the early eighth century to form the basis of the Way of Mantras there and in the Japanese Shingon and Tendai sects.’ Yet Amoghavajra (Ch. Bukong), the renowned translator monk from Sogdiana, who perhaps met his preceptor Vajrabodhi (Ch. Jin’gangzhi) at a young age in Java,14 became deeply involved in imperial state protection through Esoteric ritual.15 Amoghavajra’s official biography shows him as being called on to resolve crises of the Tang emperors—bringing rain to end droughts, repelling a Tibetan army at the border, helping to restore the Tang after the An Lushan rebellion, and then consecrating emperor Suzong as cakravartin16 and holding extended faith propagation ceremonies for thousands of courtiers and army officers.17 Charles Orzech (1998: 138) reaches the conclusion that the Buddhists’ ‘teachings were shaped both by their particular religious ideologies and by imperial needs. Chinese Esoteric Buddhism in the eighth century was driven by political utility.’ 14. Abe 1999: 222–223; Sundberg and Giebel 2011: 30. 15. See Yury Khokhlov’s chapter in this volume on how Amoghavajra’s southern seas experience left a deep imprint on Tang Buddhism that then later flowed through the Hexi corridor and Dunhuang into Tibetan Buddhism. 16. Lehnert 2012: 262. 17. ‘In the third year of Ta-li [768], the master [Amoghavajra] held a ceremony at the [Ta-]hsing-shan monastery. He was given [by the emperor] 12 quilts of embroidered brocade and 32 embroidered gauze banners, the value of which was 10 million cash. In addition, he was given enough provisions to permit the crowd to remain at the ceremony for 14 days. The eunuch attendants, the ministers, and all the commanders of the imperial army were ordered by the emperor to receive abhiṣeka at the ceremony. Altogether more than 5,000 monks and laymen attended’ (Orlando 1981: 144). See also Orzech’s account of the Esoteric turn of the mid-Tang court in the eighth century in Orzech 1998: 135–146, and Sundberg and Giebel 2011. Hiram Woodward (2004: 341) adds the nuance that ‘it is likely that the Mantrayāna—specifically the texts quoted in the Sang Hyang Kamahāyānan Mantranaya—was known in Java in 792, primarily through connections with the Abhayagiri Monastery. The reason is less because of the wording of the Ratubaka inscription than because of the relationship of these texts to the thinking of Amoghavajra. In other words, following this line of analysis, it was not Amoghavajra’s passages through Southeast Asia that tie Javanese to Chinese traditions, but the connections of both Amoghavajra, who stayed there [in Sri Lanka] in the 740s, and central Javanese Buddhists to the Abhayagiri.’

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia With their entry at court level in Maritime Asia, the Buddhists’ teachings in cosmology, soteriology and ritual technology found expression in large royal ceremonial centres, the remnants of which form the bulk of the material record that has come down to us. In addition, the diaries of travelling Indian, Chinese, and Korean monks open rare, illuminating windows into how Buddhist courts and communities along the sea lanes were linked into an interactive network. By the 10th century, images took a tantric turn to represent the fierce protagonist deities of the later Esoteric Buddhist Yoginītantras, which had developed in Ganges Valley monasteries and in Bengal under the Pāla dynasty. By the time Sena kings replaced the Pālas in 1070 CE, resplendent Heruka deities dancing on māra demons had been erected in monasteries in southern Bengal, Nālandā, Ratnagiri (Odisha), and Sarnath. No known inscription however links these icons to royal patronage. The long infusion of tantric techniques into Buddhist scriptures was manifested in saṃsāra terms as fierce-eyed deities with fangs, skull chaplets, flourishing vajras and grasping khaṭvāṅga skull staffs, like those at Śubhapur (Comilla) and the funerary stūpa at Ratnagiri (Odisha).18 Sanderson sees the Sarvabuddhasamāyogaḍākinījālaśaṃvara as marking the beginning of the full appropriation of Vidyāpīṭha Śaivism into the Yoginītantras.19

18. See Linrothe 1999, figs. 180, 188. ‘The saṃsāra is Heruka’s phenomenal aspect, and he is the Lord, the saviour of the world […]. His eyes are red with compassion; his body is black to indicate friendliness; his four legs symbolize the four means of conversion, […] his eight face the eight releases and his sixteen arms the sixteen voids. The Five Buddhas are represented by the symbolic adornments, and he is wrathful for the subduing of the evil-disposed’. Hevajratantra 10–12 (Snellgrove 1959: 117). 19. The origin myth of the Vidyāpīṭha pilgrimage sites around India holds that Śiva took the body of his wife, the goddess Satī Dākṣāyaṇī, who entered a great sacrifice of the gods to which she and Śiva were not invited and committed suicide. Śiva entered and took her body on his shoulders and began a dance of rage that threatened to engulf the cosmos. So the gods caused Satī’s body to disintegrate and the pīṭhas marked the spots where pieces of the body fell to earth (see the entry ‘pīṭhā’, Encyclopaedia Britannica, https://www.britannica.com/technology/pitha, last accessed 17 February 2020).

04_CreSouthV1_4P_22Jan22.indd 131

131

First, it introduces or brings to the fore the cult of the deity Heruka with an iconography inspired by that of the Bhairavas of the Vidyāpīṭha with their accoutrements and attributes of the cremation-ground dwelling Kāpālika Śaiva ascetic. According to the visualization given by Ānandagarbha he has four faces and eight arms, emerging as the transformation of a dark blue flaming Vajra, itself a transformation of a dark blue syllable HRĪḤ. The central face is fierce (raudram), those to its right and left expressive of delusion and erotic passion, and that behind open-mouthed to devour. In his two uppermost hands he holds the freshly flayed skin of Bhairava over his back, in the two below a bow and arrows, in the third right in descent he shakes a blazing threepronged Vajra, and in the fourth a skull-bowl filled with human blood (mahāraktam). In the third left in descent he brandishes the Kāpālika’s skull-staff (khaṭvāṅgaḥ), topped with a three-pronged Vajra and adorned with bells, and in the fourth a skull-bowl filled with human flesh (mahāmāṃsam) (Sanderson 2009: 148).

We will see below that an icon close to the Ānandagarbha model appeared, perhaps in the 13th century, on the island of Sumatra in the southern seas. In 1935 Austrian archaeologist F.M. Schnitger, working in North Sumatra, followed up a report on a 118 cm high relief identified as a Heruka by Sanskritist F.D.K. Bosch (1930), who found it broken into fifteen pieces in the sanctuary of Bahal temple II at Padang Lawas (Fig. 4.1):20 This god, identified by Dr Bosch as Heruka, bears an Akṣobhya and five skulls in his flaming hair. The uplifted right-hand brandishes a wajra, the left hand holds the half of a skull in front of the breast and also clasps a staff trimmed with ribbons. Around the throat is a double necklace, across the breast a chain and on the shoulders hangs a long cord with about 20. See Natasha Reichle’s account of the discovery of the shattered icon photographed by Bosch, followed up by Schnitger in 1935, who found some further damaged fragments in the Biaro Bahal site museum, which were reported lost without trace by archaeologist Rumbi Mulia in 1976 (Reichle 2007: 133–137).

28/1/22 5:38 PM

Peter Sharrock

132

six heads […]. The image of Bahal strongly resembles the one in the Comilla in Bengal which now stands in the museum of Dacca. The latter, however, does not stand on a corpse but on a lotus cushion. On the upawita hang far more heads, and there are also other differences. (Schnitger 1937: 26)

that any initiate in the practice of this Tantra [Guhyasamāja] is not only familiar with the Śaiva scriptures but is able to enact their rituals by assuming the role of a Śaiva Guru […]. For it tells the adept of this tradition that in order to acquire the female consort required for his post-initiatory observance he should enter the home of a family of untouchables who are observant devotees of Śiva, reveal to them one of the Saiddhāntika scriptures […], give them Maṇḍala initiation [following this scripture], and then return to them the dakṣiṇā [payment for sacrifice] that they will give him, taking a girl from them in its place. (Sanderson 2009: 144)

khmer empire

Fig. 4.1: 1930 Oudheidkundige Dienst 10591. Photograph of shattered high relief identified by Bosch as Heruka from Biaro Bahal II, Padang Lawas, North Sumatra, 118 cm. (Photo courtesy of Leiden University Library)

The fierce emanations of the supreme Buddhas sharing features with tantric Śaiva-Śākta deities signalled the incorporation into the monastic setting of a range of intense, secret rituals, which included antinomian actions, taboo substances, and sexual yoga. Sanderson claims the first monks who specialized in the new rituals had already acquired expertise and practice as Śaiva initiands: [T]he Guhyasiddhi of Padmavajra, written in all probability in the eighth century, assumes

04_CreSouthV1_4P_22Jan22.indd 132

The Yogatantras reached the Khmer domains in the 10th century when the liberal King Rājendravarman II returned the capital to Yaśodharapura (Angkor) from Koh Ker. This Śaiva monarch permitted a Buddhist revival through his polymath minister-general-architect Kavīndrārimathana, who in c. 953 CE built the Bat Cuṃ shrine at the heart of the new city on the eastern edge of Angkor and dedicated it to the Buddha, Vajrapāṇi, and Divyadevī (Prajñāpāramitā). Thereafter Rājendravarman’s son and successor, Buddhist King Jayavarman V, financed a string of āśramas for the study of the root Yogatantra STTS and Śākyamitra’s learned 9th century ṭīkā (commentary). Jayavarman’s ācārya Kīrtipaṇḍita had them brought to Angkor after ‘having searched in a foreign kingdom’.21 Jayavarman’s Takeo state temple however was dedicated to Śiva, in line with the Brahmanical tradition from the founding of the state in 802 CE. Religious tolerance and eclecticism seem to have 21. See Tadeusz Skorupski’s translation of the Sanskrit inscription K. 111 of Wat Sithor of 980 CE, with references to the STTS and Śākyamitra’s ṭīkā, in Sharrock 2012: 205. The Khmer Buddhists had contact with the early Indian Esoteric ācārya Puṇyodaya in the 7th century but ceremonies for erecting Buddhist icons disappeared abruptly from the material record and the community was presumably suppressed. Yijing reported on his travels that he heard that Buddhism had entered the Khmer territory after Hinduism and had flourished alongside it until the monks were expelled or killed by ‘a wicked king’; see Takakusu 1896: 12. Nancy Dowling (2000: 129) concluded that this king was Jayavarman I (r. 657–681).

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia mostly come easily to the Khmers (see Sanderson 2003–2004: 233, 435). Striking among the icons of this time are supreme Buddhas enthroned on the coils of giant, multi-headed serpents or Nāgas (the Khmer Vairocana?). Beside him on a series of four-sided caityas is a ferocious Vajrapāṇi-Trailokyavijaya22 and a five-headed, ten-armed Prajñāpāramitā. They were found in a traditionally Buddhist area northwest of Angkor at Kbal Sre Yeay Yin and are now in the National Museum of Cambodia in Phnom Penh (NMCPP) and in the Musée Guimet in Paris. I have suggested (Sharrock 2009: 131) that this Vajrapāṇi reflects the widely diffused STTS narrative of the fierce Bodhisattva’s subjugation of the recalcitrant Maheśvara, who refused to enter the Buddhist 22. Lobo (1997: 73) saw this figure as a very early icon of Hevajra. I resisted in favour of seeing the five Vajradhātu Jinas seated in Trailokyavijaya’s mukuṭa as reflecting the STTS subjugation of Śiva and Vairocana’s warning to the still resistant god that ‘Vajrapāṇi is the overlord of all the Tathāgatas’ (Sharrock 2006: 24). For George Cœdès (1923: 43), the name Vajrin or Vajrapāṇi subsumed in Khmer epigraphy the later names Vajradhara and Vajrasattva: ‘Mais comme ces deux noms ne se rencontrent pas dans l’épigraphie cambodgienne et que nous avons vu d’autre part Vajrapāṇi invoqué au début des inscriptions de Bàt Čuṃ et vénéré dans un des trois sanctuaires, j’ai cru pouvoir réunir toutes ce statuettes sous la domination unique de Vajrapāṇi, dont Vajradhara et Vajrasattva ne sont d’ailleurs que des succédanés.’ Woodward (2015: 244) finds that the pose, the fierce krodha facial features, and the attributes suggest the Trailokyavijaya manifestation of Vajrapāṇi. He moves to a similar triadic identification to Cœdès when he perceives an intriguing local epigraphic link in noting that the fierce figure can be seen as upholding the five seated Jinas, in the way Vajrasattva is called ‘the upholder of all the Buddhas’ (ādhāraḥ sarvavuddhānāṃ) in the later Sāb Bak inscription (K. 1158 of 1060 CE). He concludes that the krodha is therefore best identified as Vajrasattva/Vajradhara/Vajrin—all Vajrapāṇi names used in the Guhyasamājatantra that is referred to in Sāb Bak, where Vajradhara (‘vajra holder’) is an alternative name for Vajrasattva (ibid.: 246). This may gain support from Tanaka’s recent work on the maṇḍala. The krodha’s attributes are close to those of the five Buddhas of the Guhyasamājatantra, who are, according to Tanaka (2018: 182), ‘... three-headed and six-armed, and they hold a vajra, a wheel, a jewel, a lotus, a sword, and a bell in their six hands. Among these attributes, the first five are the symbols of the five Buddhas while the bell is regarded as the symbol of Vajrasattva, the sixth Buddha.’

04_CreSouthV1_4P_22Jan22.indd 133

133

maṇḍala, provoking a confrontation in Vairocana’s celestial court between Śiva in Bhairava form and Vajrapāṇi as Trailokyavijaya.23 Before the battle, won with Vajrapāṇi’s vajra, Vairocana intervened with the warning ‘Vajrapāṇi is the overlord of all the Tathāgatas’, which seems to fit the fanged figure in the Kbal Sre Yeay Yin caitya with the Vajradhātu pentad of Jinas seated in his crown and mukuṭa of ascetic hair: His eyebrows tremble with rage, with a frowning face and protruding fangs; he has a great krodha appearance. He holds the vajra, aṅkuśahook, sharp sword, a pāśa-noose and other āyudha. (Dānapāla’s Chinese translation of STTS, T 18 882.370b, Linrothe 1999: 183)

After a seemingly vigorous Buddhist resurgence under Jayavarman V (r. 968–1001), King Sūryavarman I fought his way to the throne and re-established the orthodox Khmer Śaiva tradition in a long reign that lasted through the first half of the 11th century. The Buddhist revival was treated with respect.

phimai The closing decades of the 11th century saw religious and political regime change in the Khmer domains. The Esoteric Buddhist Mahīdharapura dynasty seized power, guided by the Śaiva strategist and rājaguru Śrī Divākarapaṇḍita (Fig. 4.2), who would crown three kings (two Buddhist, one Vaiṣṇava). This religious pivot brought a northern vassal ruler to the throne as King Jayavarman VI in 1080 CE.24 Within a century the dynasty would make Angkor one of the largest cities outside China. 23. Davidson sees the Śiva subjugation myth as defensive against the rising power of the Trimūrti: ‘Buddhist monasteries at this period had become enormous landed institutions that controlled great economic resources but had a tenuous relationship to the wider society, somewhat like the medieval Christian monasteries and modern universities ... Thus, at the socio-historical level, we should understand the Maheśvara myth in the Tattvasaṃgraha as a straight-forward defensive technique of the Buddhists to establish superiority of their gods over Maheśvara, Brahmā, Viṣṇu, etc., in an attempt to retrieve some of their lost position in unsophisticated circles in India, whether at Devīkoṭa, Vārāṇasī, Patna or wherever’ (Davidson 1991: 215). 24. See Briggs 1999 [1951]: 188.

28/1/22 5:38 PM

Peter Sharrock

134

established 108 Śaiva Vidyāpīṭha holy sites in India. He published a hand-drawn map entitled ‘Sites of Indian Buddhist Tantric Activity 10th–11th centuries’ in which the Heruka emitted by the supreme Buddha Vajrasattva/Vajradhara held the centre of the maṇḍala on Mount Kailāśa above three concentric and roughly semi-circular rings of pīṭhas with eight Bhairavas each (Fig. 4.3): Adapted from a story articulated in such yoga-tantra works as the Sarvatathāgatatattvasaṁgraha, Cakrasaṃvara proponents maintained that the eternal Buddha Vajradhara emanated a form, Heruka, to control Maheśvara (Śiva) and his twenty-four Bhairavas, along with their consorts. Maheśvara was eventually humiliated and destroyed, and Heruka took his place on top of Mount Sumeru, with the 24 Bhairavas controlling the 24 pilgrimage sites of India. (ibid.: 41)

Fig. 4.2: Presumed portrait of rāja-hotā Śrī Divākarapaṇḍita, architect of the Mahīdharapura dynasty, being carried by Śaiva ascetics in the military parade of Sūryavarman II, the third king he crowned, on the southern bas-relief of Angkor Wat. This was the only Khmer guru in history to be given the title Dhuli Jeng (‘dust of the feet’ [of the god]), normally reserved for kings. (Photo by author)

In Phimai, Divākarapaṇḍita must have been steeped in both Vidyāpīṭha Śaivism and the defensive Yoginītantra strategies developed in the monasteries of 10th and 11th century Pāla-Sena Bengal and Bihar and imported into the Buddhist revival in Tibet. Davidson (2005: 41–41) sees this in Tibet as a Buddhist counterstrategy that consisted in overlaying the Cakrasaṃvara maṇḍala on the

This Buddhist aspirational model for a new sacred geography for absorbing and eventually dominating the subcontinent’s long-implanted Śaiva holy pilgrimage sites was propagated through the widely-distributed Hevajratantra itself, which lists the 24 targeted sites in the maṇḍala (I.vii 13–16): The sites with asterisk are also named in the Saṁvarodayatantra (Tsuda 1974: 252). Interestingly, the pīṭhas targeted by the Tantric masters in both of these texts include, in the outer ring, Suvarṇadvīpa (‘Golden Isle’, usually taken to intend Sumatra or the adjacent Thai/Malay Peninsula). Because of its interest in sacred geography, some Indian masters of the Cakrasaṃvara were well aware that sacred sites existed outside India as well as inside it, and the list includes Suvarṇadvīpa, a medieval kingdom variously located in Burma [probably Suvarṇabhūmi, ‘land of gold’] or Indonesia. (Davidson 2005: 41)

Table 4.1 upakṣetras

chandohas

upacchandohas

Uḍḍiyāna*

pīṭhas

Mālava*

Munmuni

Kulatā*

Harikela

Kaliṅga

Pretapuri*

Jālandhara*

Sindhu*

Kāruṇyapāṭaka

Arbuda*

Lampāka

Suvarṇadvīpa*

Gṛhadevatā*

Paurṇagiri*

Nagara

Devīkoṭa*

Godāvarī*

Kāñcika*

Kokaṇa

Karmārapāṭaka

Himādri

Saurāṣṭra*

Kāmarūpa*

04_CreSouthV1_4P_22Jan22.indd 134

upapīṭhas

kṣetras

melāpakas

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

135

Fig. 4.3: Sites of Indian Buddhist tantric activity 10th–11th centuries: Davidson 2005: 39. (Map courtesy of Ronald Davidson and Columbia University Press)2525

25 Davidson (2005: 41) elaborates on the map as follows: ‘While the exact number of deities in individual maṇḍalas differed, each featured Cakrasaṃvara (as Heruka) in the center and three concentric rings of eight Bhairavas per ring progressing out from the center, representing the maṇḍalas of mind, speech and body.’

04_CreSouthV1_4P_22Jan22.indd 135

28/1/22 5:38 PM

Peter Sharrock

136

In the burgeoning Khmer empire of the late 11th century, Divākarapaṇḍita was the leading religious advisor (rājaguru) to a Buddhist usurper, who became King Jayavarman VI, the founder of the Mahīdharapura dynasty that would, within a century, take Angkor to its apogee.26 It seems possible that he was instrumental in the selection of such a Heruka-maṇḍala concept of alignment with established major religious centres as a means of asserting the new dynasty’s support for the long-embedded Śaiva holy sites of the kingdom, as well as legitimizing a new (predominantly Esoteric Buddhist) dynasty from a powerbase in the northwest of the empire. The Esoteric Buddhist Sāb Bak inscription (K. 1158 of 1060 CE), mentioned in footnote 22 above, does mention Buddha images being erected in a tīrtha. In the year 988, on the seventh day of the bright fortnight of the month of Tapasya, this learned Vraḥ Dhanus is the one who has installed the Sugata and other images in Tenpāsnaga, the fortunate and excellent pilgrimage place (tīrtha).2727

No other epigraphic account survives of such an extension of the mantle or seal of a new dynasty across the realm of Jayavarman VI, but Sanderson records, on the basis of inscription K. 194 (A9–B17), how the wealth bestowed on Divākarapaṇḍita for the coronation of Sūryavarman II was personally borne by the Guru to key religious sites in the form of gifts to Śiva from the new (Vaiṣṇava) king: 26 ‘Jayavarman seems to have been an ambitious prince and his ambition was apparently fired by a young brahman named Divākara whom he met early in his career. Taking advantage of the unsettled state of affairs, Jayavarman seems to have raised the standard of revolt and proclaimed himself king in the north…. The first appearance of Jayavarman VI as king is found in an inscription at the temple of Nom Van, near the present Korat, which Cœdès dated in 1082. This inscription, in Sanskrit and Khmer […] was a royal order of Jayavarman, who is called king, addressed to many high civil and religious dignitaries, directing them to supervise the monasteries (devāśrama) at Ratnapura (Nom Van) […]. Some of these notables will be mentioned in later inscriptions; so it appears Jayavarman had strong support from the beginning of his reign’ (Briggs 1999 [1951]: 178). 27 Verse 13 of an unpublished 2012 translation by Tadeusz Skorupski prepared for a SOAS research group.

04_CreSouthV1_4P_22Jan22.indd 136

Divākarapaṇḍita, after serving in lesser capacities under Udayādityavarman II (r. 1050–1066) and Harṣavarman III (r. 1066–1080), is said to have followed this practice as Vraḥ Guru under Jayavarman VI (r. 1080–1107), Dharaṇīndravarman I (r. 1107–1112) and Sūryavarman II (r. 1113–c. 1150). We are told that after he had performed the royal consecration of Sūryavarman II, given him Śaiva initiation, taught him the Śaiva scriptures and other branches of learning, and been invited to perform the Koṭihoma and other annual brahmanical sacrifices for him, he was given golden palanquins and many other valuables so that he could visit various sacred sites around the kingdom and give these to the deities installed there, each engraved with a verse composed by the king himself to the effect that it was a gift to Śiva made by his revered Guru. The sites chosen for this purpose were five, of which the first three are clearly the most important: Bhadreśvara [Wat Phu], Śikharīśvara (Preah Vihear), and Śivapura Danden (Phnom Sandak). At each of these he had a water-reservoir excavated, founded a hermitage, gave it slaves and villages and made an endowment to provide for worship. (Sanderson 2003–2004: 420)

The iconography of Phimai is truly noteworthy. The icon in the central sanctuary is lost and, according to Woodward (2003: 149), was ‘probably a Nāga-protected Buddha’. After paying tribute to the scale and artistic power of the central lintels carved in high relief, evidently produced by ‘the movement of the kingdom’s finest craftsmen to this eccentric location’ (ibid.), Woodward goes on to note that they have an innovative ‘stretched-out maṇḍala form’ and share an unusual and striking theme involving ‘conquering’: ‘The other four interior lintels have Buddhist subject matter and may each involve conquering. The outer southern lintel is a standing crowned Buddha ... the northern and eastern lintels, both themselves stretched-out maṇḍalas, have at their centers forms of Vajrasattva and Saṁvara, respectively.’28 The most striking image is a Khmer Heruka on the eastern lintel which appears to be an unortho28 Conti (2014: 384) has more recently argued for seeing a maṇḍala of Akṣobhya rather than Vajrasattva.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia dox, three- or four-headed, eight-armed martial representation of Cakrasaṃvara, bursting out of a bloodied elephant hide of illusion and trampling crowned māras on the elephant’s head in an ardhaparyaṅka (half-crossed legs) dance posture (Fig. 4.4). The Heruka’s dance posture and his attributes depart from texts that describe him, such as the Sādhanamālā, and are unique to a Khmer iconography that is repeated in small terracotta tablets (Conti 2014: 389). The elephant hide, which Cakrasaṃvara normally stretches out behind his back, but within which the deity here dances, is key to his identification on this relief. Snellgrove is one of many scholars who have pointed to the deity’s Śaiva parentage.29 In his 2007 translation of the ŚrīCakrasaṃvaranāmamahāyoginītantrarāja or Laghusaṃvara, David Gray (2007: 36) translates the name Cakrasaṃvara as ‘the binding of the wheels’ of kāyavākcitta (body, speech, mind). As the representation of a spiritual embrace of the entire human saṃsāra (‘wandering through phenomena’), this could have potential application in the political world, and perhaps in the manner achieved by Guru Divākarapaṇḍita. Davidson, looking at the decline of Buddhism in India and its hardwon revival in Tibet, casts the Buddhist maṇḍala take-over response more aggressively in line with Vajrapāṇi’s oft-repeated subjugation of Maheśvara in the STTS: The Cakrasaṃvara system represented a new attempt to bring a sense of place to the tantric table ... Adapted from a story articulated in such yoga-tantra works as the Sarvatathāgatatattvasaṃgraha, Cakrasaṃvara proponents maintained that the eternal Buddha Vajradhara emanated a Heruka form to control Maheśvara and his 24 Bhairavas and consorts. Maheśvara was eventually humiliated and de29 Snellgrove (2004 [1987]: 155) sees Saṃvara as essentially a Buddhist remodelling of Śiva: ‘Despite the symbolic interpretation in exclusively Buddhist terms, the origin of this divinity must surely be clear. The naked ascetic smeared in ashes with piled up matted hair, adorned with a lunar crescent, wearing skins of elephant and tiger, garlanded with skulls, holding trident, drum and khaṭvāṅga, all these attributes indicate Śiva as lord of yogins, the very one whom Vajrapāṇi is presumed in another context to have reduced to abject submission.’

04_CreSouthV1_4P_22Jan22.indd 137

137

stroyed and Heruka took his place on Mt Sumeru, with the 24 Bhairavas controlling the 24 pilgrimage sites in India. (Davidson 2005: 41)

It appears possible that Divakārapaṇḍita and Jayavarman devised a plan to extend a new sacred Buddhist geography across the empire. If this was so, their strategy at first looks to have been circumspect and incremental. They began in Phimai, a traditionally Buddhist region in the north western wing of the empire, rather than in Brahmanical Angkor and perhaps offered gifts to Śiva at major religious centres as later took place under Sūryavarman II, as noted above. Presumably their caution reflected trepidation at the response of the Śaiva landowning establishment. David Snellgrove made this comment on the 10th-century Khmer Buddhist revival: But Buddhism was clearly at a disadvantage, especially within the confines of the capital city of Angkor ... [T]he lineages of influential Brahmans, often related to the leading aristocratic families, formed an essential part of the structure of the state at least from the time of Jayavarman II onwards. (Snellgrove 2001: 54)

Jayavarman VI was crowned in Angkor but the only construction there attributable to him—in my view—is the walled Buddhist-Brahmanical temple of Banteay Samré, with unique Khmer tantric iconography, built on the eastern edge of the city.30 Angkor itself remained under the sway of Śiva in the grand and stately Baphuon pyramidal temple, set beside the palace and facing the wide royal plaza at the heart of Angkor. This was the work of the Śaiva King Sūryavarman I (r. 1005–1050) and his sons Udayādityavarman (r. 1150–1166 CE) and Harṣavarman III (r. 1166–1180 CE). It would take another century before another Esoteric Buddhist king—Jayavarman VII—would plant Heruka in the centre of Angkor within the first Buddhist state temple. 30 The claim that Banteay Samré is contemporaneous with Phimai is based on the almost identical construction of doorway, roofs, lintels and staircases—and towering central sanctuaries with elegant antefixes which give them the shape of giant pineapples or corncobs and were the precursors of the towers of Angkor Wat.

28/1/22 5:38 PM

138

Peter Sharrock

As noted earlier, religious tolerance was the norm in the ancient Khmer domains. It was possible for Divākarapaṇḍita, the most eminent Śaiva Brahmin of his time, to officiate over the coronation of two Buddhists and one Vaiṣṇava king, and to do this with the traditional 802 CE Śaiva cakravartin coronation ritual called devarāja. The guru lived long enough to crown the ardent Vaiṣṇava King Sūryavarman II (r. 1113–1149) after he seized the throne from his Buddhist uncle and from Harṣavarman III, at the age of sixteen. Sūryavarman constructed the enormous, world-celebrated and perfectly finished Angkor Wat temple, and led military campaigns by land and sea into the neighbouring Ðai Viet in alliance with kings of Campā. His intraregional expeditions and his alliance with the Cams would set a model for regional interactions for three centuries. He was succeeded by his much younger brother Tribhuvanādityavarman, an Esoteric Buddhist who built eight Buddhist shrines, according to inscription K. 1297, which was recently auctioned in Paris and preliminarily translated by the late Claude Jacques, and whose translation and study is forthcoming by a team of scholars from the EFEO in Paris.31 This king left many icons of crowned and bejewelled Buddhas enthroned on giant Nāgas carved in what French historians defined as the ‘Angkor Wat style’. No dedication stela is extant for the enormous temple complex of Prasat Hin Phimai, presumed to be Jayavarman VI’s off-centre powerbase, but a construction on this scale could only have been financed by the royal treasury. On the western side of Phimai’s central sanctuary, there is a Cakrasaṃvara lintel in full view of all who were allowed to enter (Fig. 4.4). In the lintel facing north, Pia Conti (2014: 386) has identified the central Buddha surrounded by four Jinas as Akṣobhya in the Esoteric form of Śrī Samāja, which he takes in the Guhyasamājatantra. Szántó and Griffiths (2015: 370) see this Tantra as bridging

31 The late Professor Claude Jacques of the École Pratique des Hautes Études (EPHE) lectured on his draft translation of the stone in April 2016 at the École française d’Extrême-Orient (EFEO) in the unpublished paper entitled ‘De la fragilité de notre science historique: la stèle K. 1297’ (Jacques 2016a). See Sharrock 2018.

04_CreSouthV1_4P_22Jan22.indd 138

Fig. 4.4: Cakrasaṃvara, bursting from the elephant hide of illusion and dancing on two māras, descends from a band of cosmic Buddhas accompanied by Yoginīs and Vidyārājas in the eastern lintel of the Phimai central sanctuary. (Photo courtesy of Pia Conti)

... the gap between the type of esoteric Buddhism that still operates within the realm of ritual purity and that of transgressive, antinomian esoteric revelation ... The Śaṃvara was instrumental in introducing significant Śaiva elements into Buddhism ... enacting the pantheon in communal worship, the imagery of the cremation ground.

Further, Sanderson (2009: 170–71) notes the iconographical similarities between Śiva and Heruka: The remnant features of Śiva’s iconography evident here are the trident, the third eye, the new moon on the piled up braids, the tiger-skin lower garment, the multiple faces and arms, the skull-bowl, the skull-staff, the bleeding elephant hide representing illusion burst asunder, the severed head of Brahmā, the snake as brah-

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia manical thread, the sharp fangs, the chaplet of skulls, his dwelling in the cremation grounds, and the ashes on his limbs. All these had entered Śiva’s iconography long before the formation of the Tantras of the Cakrasaṃvara cycle.

Cakrasaṃvara is unmistakable as he dances in ardhaparyaṅka posture on two crowned māras. Though dramatic, the Phimai image is still composed and constrained within the pervasively reserved Khmer tradition, which contrasts, for example, with this more typical and extravagant Indic text: Place in the center of the lotus the hero who is the terror of Mahābhairava, who is bright and brilliant, and who makes the tremendous noise of very loud laughter. Wearing a skull rosary, divine are his three eyes and four faces. Covered with an elephant hide, his excellent eyebrows are split by a vajra. His hand wields a khaṭvāṅga staff, [and he is] ornamented with a half a hundred garlands. (Cakrasaṃvara 2.14a–16b, in Gray 2006: 26)

The Khmer reserve excluded the yab-yum icons of sexual union with the deity’s consort or mudrā, as defined in the texts and routinely reproduced in Tibet. It presumably circumscribed Khmer ritual performance. Hiram Woodward (1981: 57) once memorably remarked, while describing a Hevajra in the Bangkok Museum, that in Khmer art ‘there need be no greater facial indication of Hevajra’s fierce aspect than open eyes.’ In the sculptural corpus of the Khmer empire, Cakrasaṃvara is found on small bronze tablet moulds opposite Hevajra but the Phimai icon is one of only two known in stone. This is in contrast with the about 100 elegant dancing Hevajras in bronze that have survived. Some of the most elegant may be stylistically dated by their lithe limbs and courtly dress to the ‘Baphuon style’ that includes Jayavarman VI’s reign. The Hevajra bronzes impressed and puzzled art historian Jean Boisselier: The [tantric] bronzes from the 12th and 13th centuries constitute a fairly considerable group but with no stone statue being reported for the same period, the importance of the role these divinities could have played in Khmer

04_CreSouthV1_4P_22Jan22.indd 139

139

Mahāyānist beliefs is strongly diminished. (Boisselier 1957: 324)

Boisselier presumably did not know that sixteen years before he wrote this, the EFEO had sold the broken bust of a large, eight-headed32 stone Hevajra to the New York Metropolitan Museum, mistaking it for a damaged eleven-headed Avalokiteśvara. Boisselier also had no idea that the stone Hevajra had probably dominated the main entrance to the Bayon for perhaps 100 years (see below). Curator Alan Priest (1937) in New York however immediately recognized the bust as Hevajra. The EFEO recorded a second stone ‘Hevajra’ outside the West Gate of Angkor Thom. The eight-headed, twenty-armed figure standing (not dancing) on four legs drew only a brief unsigned ‘Chronique’ note in the Bulletin de l’École française d’Extrême-Orient in 1925 on ‘other [finds] outside the West gate of Angkor Thom ... brought interesting sculptures to light, including one image of Hevajra’ (Chronique 1925: 592 [my translation from French]; see Pl. LXVI, B). The New York bust has now at last been reunited, electronically, with its dismembered parts, as we will see.

campā: mỹ sơn and dương long A second Heruka, probably identifiable as Cakrasaṃvara, was carved in stone more than a century after the Phimai lintel. If this identification is correct, it would imply that the Esoteric Buddhist maṇḍala was stretched to the empire’s eastern extremity in Vijaya, Campā. The icon is a high relief on a tympanum believed to have been erected on the high and imposing 40 m three brick towers of Dương Long temple, located near Vijaya port during the period of a renewed alliance with Angkor from 1203–c. 1220. This alliance began under King Sūryavarman II, who enlisted the support of Cam rulers each time this restless, campaigning king attacked the Đại Việt to the north. A Khmer prince, 32 The uppermost eighth head was missing. Decades later it was found outside the Gate of the Dead of Angkor Thom and deposited in the Angkor Conservation depot in Siem Reap, only to be reported stolen in a 1992 break-in and reported missing in an ICOM booklet ‘One Hundred Missing Objects: Looting in Angkor 1 Jan. 1993’.

28/1/22 5:38 PM

140

Peter Sharrock

who supported Sūryavarman’s mission in-country and who had a relatively weak claim to the Angkor throne, was later to seize power as King Jayavarman VII, the greatest builder in Angkor history. The long residence in Vijaya of the prince who would later be crowned Jayavarman and his engagement in Cam culture and high politics came at a formative period of his adult life. The inscription at Mỹ Sơn temple group G, dated 1157/8 CE, says that King Jaya Harivarman I was born in Vijaya (urāṅ ratna bhūmi Vijaya) and was respected throughout the Campā territories as far south as Pāṇḍuraṅga (Phan Rang) (Golzio 2004: 161–162, 166–168). When Jaya Harivarman later came into conflict with Khmer King Sūryavarman II and then with his successor Tribhuvanādityavarman, it seems likely that the future Jayavarman VII fought against them on the Cam side. This occurrence offers a possible explanation for why, when he became king, the inscriptions of Jayavarman’s reign showed little respect for his distinguished Esoteric Buddhist predecessor Tribhuvanādityavarman.33 The central Dương Long tower is 42 m high— the highest constructed in Campā and the same height as the Bayon in Angkor (Fig. 4.5). The scale of the construction was an early indicator of the coastal state’s growing wealth from the maritime trade route, which would take it to its apogee in the 15th century (Whitmore 2018). The Khmer empire in contrast would soon begin to decline, and the neighbouring Thais would occupy Angkor when the Cams were at their peak. King Jayavarman VII had waged a ten-year military campaign to control Vijaya and its deep-water port, which was apparently a joint project to profit from the growing Sung dynasty maritime trade. In the years from 1203 CE to Jayavarman’s death in c. 1220 CE, the close alignment of the Khmers with the Cams left a cultural imprint on the temple art and statuary in Campā. Jean Boisselier (1963: 256–297) designated this the ‘Tháp Mẫm style’ after a nearby temple left incomplete but with a rich collection of energetic, floriated statuary prepared for installation. This 33 In an inscription written by Jayavarman’s wife, Tribhuvanādityavarman is ungraciously called ‘a servant (bhṛtya) who killed his master (bhartṛ)’. K. 485, st. LXV and LXVI. Cf. Cœdès 1942 (IC II): 161; Sharrock 2018: 114.

04_CreSouthV1_4P_22Jan22.indd 140

Fig. 4.5: Dương Long temple group near Vijaya port and the Tháp Mẫm ruin. (Photo courtesy of Trần Kỳ Phương)

style includes the Dương Long icon recently tentatively identified as Cakrasaṃvara (Fig. 4.6). The 100 cm Dương Long tympanum shows a deity dancing in a Khmer style sampot with three heads visible—the faces have square jaws and wide smiling mouths without moustache and look decidedly more Khmer than Cam. It was discovered in the 1980s with other sculptures in an excavation by Polish archaeologists shortly after the end of the Vietnam war (Phương 2018: 249). They thought it represented the god Brahmā, as still reflected in the label in today’s elegantly refurbished Bình Ðịnh provincial museum (‘Thần Brahma: the god Brahma, late 12th/early 13th century, Dương Long tower’). After visiting the museum, epigrapher and historian Anne-Valérie Schweyer recently proposed it was a representation of Cakrasaṃvara: Most likely, the temple of Heruka could be in Dương Long, where there was a relief depicting Chakrasamvara—another form of Heruka. The Dương Long temple is very much influenced by Khmer iconography, and I think it is most likely the ‘Heruka’s house’. (Schweyer 2018: 77)

The identification is considerably less secure than that of the Phimai lintel as there is no clear elephant

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

141

Fig. 4.6: Cakrasaṃvara (?) from Dương Long temple, Campā, now in the Bình Ðịnh Museum, Vietnam. (Photo courtesy of Anne-Valérie Schweyer). Fig. 4.6.1: Phimai Cakrasaṃvara. (Photo by author)

hide behind the deity. There are traces of a stylized oval shape behind the heads and beside the principal right arm but the feet that would have danced on the elephant’s head, as at Phimai, are lost. There are though clear similarities in the ardhaparyaṅka dance position, the Khmer facial features, the Khmer sampot chong kben and jewellery and the delicate and unusual hand-touching mudrā of the two principal hands. In Phimai the deity holds no attributes and six rear hands hold apart the thick, upside down elephant hide. The rear hands of the Dương Long figure hold two lotuses and two swords. If it was a Cakrasaṃvara on the great Dương Long tower, in this Khmer-influenced period, this could perhaps mean Jayavarman VII was extending to completion the Mahīdharapura dynasty maṇḍala of sacred Heruka geography under the protection of Cakrasaṃvara in Phimai, Hevajra in Angkor and Cakrasaṃvara at the eastern wing of the empire at Dương Long in central Vietnam.

04_CreSouthV1_4P_22Jan22.indd 141

mỹ sơn and angkor Recent archaeological exploration of the Bayon temple in Angkor indicates that Heruka appeared there in a prominent exoteric position under King Jayavarman VII (r. 1182–c. 1220). Jayavarman’s ascent began with his seizure power with Cam allies34 in a power vacuum left after Cam King Jaya Indravarman IV of Grāmapura crossed the mountains in c. 1180 CE ‘in chariots’ to attack Angkor.35 Buddhist King Tribhuvanādityavarman (r. 1149–c. 1180) was killed and the palace burnt. Jayavarman, allied with Cam military princes from his long sojourn in Campā fighting with Jaya Indravarman’s predecessor, emerged from hiding in an unidentified Khmer location to drive the Cams out and take control of the city. His claim to the throne was weak 34 Vickery 2006: 110. 35 Cœdès 1942b: 389, K. 485 LXVIII, Phīmãnàkàs.

28/1/22 5:38 PM

142

Peter Sharrock

and he had adopted a low profile after racing back an unpublished paper on the inscription by Ian from Campā to try to help a relative who had earlier Lowman (2016: 1). risen against Tribhuvanādityavarman (Sharrock I praise the blessed Śrī Buddha Śrīvajrasattva, 2018: 116). Once in power, Jayavarman’s vigorous who is dual and non-dual (dvayādvayaṁ); the response to the unprecedented Cam attack was to good law which protects; and the community construct 8 m high ramparts stretching 3 km × 3 that is multiple and compassionate. (Jacques km around his new capital Angkor Thom. Outside 2016a: 1) this he built six large walled complexes centred on Buddhist temples that securely housed the popu- Vajrasattva is central to what Linrothe identifies as lation. The vast building programme was greater ‘Phase Three Esoteric Buddhism’ texts as the sixth, in volume than all his predecessors combined. At supreme, primordial ‘Ādibuddha’,37 and ‘comprises the heart of Angkor Thom, he erected a stone sky- all the Buddhas and who is ultimately identical scraper with giant smiling faces on towers looking with Hevajra himself ’.38 As an icon, seated with in all directions over the city. vajra held to the chest and ritual bell (ghaṇṭā) on Inscription K. 1297 indicates that the first his left hip, he first appears in the late 10th to early face-tower, considered the greatest architectural 11th centuries.39 innovation of Angkor, was actually built by TriAt a 2008 colloquium in SOAS on the Bayon, bhuvanādityavarman as Prasat Preah Stung in the Hiram Woodward observed a parallel between the major old Khmer Buddhist complex at Preah Khan stepped Greek cross base of the Bayon in the early of Kompong Svay, home of the Buddha of Vaṃśārā- phase 1.13 recorded by archaeo-architect Olivier ma/Kaṃmrateṅ Jagat Chpār Ransi.36 Jayavarman Cunin (Fig. 4.7) and the stepped base of a welladopted the model and made the towers the hall- known gilt bronze Hevajra icon in the National mark of the Bayon style, erecting a total of more Museum of Cambodia, Phnom Penh. In Cunin’s than 100 at the Bayon and at Banteay Chhmar in account, the Bayon was built first as a set of sancthe west. After studying the many earlier attempts tuaries grouped around a high central tower on a to identify the deity in the face-towers—Brahmā, raised base in the shape of a Greek cross. It was then Avalokiteśvara, neak ta local gods—I concluded a gradually extended over three decades outwards better fit with all the evidence would be Vajrasattva, and downwards towards its outer gallery at ground the supreme eternal deity of Esoteric Buddhism. The visual effect of the multiple pacific faces on 37 Monier-Williams 1899: 136–137 defines the epithet ādi the towers, and the sublime effect of their smile, as ‘first, original, primordial’. Linrothe (1999: 238) observes suspended in stone high above the city, impress that this concept ‘brought Phase Three Esoteric Buddhism many visitors as an attempt to express a cosmic perilously close to the notion of a supreme godhead ... a vision of the supramundane. In a Vajrayāna milieu monist or pantheist conception ... anathema to Buddhism’. in the 13th century, this could perhaps have had a 38 Linrothe (1999: 27). He follows Snellgrove 1988: 1366: ‘[A]ll these later tantras refer to a 6th supreme Buddha who natural affinity with the notion of a Vajradhara or unites the fivefold set within himself and is at the same Vajrasattva as the ultimate being and prime mover time identical with the particular tutelary divinity of any of the cosmos (Sharrock 2007: 268). particular tantra, Hevajra or any other.’ A decade after risking airing this hypothesis 39 The dozen Khmer Vajrasattva icons in this posture in in my doctoral thesis, I learnt with an encourag- museum collections are relatively early and range from the ing shock that Tribhuvanādityavarman, the first late 11th century ‘Baphuon style’ to the later ‘Angkor Wat builder of a face-tower, addressed what appears style’ of the mid-12th century. Estève and Vincent (2010: 151, n. 79) list five French scholars dating them to the 12th and to be his main inscription (K. 1297) to Vajrasat- 13th centuries. Their valuable study of a 25 cm Angkor Wat tva, according to a draft translation by Jacques and style Vajrasattva bronze palanquin end piece for carrying 36 K. 1297 st. 35: ‘[also a tower with] quadruple faces [looking] in the four directions, with serpents on the side’ (trans. Jacques 2016: 11).

04_CreSouthV1_4P_22Jan22.indd 142

a deity in procession (ga 5303, bearing inscription K. 943 in the Phnom Penh Museum), surprisingly assigns the piece to the 14th century because of a single letter in the inscription judged to show the arrival of Middle Khmer. Inscription K. 1297 would suggest the reign of Tribhuvanādityavarman.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia level in the centre of Angkor Thom city. Towards the end of this cycle, multiple giant faces as high as a person were carved as addorsed heads on the central stone tower that rises up 43 m over the city (Fig. 4.7.1). The bronze with a high, stepped base Woodward pointed to has been called a miniature sanctuary40 or palace because it has a framework of columns and capitals that fits over the deity and the circle of eight dancing Yoginīs in his maṇḍala (Fig. 4.7.2). Rather like the Bayon central sanctuary with its giant faces, the bronze deity with a tower of eight heads rises from the Greek cross base high above the Yoginīs dancing around him on a lotus platform (Fig. 4.7.3). The implication appeared to be that the bronze sanctuary may offer a key to unlocking the identity of the multiple deity heads in the Bayon tower. The question is then whether the Bayon is a massive rendering of the supreme Buddha Hevajra/Vajrasattva in architectural form. My impression was that Woodward raised this intriguing observation without committing to an identification. He has not, to my knowledge, gone on to develop or publish the observation of a Greek cross/tower of heads link. To do so would entail an adjustment from the position he took in 2004, when he saw the Phimai advance in Tantrism and the Yoginītantras as later reduced to a secondary role in Angkor, where the older Mahāyāna of the Kāraṇḍavyūhasūtra, centred on Avalokiteśvara, was again ‘held in highest esteem’. For much of the eleventh and twelfth centuries [the Kāraṇḍavyūhasūtra] teachings were somewhat eclipsed by the rise of followers of the Yoginī Tantras, associated with the temple of Phimai (in northeastern Thailand) and the emergence of a new dynasty. Then, during the reign of Jayavarman VII (1181–mid 1210s), the Kāraṇḍavyūha sūtra teachings were again held in highest esteem, with the legacy of Phimai Tantrism playing a secondary role. (Woodward 2004: 348)

Yet the question of whether the Bayon face-towers are a kind of blow-up that turns a state temple into a huge icon of the supreme deity has been broached by other scholars. For example, Bruno 40 Dalsheimer 2001: 268.

04_CreSouthV1_4P_22Jan22.indd 143

143

Dagens (2003: 243) considered that ‘it is the invention of the face towers that makes architecture slide towards monumental statuary and temples towards divine image.’ T.S. Maxwell (2007: 86) went further, saying the face-towers turned the temple into a god: In terms of architectural theory, the new method of depicting the four faces on the towers of Jayavarman’s temples appears to represent ... a sweeping liberalization of architecture and iconography, because this method abandons the canonical distinctions between vāstu and śilpa ... The Khmer temple was therefore no longer merely a devagṛha, the traditional Indian ‘house of God’, it became itself a God.

Merging śilpaśāstra (statuary) and vastuśāstra (architecture) is unquestionably a major architectural innovation and many scholars have remarked on the originality of projecting supramundane faces high over the capital as the supreme architectural innovation of the classical Angkor period. Further, all entrances to the Bayon were enwrapped and embellished by emphatic friezes containing a total of approximately 6,250 celestial dancers in ardhaparyaṅka posture, which I have suggested are projectors of a Yoginītantra cult.41 Rendering the state deity and maṇḍala in architecture was not confined to the capital. Cunin (2005: 83–85) increased the number of face-towers at the Bayon by eleven to a total to fifty-nine by tracking all stones bearing parts of carved faces from the collapsed towers inside and outside the Bayon outer gallery. He went on to record forty-seven face-towers in Banteay Chhmar, a large temple complex 110 km northwest of the capital in a thinly populated zone under the Dangrek mountains.42 From the 41 Sharrock 2006: 108; Sharrock 2013: 124. 42 Banteay Chhmar’s proximity to the Dangrek chain probably gave it special role in the new hospital network as the medicinal herbs, minerals (including gold, mercury and camphor) and animal parts were gathered there. Four ‘satellite’ sanctuaries outside the main temple complex are built on the standard plan for hospitals and Banteay Chhmar may have been a centre for amassing, blessing and distribution to the large network, which included 30 hospital sites identified over the mountains in Khorat (modern Thailand). Furthermore, the largest 80 cm stone statue of the Khmer Bhaiṣajyaguru (Phnom Penh Museum ga 1714) was found in the temple.

28/1/22 5:38 PM

Peter Sharrock

144

Fig. 4.7: 3D archaeological reconstruction of the first architectural project of the Bayon temple. (Courtesy of Olivier Cunin)

Fig 4.7.1: Bayon central sanctuary. (Photo by author)

04_CreSouthV1_4P_22Jan22.indd 144

Fig 4.7.2: Gilt-bronze Hevajra maṇḍala in a palace structure on a stepped Greek cross base. (Photo by author)

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

145

Fig 4.7.3: Dancing Hevajra, without palace, encircled by an incomplete set of his eight Yoginīs. (Photo courtesy of Emma C. Bunker)

Fig 4.7.4: Part of the frieze of dancers that enwraps all the entrances to the Bayon. (Photo by author)

04_CreSouthV1_4P_22Jan22.indd 145

28/1/22 5:38 PM

146

Peter Sharrock

Fig. 4.8: Map of Cambodia with locations of the Angkor site, Preah Khan of Kompong Svay and Banteay Chhmar. (Map courtesy of Olivier Cunin)

04_CreSouthV1_4P_22Jan22.indd 146

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

147

Fig. 4.9: Banteay Chhmar Northern Gallery. King Jayavarman VII appears to consecrate prince Indrakumāra as king of Vijaya before the prince sets off to occupy the throne in Campā. (Photo by author)

viewpoint of sacred geography, the temples with the face-towers fit into a circle centred on Angkor and possibly define a new Buddhist maṇḍala of state (Fig. 4.8) that implants the face-tower deity across the kingdom, somewhat in the way Amoghavajra created the state as a field of merit on Mount Wutai in 767 (see footnote 6 above). If the face-tower sites are in this way projecting a Heruka maṇḍala over the Brahmanical Khmer state, Jayavarman can be seen as maintaining the Divākarapaṇḍita Mahīdharapura strategy mentioned above, rather than going back to an old Avalokiteśvara model. (The Bhaiṣajyaguru maṇḍala of hospitals would form another Angkor-centred circle twice as large). Davidson’s map, extracted from the Hevajratantra and Saṁvarodayatantra, places the principal

04_CreSouthV1_4P_22Jan22.indd 147

Heruka on Mount Kailāsa, which raises the question of where did Jayavarman locate his. The first Heruka in the Khmer-Cam epigraphic record suggests a challenge to this strategy, as it was raised in neighbouring Campā. Having crushed domestic resistance and fortified the capital, Jayavarman turned to restoring Campā, where he spent his formative years, to his empire. He assembled an army at Banteay Chhmar, at a safe distance from the perhaps still resistant Śaiva capital and launched its first expedition in 1190 under the command of the gifted Cam prince Vidyānandana, who had already put down the biggest recorded domestic rebellion against Jayavarman at Malyang in the western Khmer region. There is a rare detailed account of the Campā campaigns from two inscriptions— one in Old Khmer language in Banteay Chhmar

28/1/22 5:38 PM

148

Peter Sharrock

(K. 227)43 and one in Old Cam language in Mỹ Sơn. The Cam version of events in inscription C. 92B44 says Vidyānandana led the Khmer army over the mountains and defeated the Cam defenders of the port of Vijaya. He then invited Khmer prince Indrakumāra to take the Vijaya throne. Indrakumāra’s consecration in preparation for departing to Vijaya may possibly be shown in a low relief on the Northern Gallery of Banteay Chhmar (Fig. 4.9). There is as usual no inscription, but in the relief panel, the enlarged image of the king sits in ceremonial dress with the sword of state while courtiers below him raise a chalice, perhaps holding water for a royal abhiṣeka. The smaller figure kneeling before him has the facial features and distinctive necklace of the royal figure seen beside the king on the temple’s Eastern Gallery, who is usually taken to be Indrakumāra. Behind him sit three bearded Brahmins wearing high ritual headgear. Below them the cavalry looks ready to depart. When Indrakumāra arrived to take the Vijaya throne, Vidyānandana headed south to rule over Pāṇḍuraṅga as vassal of the new king in Vijaya. But the Cams soon rebelled against Indrakumāra, who fled back to Banteay Chhmar. With Indrakumāra gone, Vidyānandana marched on Vijaya to restore order and then turned against his patron in Angkor and declared himself King Sūryavarman of Vijaya and Pāṇḍuraṅga.45 Two years later in 1194 a second Khmer army reached Vijaya and was defeated by Vidyānandana, who had already shown himself to be a brilliant general and an erudite ruler. Inscription C. 92B says that in celebration he built ‘Lord Heruka’s mansion’ (paṅap rumaḥ śrī herukaharmya) after his victory at Jairamya-vijaya.46 The Heruka icon is lost (destroyed?) but the 3 m × 3 m brick base of his mansion survives in Mỹ Sơn along with the high and elegant C. 92B historical stela that stood beside it. His Heruka jinābhiṣekaḥ reaffirmed his status after two years on the throne but would also have transmitted a message of defiance and perhaps respect to his former patron in Angkor. 43 Cœdès 1929: 309–315; Jacques 2015: 141–144. 44 Finot 1904: 972; Jacques 1995: 154; Schweyer 2009: 327–328. 45 C. 92B, line 8 (Finot 1904: 972). 46 Schweyer 2018: 76.

04_CreSouthV1_4P_22Jan22.indd 148

I asked Ācārya Malcolm Smith at a conference on Tantra organized at the Victoria and Albert Museum (London) in 2018 what such an event could mean in Buddhist terms. He said similar confrontations involving a mix of sacred consecration and political expediency are known in Tibetan history, when the message was that both parties had attained participation in the supreme Heruka/Vajrasattva state of being that was beyond violence, so this should be reflected in the microcosm between the earthly rivals. War often followed, however. Jayavarman’s response was to send a third Khmer army nine years later over the mountains to Vijaya, which defeated Vidyānandana in 1203 and drove him into exile.47

heruka in the bayon With Campā again closely aligned with Angkor, Jayavarman probably raised his own Heruka in sanctuary BY55 at the main eastern entrance to the Bayon state temple, according to experts now working on the monument. It could have been the ceremonial highpoint of Jayavarman’s indrābhiṣeka or second coronation, the Indic consecration that confers cakravartin status following a foreign acquisition.48 Reconstructing the great Heruka of the Bayon that countered that in Vidyanāndana’s Mỹ Sơn mansion has taken almost a century. Some of the difficulty of uncovering the story of the statue, which must have been known to all Angkorians in the 13th century, lies in the absence of a Bayon dedication stela, which may have identified the original pantheon of the builder. Plus, the large stone Hevajra was dismembered and some of the parts dumped in the forest outside the city.49 The bust was 47 Vidyānandana fled Vijaya with his court and generals in a fleet. He sought refuge in the Ðai Việt but this was refused so his fleet took to the ocean and disappeared from history. 48 A mason’s informal inscription in Khmer on the unfinished Northern Gallery of the Bayon reads ... śānti buvana bphon hon stac vanodesa nā stac thve vraḥ indrābhiṣeka (‘... the peace of the whole universe. Then the king retires into the forest at the time when he celebrates the holy indrābhiṣeka’) beneath a sketched relief of the king going into the mountains on an elephant (Cœdès 1932: 74). 49 Olivier Cunin presents a detailed and truly remarkable

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

149

Figs. 4.10–4.13: Śivaliṅgas in situ in the Bayon today and Bodhisattvas and Buddha reliefs carved into liṅgas—the last perhaps hatched out by later Theravādin users. (Photos by author)

sold to the New York Metropolitan museum in 1937, the legs, arms and broken hands were left in the forest. The feet were finally recovered seventy-five years later from a pond in front of the temple and identified three years later by Olivier Cunin. I have attempted tying the removal and dismemberment of such major Buddhist icons from the Bayon to Sanskrit inscription K. 470. This inscription is dated 1327 CE by Cœdès and was found near the Bayon northeast corner gallery.50 (Pāli had replaced Sanskrit two decades earlier in royal epigraphy). This records the erection of a Śivaliṅga, presumably in the Bayon, in the first year of the reign of King Jayavarma-Parameśvara. There are still five Śivaliṅgas standing in situ in the Bayon today and many Bodhisattva and Buddha reliefs and high reliefs can still be seen re-carved into Śivaliṅgas (Figs 3.10–3.13). A second inscription in Sanskrit, the last recorded in a declining Angkor, is called ‘the Angkor Wat inscription’51 to a Śiva sanctuary on the Kapilapura mound just northeast of computer reconstruction of the erection and dismemberment of the Hevajra and a large stone Avalokiteśvara at the eastern entrance to the Bayon in a lecture entitled ‘About some missing images of the Bayon temple’ given in May 2018 at the ‘Exploring Angkor Symposium’ at the Asian Civilisations Museum (ACM), Singapore [https://www. academia.edu/41385160/About_some_missing_images_of_ the_Bayon_temple; slides 96–98 illustrate the breaking up of the Hevajra]. 50 Cœdès 1932: 145. 51 Cœdès 1968: 228.

04_CreSouthV1_4P_22Jan22.indd 149

the Angkor Wat moat. Briggs attributes it to 1330?52 There is a small ruined temple foundation there with a Śivaliṅga. The facts are few but if the return to Sanskrit and Śivapūjā, the erection of Śivaliṅgas in the Bayon and the refashioning of Bodhisattva and Buddha reliefs into Śivaliṅgas inside the Bayon all came together, a case can be made for the great Buddhist temple at the centre of Angkor Thom being refitted for Śaiva ritual by a Śaiva king lacking the funds to build his own dynastic shrine in the long tradition of his predecessors. William Southworth, in an important article that deserves greater attention, traces a century of attempts to account for the clear and widespread signs of desecration of Buddhist art in the Angkor temples (Southworth 2014a). In his conclusion he adds an element of caution while endorsing my own contribution to this enquiry (personal com52 Briggs 1948: 12. He adds: ‘A Sanskrit stele inscription—the last Sanskrit inscription of ancient Cambodia which has come to light—was discovered at Angkor Wat and was translated into French and published by Abel Bergaigne in 1885, but it has never been given much attention by historians. This inscription records the foundation by King Jayavarma paramesvara of a hermitage to Siva. No date is given, but from the context it seems that this king had then been ruling a few years. So, it may be said with certainty that this king was on the throne until after 1327, perhaps as late as 1330. This foundation and his name, however, seem to indicate that he was Sivaite’ (Briggs 1948: 22). Cœdès characterized the inscription as ‘redolent with Śivaite mysticism’ (Cœdès 1968: 228).

28/1/22 5:38 PM

Peter Sharrock

150

munication, September 2020). I had written (2007: 233): In 1327, in his first year on the throne, Jayavarma Parameśvara erected a Śivaliṅga in the (I would ask still Buddhist?) Bayon. Installing a Śivaliṅga in the Bayon immediately upon enthronement looks like a politico-religious act of some significance, so did Jayavarma Parameśvara convert the whole Bayon for Śaiva ritual and then order the systematic desecration of Jayavarman VII’s other temples in Angkor? The question cannot be answered with the data we have, but the question should be raised. Jayavarma Parameśvara’s 27-line record of his Bayon Śivapūjā, may quietly announce the arrival of the desecrator of the Buddhist Bayon. On the evidence we have, there is no stronger candidate.

Michael Vickery endorsed this analysis and emphasized the expression ‘a politico-religious act’ (Vickery 2006: 171). A political message does indeed seem to be reflected in the way many thousands of Buddha figures were chipped out of Jayavarman’s Buddhist temples and their long enclosure walls across the city, which contained compounds for large sections of the population. Further, a Japanese team53 uncovered a pit outside Jayavarman’s Banteay Kdei temple containing 274 Buddhist icons many of which, as Southworth says, ‘had clearly been deliberately decapitated and buried in the sculptural equivalent of a mass grave’ (Southworth 2014a: 202). Inside the temple there are five in situ Śivaliṅgas erected in outer and central inner galleries and there are many Buddhist reliefs crudely re-fashioned into liṅgas and set on yonis in central sections of the temple. The re-chiselling of multiple Bodhisattvas and Buddhas into liṅgas, along with the erection of stone Śivaliṅgas across the first, albeit tolerant, Khmer Buddhist state temple in Angkor, still suggest to me a later re-furbishing of the monument for Śaiva ritual. K. 470, according to Cœdès, firmly dates the erection of one of the liṅgas to Jayavarma Parameśvara’s first year on the throne. His reign left the last epigraphic traces of Śaivism and the last use 53 Ishizawa and Marui 2002: 206–208.

04_CreSouthV1_4P_22Jan22.indd 150

of Sanskrit in Angkor, after probably 180 years of Buddhist kings on the throne. And the declining economy would have favoured refurbishment over new construction. The earlier Khmer withdrawal from the Chao Phraya basin in the west and from Campā in the east, along with the much smaller scale of the 13th-century Theravāda temples of Preah Pallilay and Preah Pithu constructed near the palace, are among many indicators of a declining empire when Jayavarma Parameśvara came to power. Lacking the resource to maintain the long tradition of a new king constructing temples to his ancestors and to the state may have obliged him to reconfigure the grandiose Bayon at the centre of his capital. This hypothesis has some coherence, though it rests on limited evidence. Until more evidence emerges this king remains, in the view of the present author, the strongest candidate for what Southworth calls ‘The Alteration and Destruction of Buddhist Images at Angkor’. The French archaeological archive holds photographs of most of the dismembered parts of the Heruka statue excavated after workers uncovered them by accident in 1925 under a pile of earth in the forest beyond Angkor Thom, some 500 m outside the eastern Gate of the Dead in the king’s Angkor Thom city ramparts and close to the Siem Reap River.54 Most were taken to the Depot of the Conservation d’Angkor and photographed. With the arms broken off and the feet missing, the statue was at first identified simply as ‘a giant’55 and 54 Bruno Dagens gave a brief report on the site in the forest in 1969, when he assumed the presence of a pedestal and suggested it was a destroyed ‘chapel’. Subsequent excavation by APSARA in 2009 showed this not to be the case as there was no sign of a building and the square stone pedestal was also assumed to have been removed from the Bayon. Dagens wrote: ‘In Jayavarman VII’s time, this dancing god [Hevajra] of the Mahāyānist pantheon enjoyed a certain popularity that is attested, if not in the epigraphy, at least by numerous representations in bronze and also in a large statue in stone: this latter, sadly very mutilated, represents the god dancing as in the bronze images. It was found in a chapel to the east of the Gate of the Dead of Angkor Thom, where several statues of Lokeśvara were also found’ (Dagens 1969: 143). 55 Journal de Fouilles, 8 March 1925 (EFEO Centre, Siem Reap): ‘Marchal excavated Avalokiteśvaras, the bust of “a giant”, his huge legs “in squat position”, a 1m pedestal

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia the 1.3 m bust (Fig. 4.14) eventually sold to the NY MET as an eleven-headed Avalokiteśvara with the uppermost head missing. After studying a photocopy of the hand-written French Journal de Fouilles in the EFEO office in Siem Reap, I scrambled through the undergrowth and found the earth mound and the legs in 2009 (Fig. 4.15). The conservation authority APSARA rapidly excavated the site and moved the legs to the Norodom Sihanouk Museum in Siem Reap. If the Hevajra was erected in the Bayon, where did it stand? One possibility is Bayon sanctuary BY52, which is considered a late addition to the Bayon because it has a circle of Garuḍas below the lotus on its summit, an innovation from Banteay Chhmar brought down to Angkor. Archaeo-architect Olivier Cunin measured BY52, which has large wooden doors that open inward, and concluded it was too small for the 3.6 m stone Hevajra (personal email, December 2009). JASA (Japan-Apsara Safeguarding Angkor) expert Robert McCarthy later proposed (personal email) that the statue could have been erected in the larger adjacent sanctuary BY55. McCarthy photographed an oval trace in the sandstone floor of BY55, which could have been caused by a heavy rounded statue pedestal (Fig. 4.16). In 2013 the JASA group, working for many years to stabilize the Bayon, uncovered a large but obscure statue base with a tenon in a pond outside the main eastern entrance. Olivier Cunin, whose prodigious knowledge and database of many thousands of Angkor stones is unmatched, came across the block three years later and identified it as the Hevajra’s four feet poised in a dance posture.56 Cunin conducted a detailed photographic study of BY52 and BY55 and also settled on the latter as the site of the Hevajra. The uppermost head of the Hevajra was reported stolen in a 1992 break-in at the Siem Reap conservation depot (see footnote 32 above). carved from a single block, and 16 hands’ (my translation from French). 56 The broken off feet can be seen in slides 18, 21–23 in Cunin’s 2018 lecture in Singapore, https://www.academia.edu/41385160/About_some_missing_images_of_the_ Bayon_temple.

04_CreSouthV1_4P_22Jan22.indd 151

151

Fig. 4.14: Broken bust of Hevajra in NY MET Khmer Gallery. (Photo by author)

Fig. 4.15: Heruka legs the author located in the forest outside Angkor Thom. (Photo by author)

28/1/22 5:38 PM

Peter Sharrock

152

Fig. 4.16: Bayon sanctuary BY55 with circular trace of a pedestal base in the sandstone. (Photo by author)

04_CreSouthV1_4P_22Jan22.indd 152

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

153

Fig. 4.17: Simulation of the archaeological reconstruction of the Hevajra (DCA.751) and Lokeśvara (Ka.1695) statues into the towers 55 and 52 of the Bayon temple presented at the Asian Civilisations Museum, Singapore. The image combines Cunin’s computer restitution of the icon with a 3D scan by photogrammetry of the floors of the towers. (Courtesy of Olivier Cunin)5757

With the scattered pieces mostly reunited, Cunin or pledge ceremony with water would be required. produced a 3D computer simulation of the whole This could be the explanation long sought by art statue erected in BY55, with a large Avalokiteśvara historians for why so many bronze conch-holders (also excavated from the mound of earth in the embossed with a dancing Hevajra, sometimes in a forest, and now in the National Museum) stationed circle of eight Yoginīs, were discovered in Angkor behind it in BY52 (Fig. 4.17). and reached museum collections around the world. This huge stone Heruka erected in a prominent At least a dozen are in the material record and they exoteric position in the state temple faced east would have more than supplied the samaya needs across the mountains to Mỹ Sơn and Vijaya, where of all the temples Jayavarman built in Angkor (Figs. Vidyānanda had raised his Heruka. It also present- 4.18–4.20). Such a water consecration is described ed difficulties for anyone entering the temple for in a Tibetan manual Abhiṣekavidhi written by a the long annual calendar of state festivals. Ācārya scholar named as Prajñāśrī and translated by SnellSmith took the view that only people initiated grove (1987: 254). The initiand asks: into the Heruka would be allowed to pass through Just as Vajrasattva has consecrated former conBY55 into the temple. He suggested some samaya 57 Cunin presented his findings in a lecture ‘About some missing images of the Bayon temple’ in May 2018 at the ‘Exploring Angkor Symposium’ at the Asian Civilisations Museum (ACM), Singapore.

04_CreSouthV1_4P_22Jan22.indd 153

querors, do thou, well qualified, feel love for me and bestow the Vidyā consecrations!    Then the master collects water from the Jar of Victory and the other jars, pouring it into a scoop made of shell. He directs his thoughts towards Akṣobhya, worships and praises him and then envisions him dissolved in light ...

28/1/22 5:38 PM

Peter Sharrock

154

Figs. 4.18–4.20: Bronze lustration conch-holders cast with Hevajra’s maṇḍala in the National Museum of Cambodia, Phnom Penh. (Photos by author)

Thus he envisages it, as he takes the water in the scoop and bestows the Water Consecration, reciting the mantra: OṂ Vajra-Jar consecrate HUṂ!

tion on the base of the so-called ‘Joko Dolog’ statue of the Jina Akṣobhya, now in a park in Surabaya, commemorates both the reunification of Java by his father and his own second consecration of this image of Akṣobhya carried out in 1289 CE:

east java

This second consecration of the image of Akṣobhya was accompanied by a ritual in which Kṛtanagara was initiated according to the rites of Mahākṣobhya ... It is clear that this symbolic and ritual act was meant to have far-reaching effects. As scholars have largely agreed since the idea was first advanced by J.L. Moens,5959 Kṛtanagara’s consecration into the ritual path of Akṣobhya (and hence the Guhyasāmaja Tantra) appears to have been intended to counteract the ritual power gained by the Yuan emperor Kublai Khan through his consecrations in 1264 and 1269 as Hevajra, the central Jina of the Hevajra Tantra. This connects the event with Kṛtanagara’s earlier gift to

A political confrontation between Java and China that included supreme Buddha consecrations occurred half a century later and has drawn the attention of several scholars. Siṅhasāri King Kṛtanagara (r. 1268–1292 CE) underwent a consecration after receiving envoys from the Yuan emperor Kublai Khan (r. 1260–1294 CE), who had constructed a fleet and was threatening a naval invasion.58 Thomas M. Hunter’s reading of the Sanskrit inscrip58 David Bade points out that many scholars have seen Kublai’s prime motivation for undergoing the Vajrayāna empowerment as ‘an instrument of rule in order to justify his military conquest and reign over Tibet’ (Bade 2016: 141). But the motivation would also have underwritten the aggressive messages of Kublai’s envoys to Java from 1279 to 1286 CE calling for submission and the sending of royal hostages to Beijing as recorded in the Yuan shi. With his fleet, Kublai invaded Campā and the Đai Việt as well as Java and attempted to invade Japan.

04_CreSouthV1_4P_22Jan22.indd 154

59 Moens 1924: 544. In Bade’s translation of the key passage: ‘The most well-known of these rulers of China, Kublai Khan, contemporary of Kṛtanagara, was consecrated as Jina through the Hevajrābhiṣeka ... The iṣṭadevatā of Kṛtanagara, who thought of himself as Kublai’s match, should not have been less demonic’ (Bade 2016: 141, n. 2).

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia Mauliwarmadewa, for this monarch controlled the Melaka Straits, which was the linch-pin in the all-important international sea trade, and an obvious point that needed to be defended at all costs from the expansionist politics of Kublai Khan. (Hunter 2007: 40)

Max Nihom (1986: 497) found that the text yielded a somewhat obscure description of the king as Bhaṭāra Śiva-Buddha reflecting the dharmakāya of Mahākṣobhya, who issues as ... Vajradhara himself, the teacher of the Hevajratantra, [who] when he does appear … is an extravagance with 8 faces, 4 legs and 16 arms (II.v.8) [i.e. Hevajra] ... Although many aspects of the cult are as yet opaque—for example, the precise nature of the gaṇacakra of Nāg[arakṛtāgama]. 43.3d—based on the results offered above there is good reason to suppose that a cult of Hevajra was practised by the royal house. As such, the consecration of Kṛtanagara was indeed in all probability a consequence, in syncopation or reaction, to the tantric consecrations of Kublai Khan.

155

which Grünwedel traced in Tibetan accounts over relative chair heights for the consecration ceremony.61 The Mongol emperor responded by conferring on ’Phags pa suzerainty over three regions of Tibet. Another form of the emperor’s Hevajrābhiṣeka62 is suggested by a remarkable set of maṇḍalic aids for submission to the deity that was once on public display in Beijing (the railing behind the figures still stands outside the Forbidden City): six large bronze yab-yum couples sitting on mounts that include an elephant, a man, a buffalo, a bull, a goat and a gazelle. They were photographed in the 1920s before they disappeared without trace.63 Pott (1966: 70) thought a maṇḍala on this scale could only have ‘featured in the consecration of a sovereign, perhaps at Kublai’s Hevajravaśitā [“submission to Hevajra”].’64 David Bade’s scalpel cuts neatly through the historiographical presuppositions in the accounts of realpolitik motivation assumed in much of the recent scholarly consensus on the Kublai–Kṛtanagara confrontation. He counters this by arguing that both rulers turned to the Esoteric Buddhists for a

The Yuan emperor’s empowerment in the Hevajra 61 Albert Grünwedel wrote the first European account of cult in 1264 and 1268 CE was performed by ’Phags pa travelling under duress (as a hostage?) to Dadu Tibetan Sa skya master ’Phags pa, and inaugurat- (Beijing) where he eventually conferred the initiation of Hevajrābhiṣeka on the emperor. According to Tibetan ed a century of Tibetan Buddhism as the official legend, the Chinese had appropriated the text of the Hevareligion of China under Mongol rulers, before the jratantra before ’Phags pa had read it, so the god Mahākāla Sakya transmission was deemed heterodox under brought it to him for overnight study. The Lama, after agreethe Ming (1368–1644 CE) (Elacqua 2017: 1304). The ing to compromise on having chairs of the same height, ceremony, recorded from a Tibetan Sa skya view- initiated the emperor and was granted in return political point in a 16th–17th century painting at the Rubin and religious sovereignty over much of Tibet (Grünwedel 1900: 63–65). Museum, New York, shows Kublai as a tiny figure 62 I am grateful for Iain Sinclair’s elucidation of the terms with hands pressed together in respect, beside a ‘Hevajrābhiṣeka’, ‘Hevajravaśitā’, and ‘Hevajrasekaprakriyā’ much bigger ’Phags pa, the Imperial Preceptor or in Sanskrit, Tibetan and Chinese for naming this crucial Ti-shih and nephew of the leading Tibetan scholar consecration ritual. and Sa skya Paṇḍita.60 ’Phags pa, whom Willemen 63 See E.E. Schlieper’s photograph in Laufer 1922: 14, (1983: 16, n. 28) says ranked second only to the accessible at https://archive.org/details/milaraspatibetisemperor in power, is seated for the ceremony on 00milauoft/page/14/mode/2up (last accessed 10 September the emperor’s throne in the palace in Dadu (Beijing) 2020). I am grateful to a peer reviewer for locating this image. with the diminutive Kublai beside him. The rela64 Pott 1966: 70. Alternatively, the bronzes may have been tive sizes of the figures appear to reflect a dispute, created for a performance of the Cakrasaṃvaratantra, for 60 ‘Initiation of Qubilai Khan and offering Tibet to Phakpa in 1264’, 16th–17th century: https://rubinmuseum. org/collection/artwork/initiation-of-qubilai-khan-and-offering-tibet-to-phakpa-in-1264 (last accessed 5 April 2020).

04_CreSouthV1_4P_22Jan22.indd 155

with the addition of a horse and a makara, these mounts would correspond exactly with the eight ksetrapāla of this Tantra, in which Saṃvara, another Heruka contemporary with Hevajra, is described ripping open an elephant hide representing illusion (Avalon, Kazi Dawa Samdup and Lokesh Chandra 1987: 20).

28/1/22 5:38 PM

Peter Sharrock

156

new vision when they were faced with international Lokesh Chandra (1995: 156) sees Kṛtanagara’s issues of a scope and complexity unknown in their response as aimed to both arrest the Mongol threat own cultures. They put their trust in a compelling ‘... and to pay homage to the Mongol Emperor’. image of a cakravartin at the pinnacle of a maṇḍala Paying homage seems aligned with Ācārya Smith’s that mirrored the organization of the macrocosm. view (above) that participation in the supreme Bade (2016: 158) rightly commends Kate O’Brien’s Heruka/Vajrasattva state of being took initiates reading of Mpu Tantular’s classic 14th-century beyond violence and that this should be reflected in kakavin Sutasoma65 as a poetic rendering of the mundane inter-state actions. Kṛtanagara was killed spiritual meaning of the confrontation as it was in 1292 CE before the huge Mongol fleet landed, perhaps understood in 14th-century Java. O’Brien but the invasion failed (as the others did in Japan sees the kakavin hero prince Sutasoma as Kṛtana- and Vietnam) and the East Javanese dynasty ‘was gara, whose mission is to save humanity from the restored in 1294 CE by [Kṛtanagara’s son-in-law] man-eating demon king Poruṣāda (= Kublai). Suta- Raden Wijaya, who tricked Kubilai’s invading forces soma discovers himself through a series of battles— into attacking [his rival] Jayakatwang, then turned won with ‘Buddha-Mind, the weapon of Absolute on them and destroyed them as they celebrated Knowledge’, and not by violence—to become ‘Wai- victory’ (Hunter 2007: 28).66 Hunter attributes the rocana in mortal embodiment’ and the lord of the ongoing Esoteric Buddhist tradition in Java and maṇḍala after defeating Poruṣāda (ibid.). In Bade’s Sumatra to the lasting impression made by Kṛtaview, O’Brien here seems to approach a more au- nagara’s fusion of Śaivism and Buddhism to attain thentic 14th-century Javanese view of the nature political and spiritual unity both in Java and in of what he dubs ‘(spi)ritual warfare—as well as a his growing empire. In 1292 CE Kṛtanagara was naval campaign and land battles’ (Bade 2016: 158): enshrined at Candi Jawi in a single Śiva-Buddha image after his ‘release into the realms of Śiwa and Tantric Buddhism, with its imagination of a Buddha (mokteng siwa-buddha-loka)’ (ibid.: 41). just and universal world ruler bringing about a state of affairs on earth that would mirror the heavenly state of affairs—the world as it ought to be—was not just a fantasy world of escape for monks and a useful mask hiding ugly realities for kings; it was a powerful motivating force that shaped political and social conditions throughout Asia.… The development of Tantric Buddhism and modern science were both responses to the world in a particular time and place, and both were/are oriented by the idea that knowledge is power, total knowledge being total power. (ibid.: 155–56)

65 Mpu Tantular was a poet and purohita at the court of King Hayam Wuruk of Majapahit (r. 1350–1389 CE). The Sutasoma holds one mention of Heruka in canto 124b.9d–10: ‘Having attained that [knowledge (gnosis) of Non-duality], he is considered to be Jinapati, the highest point of the State of the Absolute. […] His behaviour had many aspects—among other things he held a corpse (over his head) as if it were a parasol while performing concentration of the mind[.] Smelly blood flowed; drops of it fell on his head and ran down his chest. Its intestines were a tangled mass, green flies on (his) face crawled into (his) eyes, [n]evertheless his mind was unswerving in its quest to bring forth the Lord Heruka’ (trans. O’Brien 2008: 137).

04_CreSouthV1_4P_22Jan22.indd 156

sumatra Andrea Acri (2016: 8, 19–22) sees the Esoteric Buddhist cults of Jayavarman VII, Kṛtanagara and Kublai Khan as a ‘second wave’ that long remained vital in Maritime Asia. In the early 11th century, the collapsing Pāla dynasty in northern India led to an exodus from the Ganges Valley monasteries in Bengal and Bihar. The future Tibetan Buddhist patriarch Atiśa (982–1054 CE) undertook what was later called a dangerous sea passage to Śrīvijaya where he studied at the feet of the guru Dharmakīrti from 1013–25 CE, presumably at the pīṭha named as Suvarṇadvīpa (‘Isle of Gold’) in the lists cited in the Hevajratantra and the Saṃvarodayatantra. Tibetan archives credit Dharmakīrti with a commentary on the Tantra. Schoterman (2016 [1986]: 115) notes attacks from Kabul in 1018 by Mahmūd of Ghaznī ‘almost reaching Benares (which he plundered in 1033) and attacking Somnāth (Kathiawar) 66 Munoz (2006: 260–274) says the Chinese, battered and running short of supplies, cut their losses and left rather than wait months for the winds to turn.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia in 1024.’ For a time Suvarṇadvīpa emerged as a haven for Buddhist study: Although the Buddhism in northeastern India during this period was not yet directly confronted with Islam, it was highly susceptible to infiltration by various Tantric currents, as a result of which ‘classical’ Buddhism virtually disappeared. It is not difficult to imagine that Śrīvijaya took over the role of northeast India in this period. There one could study classical Buddhism in peace. (Schoterman 2016 [1986]: 115)

But this was soon to change. The Cōḻa army marched from south India to the north and then became the first Indian dynasty to expand its empire by sea. In 1025 CE the Cōḻa fleet attacked and largely destroyed Palembang and Jambi on Sumatra, as well as Kedah on the Malay Peninsula. Months before this, Atiśa returned to Bengal to become abbot of Vikramaśīla, before being eventually invited to Tibet in 1040 CE to begin the celebrated reform mission of his remaining fourteen years. Schoterman (2016 [1986]: 115) paints a bleak picture of Sumatra in the late 11th and 12th centuries: Weakened by the attacks from southern India at the hands of the Cōḻas ... Śrīvijaya lost its prominent position in the Buddhist world. Perhaps this is borne out in another manuscript of the previously mentioned Aṣṭasāhasrikā Prajñāpāramitā ... which dates from 1071. In this manuscript we also find a series of miniatures of Buddhist sanctuaries in India and beyond. The island of Sri Lanka is again mentioned thrice, which suggests that little changed in the intervening period. As regards the Indonesian Archipelago, however, the picture has radically altered. Only Java is still mentioned. Both Śrīvijaya and Kedah have disappeared from the list of important Buddhist sanctuaries.

In neighbouring East Java, there is little evidence of the Buddhist Tantras being influential under the Kaḍiri dynasty (929–1222) but Woodward (2004: 344) thought some maṇḍala bronzes in the Surocolo hoard could suggest an early 11th-century

04_CreSouthV1_4P_22Jan22.indd 157

157

entry for the Hevajratantra. Lokesh Chandra and Sudarshana Devi Singhal (1995) considered half the pieces as belonging to Hevajra’s maṇḍala— three of his eight Yoginīs, four heavenly musicians and two of the animal-headed female door guardians. Tanaka (2018: 219) says the Surocolo bronze statuettes are identified by Buddhist iconographer Matsunga Keiji ‘... as a three-dimensional maṇḍala of the Vajrasattva family of the Samāyogatantra.’ Sumatra was recovering from the Cōḻa onslaught in the late 13th century when Kṛtanagara expanded the alliances of the Siṅhasāri dynasty to the Malayu King Maulivarmadeva. Kṛtanagara’s move eastwards possibly provoked the erection, as a defence mechanism, of the Padang Lawas Heruka mentioned earlier (Fig. 4.1). The complex of temples there was part of the polity of Panai, a Malayu/Jambi-dependency, and was active from the 10th to the 14th centuries. The Heruka with almost identical pose, ritual accoutrements, flaming aureole around Akṣobhya in the mukuṭa and robe flying up the backslab, is carved in a local tufa stone and is indeed very close in style and arrangement to those in Śubhapur, Nālandā, and Ratnagiri. All hold the skull-staff khaṭvāṅga, skull-bowl padmabhājana, and drum (ḍamaru) bestowed after the fourfold empowerment ritual. Inscribed stones at the site (some incorporating phrases from the Hevajratantra, according to Griffiths 2014: 18–22) near the three main towers bear dates of 1179, 1245, and 1372 CE (Bautze-Picron 2014: 108 n. 3). Dating the Heruka is difficult, the 13th century being the most likely option. Relations with the expanding Rājarāja Cōḻa dynasty in Tamil Nadu were cordial until Panai was hit in a surprise raid, through the western Sumatran port of Barus that heralded the start of King Rājendra Cōḻa’s devastating invasion of 1025. But despite the close stylistic parallels with the 11th-century north Indian icons, the erection of the Panai Heruka is more likely to have been long after the Cōḻa invasion. The presence in Padang Lawas of a small stone icon of Mahākāla, one of the Bhairava-like defenders of Buddhism, matches similar icons of the Siṅhasāri dynasty. Having suffered under the Cōḻas, the Panai kings may have sought supernatural protection against the expanding

28/1/22 5:38 PM

158

Peter Sharrock

empire in neighbouring Java. Its spiritual/political message may have been similar to that of the Heruka dedicated in Campā (inscription C. 92B) by Cam prince Vidyānandana, who broke with his Khmer backers and proclaimed sovereignty (see above). We will return later to Sumatra. During the subsequent expansion of the Majapahit golden age under King Hayam Wuruk (r. 1350–1389 CE), celebrated today as setting the bounds of modern Indonesia, one of the king’s cousins moved to Sumatra as King Ādityavarman (c. 1347–1379), securing control of the gold mines and the trade through the Melaka Straits.67 He eventually asserted autonomy within the now sprawling empire by adding an epigraph re-dedicating to himself the famous Amoghapāśa stela sent by Kṛtanagara decades earlier to celebrate a marriage alliance. Ādityavarman moved the stela to Padang Candi, leaving behind Kṛtanagara’s inscribed pedestal, and in an unpolished Sanskrit on the back upgraded his title to the supreme title maharājādhirāja, a title reserved for his cousin in East Java. A few years later he ‘moved the capital to Saruaso ... a safe haven in Malayu’s fertile mountainous hinterland’ (Kozok and van Reijn 2010: 136) that controlled the output of the gold mines. At Sungai Langsat he raised Indonesia’s largest known statue, a 4.4 m ‘Bhairava’ with bulging eyes, a knife at his chest, standing on a skull pedestal and with Akṣobhya set in his huge ball of ascetic hair (Figs. 4.21, 4.22). Bautze-Picron considers ‘Bhairava’ a misidentification and sees it as the fierce Buddhist deity Mahākāla, protector of Kublai Khan’s Yuan court (Bautze-Picron 2014: 113–114). Could this image represent another link back to Kublai’s Javanese contemporary Kṛtanagara? The only referent in this period for the scale of the statue, the pedestal of skulls, the fine quality of the stone and 67 ‘Adityavarman, a devotee of Tantric Buddhism, became immensely powerful and was in effect independent as king of the Minangkabau highlands, which produced gold and also had developed intensive rice cultivation. Subsequent attempts by Java to reassert its authority were to lead in about 1391 to the flight of a prince of Melayu to Temasek and then to the founding of Melaka’ (Bowring 2019: 114–115).

04_CreSouthV1_4P_22Jan22.indd 158

sophisticated carving is Kṛtanagara’s kingdom; nothing comparable is known from Sumatra (Acri and Wenta 2022). Hunter (2007: 40–41) describes the Śiva-Buddha image in Kṛtanagara’s mortuary temple as follows: Perhaps the most spectacular evidence of Kṛtanagara’s drive to carry forward the cause of political and symbolic unity initiated by his predecessors is the exceptional case of his post-mortem enshrinement at Candi Jawi as a single ‘Śiwa-Buddha image’ (siwa-buddhârca), that the Desawarnana commemorates as his ‘release into the realms of Śiwa and Buddha’ (mokteng siwa-buddha-loka) in the year 1292 CE. In its outward appearance the shrine at Candi Jawi reveals a unique linkage of Shaivite and Buddhist elements. The slim, tapering spire of the main shrine links it with the form of temple spire favoured in Shaivite sanctuaries of East Java, while its capstone, in the form of a stupa, represents the Buddhist element. Most remarkably, the image within the main shrine also combined Shaivite and Buddhist elements: mwang ri jro Siwa-wimba sobhita halep-niraparamita Aksobhya-pratime ruhur (m)makuta tan hanoly n-tika And within there was a splendid Shiva-image, its beauty fitting, without parallel (With) an image of Aksobhya above the crown, of that there can be no doubt. [Robson, Deśavarnana (Nāgarakṛtāgama), 56.2]

The Saruaso II inscription mentions Ādityavarman’s son, the crown prince Anaṅgavarman as daily practising ‘meditation on Hevajra’ (hevajranityāsmṛtiḥ). This concludes the overt signs of the long ‘second wave’ of royal Esoteric cults in Maritime Asia. Having virtually died out in the Indian Subcontinent by the late 13th century, it continued to live or even thrive—in its localized adapta-

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia

159

Figs. 4.21, 4.22: Ādityavarman’s Bhairava/Mahākāla with Akṣobhya, Jakarta National Musuem. (Photos by author)

tions—until the 15th century in Java and Sumatra (e.g., under King Ādityavarman, r. ?–1375), and to the present day in Nepal, Tibet, Bali, and Japan. (Acri 2016: 8)

The ritual technology of the Heruka-maṇḍalas, the fruit of an historical fusion of Indic tantric Śaiva

04_CreSouthV1_4P_22Jan22.indd 159

and Buddhist ideologies, attained its political florescence not in the subcontinent, where its longterm application was cut short by the rise of armed Islam, but in the politico/religious strategies of the royal courts of Maritime Asia from Phimai to Angkor to Mỹ Sơn to Siṅhasāri and Sumatra over more than three centuries.

28/1/22 5:38 PM

160

Peter Sharrock

. 1963. La Statuaire du Campà: recherches References sur les cultes et l’iconographie. Paris: École franAbé, Ryūichi. 1999. The Weaving of Mantra: Kūkai çaise d’Extrême-Orient. and the Construction of Esoteric Buddhist DisBosch, F.D.K. 1930. ‘Verslag van een reis door course. New York: Columbia University Press. Sumatra’ (Report of a Journey through Acri, Andrea. 2016. ‘Introduction: Esoteric BudSumatra), Oudheidkundig Verslag, Bijlage C: dhist Networks along the Maritime Silk Routes, 133–157. 7th–13th Century AD’, in Andrea Acri (ed.), Bowring, Philip. 2019. Empire of the Winds: The Esoteric Buddhism in Mediaeval Maritime Asia: Global Role of Asia’s Great Archipelago. London: Networks of Masters, Texts, Icons, pp. 1–25. SinI.B. Tauris, Bloomsbury Publishing. gapore: ISEAS – Yusof Ishak Institute. Briggs, Lawrence P. 1948. ‘Siamese Attacks on Acri, Andrea and Aleksandra Wenta. 2022. ‘A Angkor Before 1430’, The Far Eastern QuarBuddhist Bhairava? Kṛtanagara’s Tantric terly 8/1: 3–33. Buddhism in Transregional Perspective (Thir. 1999 [reprint 1951]. The Ancient Khmer teenth–Fourteenth Centuries CE)’, in EntanEmpire. Bangkok: White Lotus. gled Religions. In press. ‘Chronique’. 1925. Bulletin de l’École française d’ExAvalon, Arthur, Kazi Dawa Samdup, and Lokesh trême-Orient 25: 543–595. Chandra (eds.). 1987. Shrichakrasambhara Cœdès, George. 1923. Bronzes khmèrs. Paris: Ars Tantra/Sri-Cakrasamvara-Tantra. A BudAsiatica. dhist Tantra. New Delhi: Aditya Prakashan. . 1929. ‘Études Cambodgiennes XXIV. Nou[Reprint of Avalon and Kazi Dawa Samdup velles donnés chronologiques et généalogiques 1919, London, Luzac & Co.] sur la dynastie de Mahīdharapura’, Bulletin de Aymonier, Étienne F. 1901. Le Cambodge vol. II: l’École française d’Extrême-Orient 29: 309–315. les provinces siamoises. Paris: Ernest Leroux. . 1932. ‘Études Cambodgiennes XXVIII. Bade, David. 2016. ‘(Spi)ritual Warfare in 13th-CenQuelques suggestions sur la méthode à suivre tury Asia? International Relations, the Balance pour interpréter les bas-reliefs de Bantéay of Powers, and the Tantric Buddhism of KṛtanaČhmàr et de la galerie extérieure du Bàyon’, gara and Khubilai Khan’, in Andrea Acri (ed.), Bulletin de l’École française d’Extrême-Orient Esoteric Buddhism in Mediaeval Maritime Asia: 32: 71–112. Networks of Masters, Texts, Icons, pp. 141–159. . 1942a. ‘Inscription du Bàyon K. 470’, InsSingapore: ISEAS – Yusof Ishak Institute. criptions du Cambodge, Vol. 2, pp. 187–189. Bautze-Picron, Claudine. 2014. ‘Buddhist Images Hanoi: École française d’Extrême-Orient. from Padang Lawas Region and the South . 1942b. ‘Grande stèle du Phīmãnàkàs’ (K. Asian Connection’, in Daniel Perret (ed.), 485), Inscriptions du Cambodge, Vol. 2, pp. History of Padang Lawas II. Societies of Padang 161–392. Hanoi: École française d’ExtrêmeLawas (mid-9th–13th century CE), Cahiers Orient. d’Archipel vol. 43, pp. 107–128. Paris: Asso. 1968. The Indianized States of Southeast ciation Archipel. Asia (ed. Walter F. Vella, trans. Sue Brown Boeles, J.J. 1966. ‘Two Yoginīs of Hevajra from Cowing). Honolulu: University of Hawai‘i Thailand’, in Essays Offered to G.H. Luce by Press. His Colleagues and Friends in Honour of His Conlan, Thomas D. 2011. Sovereign to Symbol: Seventy-Fifth Birthday. Volume 2: Papers on An Age of Ritual Determinism in 14th Century Asian Art and Archaeology: Artibus Asiae. SupJapan. New York: Oxford University Press. plementum, vol. 23, pp. 14–29. Zurich. Conti, Pia. 2014. ‘Tantric Buddhism at Prasat Hin Boisselier, Jean. 1957. ‘Vajrapani dans l’art du Phimai: A New Reading of Its Iconographic Message’, in Nicolas Revire and Stephen A. Bàyon’, in Zeki Velidi Togan (ed.), ProceedMurphy (eds.), Before Siam: Essays in Art and ings of the 22nd Congress of Orientalists, vol. 2, Archaeology, pp. 374–395. Bangkok: River Books. Communications, pp. 324–332. Leiden: Brill.

04_CreSouthV1_4P_22Jan22.indd 160

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia Cunin, Olivier. 2005. De Tà Prohm au Bàyon: contribution à l’histoire architecturale du temple du Bàyon. PhD thesis, Institut National Polytechnique de Lorraine. . 2018. ‘About some missing images of the Bayon temple’. Lecture at the ‘Exploring Angkor Symposium’ at the Asian Civilisations Museum (ACM), Singapore. [https://www. academia.edu/41385160/About_some_missing_ images_of_the_Bayon_temple] Dagens, Bruno. 1969. ‘Étude sur l’iconographie du Bàyon (frontons et linteaux)’, Arts Asiatiques 19: 123–167. . 2003. Les Khmers. Paris: Éditions les belles lettres. Dalsheimer, Nadine. 2001. Le collections du musée national de Phnom Penh. Paris: École française d’Extrême-Orient/Magellan & Cie. Dasgupta, S.B. 1950. An Introduction to Tantric Buddhism. Calcutta: University of Calcutta. Davidson, Ronald M. 1991. ‘Reflections on the Maheśvara Subjugation Myth: Indie Materials, Sa-skya-pa Apologetics, and the Birth of Heruka’, Journal of the International Association of Buddhist Studies 14/2: 197–235. . 2005. Tibetan Renaissance: Tantric Buddhism in the Rebirth of Tibetan Culture. New York: Columbia University Press. Dowling, Nancy. 2000. ‘New Light on Early Cambodian Buddhism’, The Journal of the Siam Society 88/1–2: 122–155. Elacqua, Joseph P. 2017. ‘Vajrayāna (Buddhism)’, in K.T.S. Sarao and Jeffery D. Long (eds.), Encyclopedia of Indian Religions: Buddhism and Jainism, pp. 1299–1308. Dordrecht: Springer. Estève, Julia and Brice Vincent. 2010. ‘L’about inscrit du musée national du Cambodge (K. 943): nouveaux éléments sur le bouddhisme tantrique à l’époque angkorienne’, Arts Asiatiques 65: 133–158. Evans, Damian. 2016. ‘Airborne laser scanning as a method for exploring long-term socio-ecological dynamics in Cambodia’, Journal of Archaeological Science 74: 164–175. Finot, Louis. 1904. ‘Les inscriptions’, in Henri Parmentier and Louis Finot, Le cirque de Mĩ-Sợn (Quang-Nam), pp. 83–173. Hanoi: F.-H. Schneider.

04_CreSouthV1_4P_22Jan22.indd 161

161

. 1934. ‘Manuscripts sanskrits de sādhana’s retrouvés en Chine’, Journal Asiatique 225: 1–85. Goble, Geoffrey. 2016. ‘The Politics of Esoteric Buddhism: Amoghavajra and the Tang State’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons, pp. 123–139. Singapore: ISEAS – Yusof Ishak Institute. Golzio, Karl-Heinz (ed.). 2004. Inscriptions of Campā based on the editions and translations of Abel Bergaigne, Étienne Aymonier, Louis Finot, Édouard Huber and other French scholars and of the work of R.C. Majumdar. Aachen: Shaker Verlag. Gray, David B. 2006. ‘Skull Imagery and Skull Magic in the Yoginī Tantras’, Pacific World: Journal of the Institute of Buddhist Studies, 3rd series, 8: 21–39. . 2007. The Cakrasaṃvara Tantra (The Discourse of Śri Heruka): Study and Annotated Translation. New York: Columbia University Press. Griffiths, Arlo. 2014. ‘Inscriptions of Sumatra III. The Padang Lawas corpus studied along with inscriptions from Sorik Merapi (North Sumatra) and from Muara Takus (Riau)’, in Daniel Perret (ed.), History of Padang Lawas II. Societies of Padang Lawas (mid-9th–13th century CE), pp. 211–237. Paris: Association Archipel . Grünwedel, Albert. 1900. Mythologie des Buddhismus in Tibet und der Mongolei. Leipzig: Brockhaus. Hall, Kenneth R. 1996. ‘Ritual Networks and Royal Power in Majapahit Java’, Archipel 52/1: 95–118. Hunter, Thomas M. 2007. ‘The Body of the King: Reappraising Singhasari Period Syncretism’, Journal of Southeast Asian Studies 38/1: 27–53. Ishizawa, Yoshiaki and Masako Marui. 2002. ‘La découverte de 274 sculptures et d’un caitya bouddhique lors des campagnes de fouilles de 2000 et 2001 au temple de Banteay Kdei à Angkor’, Arts Asiatiques 57: 206–218. Jacques, Claude. 1995. Études épigraphiques sur le pays cham, de Louis Finot, Edouard Huber, George Cœdès et Paul Mus réunies par Claude Jacques. Paris: Réimpression de l’École française d’Extrême-Orient N°7.

28/1/22 5:38 PM

162

Peter Sharrock

. 2015. ‘The History and Inscriptions’, in Peter D. Sharrock (ed.), Banteay Chhmar: Garrison-temple of the Khmer Empire, pp. 135–150. Bangkok: River Books. . 2016a. ‘De la fragilité de notre science historique: la stèle K. 1297’. Unpublished paper given at the École française d’Extrême-Orient, Paris. . 2016b. ‘K. 1297 Face A. Maṅgala’. Unpublished paper. Kozok, Uli and Eric van Reijn. 2010. ‘Ādityawarman: Three inscriptions of the Sumatran “king of all supreme kings”’, Indonesia and the Malay World 38: 135–158. Laufer, Berthold. 1922. Milaraspa, Tibetische Texte. Hagen and Darmstadt: Folkwang Verlag. Lehnert, Martin. 2012. ‘Tantric Threads between India and China’, in Ann Heirman and Stephan Peter Bumbacher (eds.), The Spread of Buddhism, pp. 247–276. Leiden: Brill. Linrothe, Rob. 1999. Ruthless Compassion: Wrathful Deities in Early Indo-Tibetan Esoteric Buddhist Art. London: Serindia. Lobo, Wibke. 1997. ‘L’image de Hevajra et le bouddhisme tantrique’, in Helen Jessup and Thierry Zéphir (eds.), Angkor et dix siècles d’art khmer, pp. 70–78. Paris: Réunion des musées nationaux. Lokesh Chandra (in collaboration with Sudarshana Devi Singhal). 1995. ‘The Jaka Dolog inscription of King Kṛtanagara’, in Lokesh Chandra (ed.), Cultural Horizons of India, Vol. 4, pp. 148–166. New Delhi: International Academy of Indian Culture/Aditya Prakashan. Lokesh Chandra and Sudarshana Devi Singhal. 1995. ‘The Buddhist bronzes of Surocolo’, in Lokesh Chandra (ed.), Cultural Horizons of India, Vol. 4, pp. 121–147. New Delhi: International Academy of Indian Culture/Aditya Prakashan. Lowman, Ian. 2016. ‘The Stele K. 1297’. Unpublished paper containing translation excerpts and exegesis of this inscription dated Feb. 17, 2016. Mallmann, Marie-Thérèse de. 1986. Introduction à l’iconographie du Tântrisme bouddhique. Paris: Librairie d’Amérique et d’Orient.

04_CreSouthV1_4P_22Jan22.indd 162

Maxwell, T.S. 2007. ‘Religion at the time of Jayavarman VII’, in Joyce Clark (ed.), Bayon: New Perspectives, pp. 72–135. Bangkok: River Books. Moens, J.L. 1924. ‘Het Boeddhisme op Java en Sumatra in zijn laatste bloeiperiode’, Tijdschrift van het Bataviaasch Genootschap van Kunsten en Wetenschappen 64: 521–579. Monier-Williams, M. 1899. Sanskrit-English Dictionary. Oxford: Clarendon Press. Munoz, Paul Michel. 2006. Early Kingdoms of the Indonesian Archipelago and the Malay Peninsula. Singapore: Editions Didier Millet. Nihom, Max. 1986. ‘The Identification and Original Site of a Cult Statue on East Java: The Jaka Dolog’, Journal of the American Oriental Society 106/3: 485–501. O’Brien, Kate. 2008. Sutasoma: The ancient tale of a Buddha-Prince from 14th century Java by the poet Mpu Tantular. Bangkok: Orchid Press. Orlando, Raffaello. 1981. A study of Chinese documents concerning the life of the Tantric Buddhist patriarch Amoghavajra (A.D. 705–774). PhD thesis, Princeton University. Orzech, Charles D. 1998. Politics and transcendent wisdom: The scripture for humane kings in the creation of Chinese Buddhism. University Park: Pennsylvania State University Press. Phương, Tran Ky. 2018. ‘Object 76. Part II’, in Trần Kỳ Phương, Võ Văn hắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Ðà Nẵng Museum of Cham Sculpture, pp. 247–249. Bangkok: River Books. Pott, Philipp H. 1966. Yoga and Yantra: Their Interrelation and their Significance for Indian Archaeology. The Hague: Martinus Nijhoff. Priest, Alan. 1937. ‘A collection of Cambodian sculpture’, Bulletin of the Metropolitan Museum of Art 9: 84–88. Reichle, Natasha. 2007. Violence and Serenity: Late Buddhist Sculpture from Indonesia. Honolulu: University of Hawai‘i Press. Samuel, Geoffrey. 2008. The Origins of Yoga and Tantra: Indic Religions to the Thirteenth Century. Cambridge: Cambridge University Press. Sanderson, Alexis. 2003–2004. ‘The Śaiva Religion among the Khmers (Part I)’, Bulletin de l’École française d’Extrême-Orient 90–91: 349–462.

28/1/22 5:38 PM

Heruka-Maṇḍalas across Maritime Asia . 2009. ‘The Śaiva Age: The Rise and Dominance of Śaivism during the Early Medieval Period’, in Shingo Einoo (ed.), Genesis and Development of Tantrism, pp. 41–350. Tokyo: Institute of Oriental Culture, University of Tokyo Special Series 23. Schnitger, F.M. 1937. The Archaeology of Hindoo Sumatra. Leiden: Brill. Schoterman, Jan. 2016. ‘Traces of Indonesian Influences in Tibet’, in Andrea Acri (ed.), Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons, pp. 113–122. Singapore: ISEAS – Yusof Ishak Institute. [English translation of Indonesische Sporen in Tibet. Leiden: Brill, 1986] Schweyer, Anne-Valérie. 2018. ‘Buddhism in Campā’, in Trần Kỳ Phương, Võ Văn hắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cam Sculpture, pp. 71–78. Bangkok: River Books. Sharrock, Peter. 2006. The Buddhist pantheon of the Bàyon of Angkor: An historical and art historical reconstruction of the Bàyon temple and its religious and political roots. PhD thesis, SOAS. . 2007. ‘The Mystery of the Face Towers’, in Joyce Clark (ed.), Bayon: New Perspectives, pp. 232–281. Bangkok: River Books. . 2009. ‘Garuḍa, Vajrapāṇi and Religious Change in Jayavarman VII’s Angkor’, Journal of Southeast Asian Studies 40/1: 111–151. . 2012. ‘Kīrtipaṇḍita and the Tantras’, Udaya: Journal of Khmer Studies 10: 203–237. . 2013. ‘The yoginīs of the Bayon’, in István Keul (ed.), ‘Yoginī’ in South Asia: Interdisciplinary Approaches, pp. 117–129. London: Routledge.​ . 2018. ‘Cam-Khmer interactions 1113–1220 CE’, in Trần Kỳ Phương, Võ Văn hắng, and Peter D. Sharrock (eds), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cam Sculpture, pp. 111–119. Bangkok: River Books. Sharrock, Peter and Claude Jacques. 2017. ‘The Grief of Kings Is the Suffering of their Subjects: A Cambodian King’s Twelfth-Century Network of Hospitals’, in C. Pierce Salguero (ed.), Buddhism and Medicine: An Anthology of Premodern Sources, pp. 226–232. New York: Columbia University Press.

04_CreSouthV1_4P_22Jan22.indd 163

163

Snellgrove, David L. 1959. The Hevajra Tantra: A Critical Study. London: Oxford University Press. . 1987 [revised 2004]. Indo-Tibetan Buddhism: Indian Buddhists and Their Tibetan Successors. Bangkok: Orchid Press. . 1988. ‘Categories of Buddhist Tantras’, in G. Gnoli and L. Lanciotti (eds.), Orientalia Josephi Tucci Memoriae. Seria Orientali Roma 56/3: 1353–1384. . 2001. Khmer civilization and Angkor. Bangkok: Orchid Press. Southworth, William A. 2014a. ‘The Alteration and Destruction of Buddhist Images at Angkor’, in Julia A.B. Hegewald (ed.), In the Shadow of the Golden Age: Art and Identity in Asia from Gandhara to the Modern Age, pp. 197–225. Berlin: EB-Verlag. Sundberg, Jeffrey and Rolf Giebel. 2011. ‘The life of the Tang court monk Vajrabodhi as chronicled by Lü Xiang 呂向: South Indian and Śrī Laṅkān antecedents to the arrival of the Buddhist Vajrayāna in 8th century Java and China’, Pacific World, Third Series, 13 (Fall): 129–222. Szántó, Péter-Dániel and Arlo Griffiths. 2015. ‘Sarvabuddhasamāyogaḍākinījālaśaṃvara’, in Brill’s Encyclopaedia of Buddhism, vol. 1, Literature and Languages, pp. 367–372. Leiden: Brill. Takakusu, J. 1896. A Record of the Buddhist Religion as Practised in India and the Malay Archipelago (A.D. 671–695). Oxford: Clarendon Press. Tanaka, Kimiaki. 2018. An Illustrated History of the Mandala: From its Genesis to the Kālacakratantra. Boston: Wisdom publications. Thapar, Romila. 1961. Asoka and the Decline of the Mauryas. Oxford: Oxford University Press. Tsuda, Shinichi. 1974. The Saṃvarodaya-tantra: Selected Chapters. PhD thesis, Australian National University. Vickery, Michael. 2006. ‘Bayon: New Perspectives Reconsidered’, Udaya: Journal of Khmer Studies 7: 101–173. Whitmore, John K. 2018. ‘Nagara Campa and the Vijaya turn’, in Trần Kỳ Phương, Võ Văn hắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cam Sculpture, pp. 31–36. Bangkok: River Books.

28/1/22 5:38 PM

164

Peter Sharrock

Willemen, C. 1983. The Chinese Hevajratantra: The Scriptural Text of the Ritual of the Great King of the Teaching, the Adamantine One with Great Compassion and Knowledge of the Void. Leuven: Peeters. [repr. 2004, New Delhi: Motilal Barnarsidass] Woodward, Hiram H. 1981. ‘Tantric Buddhism at Angkor Thom’, Ars Orientalis 12: 57–71. . 2003. The Art and Architecture of Thailand from Prehistoric Times through the Thirteenth Century. Leiden: Brill.

04_CreSouthV1_4P_22Jan22.indd 164

. 2004. ‘Esoteric Buddhism in Southeast Asia in the Light of Recent Scholarship’, Journal of Southeast Asian Studies 35/2: 329–354. . 2011. ‘Cambodian Images of Bhaiṣajyaguru’, in Emma C. Bunker and Douglas Latchford (eds.), Khmer Bronzes: New Interpretations of the Past. Chicago: Art Media Resources. . 2015. ‘Aspects of Buddhism in Tenth-Century Cambodia’, in D. Christian Lammerts (ed.), Buddhist Dynamics in Premodern and Early Modern Southeast Asia, pp. 218–260. Singapore: Institute of Southeast Asian Studies.

28/1/22 5:38 PM

Chapter 5

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor Jinah Kim

T

introduction

his chapter explores the religious and political significance of the goddess Prajñāpāramitā during Jayavarman VII’s reign (c. 1181–1220) from trans-regional comparative perspectives. The unique iconography of Angkorian Prajñāpāramitā helps to illustrate her relationship with other Mahāyāna cultic deities such as Avalokiteśvara and Tārā in known textual sources. Here, it is important to underscore that Khmer examples are not ‘outcasts’ in the iconographic genealogy of the goddess Prajñāpāramitā, but innovative yet politically sensitive creations that can help us understand the historical process behind the formation of a religious iconography. While a recent study on Prajñāpāramitā by Multzer o’Naghten (2016) puts a due emphasis on the Mahāyāna characteristics of the goddess Prajñāpāramitā, her importance in the Esoteric Buddhist tenets demands reconsideration.1 Exploring Prajñāpāramitā through a trans-regional lens supports the importance of Esoteric Buddhism in Jayavarman VII’s state operation that scholars like Hiram W. Woodward (1981, 2004) and Peter 1. My own study of Sanskrit manuscripts of the Prajñāpāramitā in mediaeval South Asia and recent studies on the Prajñāpāramitā in tantric Buddhist contexts like Seton (2015, 2017) inform my suggestion to consider Prajñāpāramitā’s importance in the Esoteric Buddhist context. It also echoes Philip Green’s recent proposal (2014) to reevaluate the quintessential Mahāyāna Bodhisattva Avalokiteśvara within an ‘emerging tantric context beginning around the tenth century’ in Cambodia instead of confining the bodhisattva to ‘the overgeneralized category of Mahāyāna Buddhism’.

05_CreSouthV1_9P_11Feb22.indd 167

Sharrock (2009, 2012, 2013) have brought to light. It also helps to locate Angkor in the trans-regional network of Esoteric Buddhism of the 12th–13th centuries across South and Southeast Asia, from Alchi in the Western Himalayas to Java. Prajñāpāramitā is the title of the eponymous text of Mahāyāna Buddhism, and the Prajñāpāramitā related texts have been extensively examined in the Buddhological studies.2 With more studies on the early manuscript corpus from Gandhāra underway, a refined picture of the early historical development of the Prajñāpāramitā as a Mahāyāna text has emerged (Falk and Karashima 2013; Harrison and Hartmann 2014; Harrison 2018). It remains unclear when the Goddess Prajñāpāramitā, the female personification of the famed text, and her worship appeared in the historical horizon. It seems to have been some centuries after the suggested date of the text’s formulation at the beginning of the Common Era.3 The earliest surviving identifiable image of Prajñāpāramitā may be a standing bronze statue of a two-armed goddess holding a manuscript in her left hand (von Hinüber

2. They are too numerous to list here. See Conze (1978) for the list of Edward Conze’s publication on the Prajñāpāramitā. 3. A Chinese pilgrim Faxian’s (ca. 377–422 CE) report of seeing Prajñāpāramitā in worship is taken as evidence of the worship of the goddess in practice in the early 5th century. However, Faxian’s text does not specify that it was an image of the goddess. The object of devotion here may well have been a manuscript. See Kim 2013: 30–31. Also discussed in Shaw 2006, fn. 31.

11/2/22 3:34 PM

168

Jinah Kim

2007) (Fig. 5.10).4 Prajñāpāramitā and her worship Wisdom in eight thousand verses, henceforth AsP), was an important aspect of Pāla Buddhism in India one of the oldest recensions, and the central piece between the 9th and 12th centuries (Kinnard 1999; of the Mahāyāna Buddhist canon, holds an imKim 2013). As an embodiment of a foundation- portant key to locate Angkorian Prajñāpāramitā’s al Mahāyāna Buddhist text and the wisdom ex- Esoteric Buddhist connection in the trans-regional pounded therein, the goddess Prajñāpāramitā is Esoteric Buddhist network. frequently treated as a quintessential Mahāyāna At the outset, it should be noted that my ingoddess (Conze 1948, 1949–50, 1951; Shaw 2006). sistence on comparative trans-regional analysis of Thus, Hedwige Multzer o’Naghten (2016) takes iconographic forms does not mean a return to a cothe appearance of the goddess Prajñāpāramitā as lonial-era generated centre-periphery diffusionist a sole Buddhist female deity in Angkor as suggest- model. A comparative analysis of forms based on ing ‘doctrinal trends of Mahāyāna Buddhism’. We close looking and sustained physical observation of should not forget, however, the importance of the objects can help explore little recognized features Prajñāpāramitā in Indian Esoteric Buddhism—that, of local artistic traditions. By emphasizing formal too, in all of its multivalency: as a text, transcend- analysis, I am not arguing for an alliance with ent insight, mother of all Buddhas, a manuscript, universalist approaches like that taken by André a goddess, and an image (Seton 2015, 2017; Kim Malraux in Le musée imaginaire and Les voix du 2013, 2021). silence. Rather, it is to underscore local aesthetics It is this new perspective on the Prajñāpāramitā and visual strategies and their trans-regional values in the mediaeval Indic context that this study to understand them in a larger network without brings to bear on the discussion of the Angko- invoking a diffusionist model (e.g. Acri 2016). rian Prajñāpāramitā. In particular, it contends that Prajñāpāramitā’s Esoteric Buddhist connecjayavarman vii’s buddhist family tion provides further circumstantial evidence to triad: avalokiteśvara-buddhademonstrate the importance of Esoteric/tantric prajñāpāramitā Buddhism in Jayavarman VII’s Angkor that scholars like Woodward and Sharrock among others In the famous stele of Ta Prohm, the son of Jayavarhave demonstrated (Woodward 1981, 2004; Lobo man VII, Śrī Sūryakumāra, recorded that Jayavar1997; Sharrock 2006, 2007; Estève and Vincent man VII had installed an image of his mother, Śrī 2010). The goddess Prajñāpāramitā’s Esoteric BudJayarājacūḍāmaṇī that was the image of the mother dhist connection has been suggested in previous of Jinas. From this phrase, the mother of Jina, or scholarship, but the connection has been simply jinamātṛ in this Sanskrit inscription, scholars have assumed on the basis of feminine gender, taking the suggested that the central shrine of Ta Prohm was gender and the femininity in a visual representadedicated to the mother of Jayavarman VII in the tion as signalling ‘a sexual and hence Tantric overform of Prajñāpāramitā. No sculpture that could tone’ (Huntington and Bangdel 2003: 124). The be identified as this very image mentioned in the importance of the Prajñāpāramitā teaching in the inscription has been found in situ. The name development of Buddhist Tantras has long been ‘Prajñāpāramitā’ appears often in the discussion acknowledged (Conze 1956; Wayman 1977; Snellof Buddhist inscriptions of Angkor, and it is a fagrove 1987; Seton 2015; Kim 2013). At least one miliar name assigned to a few images of female recension of the Prajñāpāramitā text, the Adhydeities in the Buddhist context of the Angkor ardhaśatikā Prajñāpāramitā (Prajñāpāramitā period in art historical discussions. Yet, the term in 150 verses) is clearly associated with Esoteric Buddhist circles (Linrothe 1999; Davidson 2002; ‘Prajñāpāramitā’ is not used in the Ta Prohm stele Woodward 2015; Bianchini 2020). As we shall see, or in any other known royal inscriptions from the the Aṣṭasāhasrikā Prajñāpāramitā (Perfection of Jayavarman VII’s reign, such as the Preah Khan stele inscription. Instead, her epithets as the mother of the Buddhas are used: i.e. ‘jinamātṛ’ and 4. Oskar von Hinüber (2007) dates this image to the ‘jinanāṃ jananῑ’ at Ta Prohm, and ‘jinanāṃ jananῑ’ early 7th century during the reign of the Palola Ṣāhi King at Preah Khan (Cœdès 1906, stanzas v, xxxvi; 1941, Vajrādityanandin.

05_CreSouthV1_9P_11Feb22.indd 168

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

169

stanza v). After almost two centuries of lacunae in inscriptional evidence, the Ta Prohm stele, which is dated 1186 CE, drops the name ‘Prajñāpāramitā’, which was commonly used in earlier 10th-century inscriptions, and uses only her epithets as the mother of the Buddhas. This usage of the epithets over the actual name seems to have been Jayavarman VII’s deliberate political choice, though his motivation remains unclear. Today, Ta Prohm is well known for its image of an ancient temple taken over by the tropical forest, a battleground between nature and culture, through the tales of tourists and their romantic photos of ruins.5 The importance of Ta Prohm as the temple dedicated to the mother of the Buddha, or Prajñāpāramitā, first came to scholarly attention when George Cœdès put forth his argument about the cult of buddharāja by considering the trinity of Lokeśvara-Buddha-Prajñāpāramitā as being realized in the three state temples of Jayavarman VII, Preah Khan, Bayon, and Ta Prohm in his analysis of the Preah Khan stele (Cœdès 1941, 1989). The Fig. 5.1: Buddhist Triad (Lokeśvara-Buddhastele of Preah Khan reports that Jayavarman VII Prajñāpāramitā), found at Roluos, Siem Reap, consecrated (opened the eyes of) the central statue Cambodia, ca. late 12th–early 13th century. Gilded of Lokeśvara named Jayavarmeśvara that was an bronze. H: 49.5 cm. National Museum, Phnom Penh, image of his father (pitṛmūrti). The contents of Ga2424. the inscriptions from Ta Prohm and Preah Khan are similar, and the first thirty-six lines of ‘Face As observed by Woodward (2015), the idea A’ of the Ta Prohm stele are verbatim repeated in of the Avalokiteśvara (Lokeśvara)-Buddha-Prathe Preah Khan stele, also on ‘Face A’. The two jñāpāramitā (henceforth ABP) trinity seems to steles are almost identical in appearance: along have been everywhere early in the reign of Jayavarwith similar arrangements of text in two columns, man VII. The trinity is invoked at the beginning of both feature a moulded base, stylized lotus flower most royal inscriptions. In addition, a good number decoration on the top, and are almost identical in of images depicting the trinity survive from the size (Ta Prohm: 60 cm by 200 cm; Preah Khan: 58 period, and artistic calibers and common stylistic cm by 185 cm). The similarity between the two intraits evident in these sculptures suggest a concertscriptions suggests deliberate planning of these two ed royal patronage of this type. One such example temples dedicated to his mother and father. Cœdès is the Prei Monte temple bronze sculpture of the proposed that the huge Buddha of the Bayon state triad now in the Phnom Penh National Museum, temple found shattered in the central pit of the datable to Jayavarman VII’s reign based on its style temple must have been the missing Buddha of the and iconography (Fig. 5.1). It is a sizeable bronze trinity that reflects the projection of Jayavarman piece nearly half a meter in height (h. 49.5 cm, w. VII’s self-image as Buddharāja or the Buddha-king 41.5 cm, d. 16 cm). Each deity and the pedestal born of the mother Prajñāpāramitā and the father were cast separately and assembled together.6 The Lokeśvara. 5. It is famous as the set for the film ‘Lara Croft: Tomb Raider’ (2001), directed by Simon West.

05_CreSouthV1_9P_11Feb22.indd 169

6. On the material composition and technique of metallurgy, see Bougarit et al. (2003) and Vincent, Bourgarit and Jett (2012). A recent study (Polkinghorne et al. 2014)

11/2/22 3:34 PM

Jinah Kim

170

Buddha in the centre sits in meditation on a lotus seat installed on a serpent’s coil. The seven-headed serpent’s dramatic hood rises above him in a shape that may evoke the Bodhi tree and its leaves under which the Buddha attained enlightenment. Lokeśvara stands firmly with his robust legs and holds a rosary, a book, a lotus, and a flask in his four hands. Amitābha in front of his matted hair is visually pronounced. Prajñāpāramitā, or the female deity who holds a lotus bud in her left hand and a book in her right, stands on the left side of the Buddha. All three of them feature downcast eyes and thick lips with gentle smiles, one of the distinct stylistic features of Jayavarman VII’s sculptural workshops. They collectively exude serenity. In its original state, this object would have been entirely gilded (i.e. fire gild on unleaded copper alloy, see Vincent, Bourgarit and Jett 2012: 140), suggesting splendour and import assigned to such a piece.

precedents of the abp triad The ABP triad became one of the most prominent types of images produced and installed in Angkor by the end of the 12th century, perhaps because of Jayavarman VII’s propagandic strategy. The idea of the ABP triad, however, was not Jayavarman VII’s invention. It was already in circulation in Angkor at least two centuries before Jayavarman VII’s time. There survives an earlier 10th-century sculptural representation of a triad. A stone stele in Honolulu shows an interesting triad configuration with the same three-some: Avalokiteśvara, Prajñāpāramitā, and the Buddha (Fig. 5.2). The number of arms and the postures held by each figure remain identical: four-armed Avalokiteśvara, two-armed Prajñāpāramitā, both standing, and Buddha seated in meditation. The composition of the triad differs considerably from Jayavarman VII’s ABP triad. In the Honolulu stele, the Buddha sits above a bejewelled pillar that rises in between Avalokiteśvara and Prajñāpāramitā, as if to articulate the Esoteric Buddhist conception of enlightenment rising from the union between compassion (Avalokiteśvara) and wisdom (Prajñāpāramitā). The emphasis here explores the importance of bronze casting at Angkor with an archaeological identification of a bronze casting workshop within the central area of Angkor Thom.

05_CreSouthV1_9P_11Feb22.indd 170

Fig. 5.2: Stele with Buddha, Lokeśvara and Prajñāpāramitā, c. 10th century. Bakheng style. Cambodia. Pinkish sandstone. Collection of the Honolulu Museum of Art. Gift of John Young, 1991 (6691.1). Photo courtesy of the Honolulu Museum of Art.

is the path signalled by the two standing figures, Avalokiteśvara and Prajñāpāramitā, the importance of which in 10th-century Angkorian Buddhism are well attested in art historical and epigraphic evidence.7 In comparison, the composition of Jayavar7. Nancy Dowling (1996) argues for the triratna (three jewels: Buddha-Dharma-Sangha; B-P-A respectively) interpretation of the stele against Woodward (1981)’s Esoteric Buddhist argument based on the Mahāvairocanasūtra. While Dowling’s effort to localize the interpretation based on epigraphic sources is laudable, perhaps her insistence on Mahāyāna interpretation is too generic and inadvertently erases the dynamism of Angkorian Buddhism that paralleled rapidly developing Esoteric Buddhist methods and means being incorporated into Buddhist practices elsewhere. Also, see Hiram Woodward (2007, 2015) and Philip Green (2014) for the importance of the Kāraṇḍavyūhasūtra, which puts Avalokiteśvara in the central position in Angkorian Buddhism.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

Fig. 5.3: Avalokiteśvara-Tārā with Buddha in earthtouching gesture, bronze, Kurkihar, Bihar, India, c. late 9th–10th century. Patna Museum, India. Photo © Jinah Kim.

man VII’s ABP triad emphasizes the fruit, the completed state, the Buddhahood that was harnessed in conceptualizing Jayavarman VII’s kingship. As a precedent to Jayavarman VII’s triad, it is important to recognize that the Honolulu stele represents a unique Khmer artistic interpretation of the importance of compassion and wisdom in Buddhist practices. Such a configuration is rarely seen in Buddhist art of India where a triad was almost always composed of male members, i.e. Buddha with two male bodhisattvas or attendants. The conception of the ABP triad appears in Angkorian epigraphic sources from the 10th century onwards, but this formulation remains unidentified in known contemporary textual sources as Woodward (2015: 231) points out.8 A rare example depicting male and female bodhisattvas as a couple with the Buddha rising above in 8. Woodward characterizes the triad of ABP as an issue that ‘cannot be easily resolved’.

05_CreSouthV1_9P_11Feb22.indd 171

171

the middle survives among the hordes of bronzes found at Kurkihar, an important Buddhist site in Bihar that seems to have catered to monks and pilgrims from South India, like Kañci and Malaya country (malayadeśa) (Pal 1988; Kim 2013: 232; Bautze-Picron 2015: 17). A bronze sculpture now housed in the Patna Museum (Acc No. 9642, h: 15.5 cm) renders Avalokiteśvara and Tārā both seated while displaying varadamudrā in their right hands and holding their respective lotuses, padma and utpala, in their left hands, in harmony (Fig. 5.3).9 The two deities are framed by a single throne-back with double openings. This framing piece was cast separately and soldered behind their heads. On this framing structure, rising in between them on a lotus seat sits Buddha in earth-touching gesture (bhūmisparśamudrā), signalling the epitome of enlightenment. A halo with flaming outer rim surrounds this Buddha. A bejewelled umbrella above him marks the peak of the sculpture while bodhi leaves drape like tendrils on either side. Two stupas on double-lotus seats are carved on the flat bar across the throne back, as if doubly confirming the Buddhist identity of the iconography. The image bears the so-called ‘Buddhist creed (ye dharma hetu…)’ inscribed on the back of the throne and a now abraded donor inscription on the front of the pedestal (Akhtar 2001: 193). The Buddha on a lotus seat is positioned on top of a vertical bar in between the couple, displaying a similar composition to the Honolulu stele. The relatively smaller size of the Buddha in comparison to the couple below is also a shared feature. The Kurkihar bronze may date to the 10th century.10 There is no direct connection between the Kurkihar bronze and the Angkorian stele in Honolulu, except for tenuous formal ones. Yet, the unique Cambodian formulation begs us to look at a group of Indian sculptures with a fresh perspective. There survives a small, somewhat enigmatic corpus of sculptures in India that represent a similar triadic formulation of male-female couple with a transcendental figure rising above in the 9. Bautze-Picron 2015: B40. 10. The sculptural quality is relatively crude in comparison to better finished and defined bronze sculptures from the Kurkihar hoard of the 9th century.

11/2/22 3:34 PM

172

Jinah Kim

Fig. 5.4: A Jain Family Group, Uttar Pradesh, India, c. 550–600 CE. Cream-coloured sandstone, 50.8 × 28.57 × 10.79 cm. Los Angeles County Museum of Art, Gift of Mr and Mrs Harry Lenart, M.77.49.

centre. Most of these images are identified in a Jain context, which rarely finds counterparts outside the South Asian subcontinent. They are often described as a ‘family group’. In a stone stele from northern India now in the Los Angeles County Museum of Art (M.77.49), for example, two large figures, male and female, sit in the centre with a tree rising above them on top of which sits a meditating Jina (Fig. 5.4). The identity of the couple remains unresolved between ‘parents of the Jina’, and a yakṣa-yakṣῑ couple who are the protective attendants (Shah 1987: 47–52). The triangular formulation with a couple and their ‘offspring’ in the centre above them seems to have become popular in Jain circles during the 10th century as seen in the bronze image (copper alloy) from eastern Maharashtra (Rajnapur Khinkini), now in the Nagpur Central Museum (Fig. 5.5).

05_CreSouthV1_9P_11Feb22.indd 172

Fig. 5.5: Yakṣa couple, c. 10th century. Rajnapur Khinkini, Maharashtra. H: 23.2 cm. Nagpur Central Museum, CMN/C/148/10. Image: Regents of the University of Michigan, Department of the History of Art, Visual Resources Collections.

According to Lisa Owen (2013), a triadic formulation with Jain yakṣa (Kubera) and Jain yakṣῑ (often called Ambikā) flanking the solitary Jina Tīrthaṅkara is found in later additions (‘intrusive panels’) to the Jain caves at Ellora, mostly carved during the mid-10th century.11 The significance of this iconographic formulation merits further investigation. As a parallel, the Avalokiteśvara-Prajñāpāramitā couple in the Honolulu stele suggest an alternate meaning of the yakṣa-yakṣī couple in Jain examples, not only as the guardians of Jain teaching 11. Owen (2013: 114–115) considers these additions as signalling the continued ‘growth and development of Jain iconography’.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

173

(śāsanadevatās) but also as an important conduit priated the concept of the earlier triad and that for practitioners in their pursuit of the higher state of Prajñāpāramitā as the mother of Buddhas, as of being. At least in the Buddhist context, it may evidenced by the exclusive use of her epithet of relate to the circulation of certain Tantras that ar- ‘Buddha-mother’. Her ‘motherly’ identity seems ticulate the spiritual relationship through familial to have nearly eclipsed her doctrinal identity as terms (kula), especially those that introduce the the personification of the teaching (dharma) in language of the union of gendered principles in the Jayavarman VII’s inscriptions at Ta Phrom and spiritual quest, with prajñā identified as the female Preah Khan. force, as in the Guhyasamāja (Wayman 1973: 6–11; Linrothe 1999: 147, 225). development of the prajñāpāramitā Going back to the political significance of iconography Jayavarman VII’s ABP triad, inscriptional evidence also suggests a modified formulation of the rela- Surviving art historical records suggest images tionship among the trinity during Jayavarman VII’s of the goddess Prajñāpāramitā came into existtime. For example, both Ta Prohm and Preah Khan ence centuries after the initial formulation of the steles invoke Buddha, Dharma, and Sangha in the Prajñāpāramitā texts.14 When, how, or why this first three stanzas before invoking Lokeśvara and transformation happened remains an art historical Prajñāpāramitā. This is a big change from the late puzzle. The iconography of Prajñāpāramitā was 10th-century inscriptions where Prajñāpāramitā fluid and experimental in its early stage of develand Lokeśvara were invoked in place of Dharma opment during the 7th through the 9th centuries and Sangha respectively.12 What is visually insinu- in South Asia. The most common attribute that ated in the Honolulu stele, that is, the conception of leads modern scholars to identify a female deity as Lokeśvara and Prajñāpāramitā as a couple bearing Prajñāpāramitā is a book; not a codex, but a pothi the Buddha was fully embraced and appropriated format manuscript.15 A pothi manuscript was often for political ends during Jayavarman VII’s reign. made with palm-leaf or birch bark cut to uniform In other words, the ABP triad prepared during size, and depicted as a narrow rectangular bar in Jayavarman VII’s reign is no longer a representa- visual representations. A book as an iconographic tion of triratna.13 Rather, as a visual representation attribute signals wisdom (prajñā) or knowledge of the path to the Buddhahood, the emphasis is (vidyā) in ancient and mediaeval South Asian art. Two main modes of presentation of this attribin the nuclear family relationship, father-mother-son with the focus on the son. In the Honolulu ute exist in the Prajñāpāramitā iconography: in one, stele, Prajñāpāramitā and Lokeśvara hold a more the book is held in the hand of the goddess as seen central position signalled by their firm stance and in Angkorian examples, and in the other, the book the prominent size in comparison to the Buddha is placed above an utpala (water-lily, nymphea) who is supported by the post. This Buddha, the that she holds. The latter strategy appears at Ellora son of their union, is the protagonist of the triad in India where we find early surviving examples of of the late 12th century in terms of his size and Buddhist goddesses that may signal changing docpresence. It seems that Jayavarman VII appro- trinal and ritual needs of Mahāyāna and emerging 12. For example, the stele of Phnom Banteay Nan invokes Buddha, Lokeśvara and Prajñāpāramitā in the first three stanzas. IC, VI: 202–206. The inscription of Thma Puok also invokes Buddha, Prajñāpāramitā and Lokeśvara in the first three stanzas. IC, III: 66–67. 13. It is noteworthy that a slightly modified triad of ‘Prajñāpāramitā-Buddha (Śākyamuni)-Ṣaḍākṣari Lokeśvara’ serves as an important visual representation of triratna in post-16th century Nepal. See Woodward 2007, fn. 51; Kim 2019: 106.

05_CreSouthV1_9P_11Feb22.indd 173

14. Faxian’s (337–c.422) report of seeing Prajñāpāramitā in worship is often taken as evidence of the existence of the image of the goddess Prajñāpāramitā in the 4th century as recently cited by Multzer o’Naghten (2016), but the passage in question does not necessarily mean there was an image of the goddess (Kim 2013: 31). 15. Pothi means book in Hindi, and especially refers to a manuscript in the traditional ‘Indic’ format as described. It is used here for convenience of reference.

11/2/22 3:34 PM

174

Jinah Kim

Fig. 5.6: Prajñāpāramitā and Avalokiteśvara. Intrusive panel outside Cave 10 second story balcony area, Ellora, ca. 7th century. Photo © Jinah Kim.

Fig. 5.7: Tārā or Bhṛkuṭī and Buddha, Cave 3A Ellora, c. 7th century. Photo © Jinah Kim.

Esoteric Buddhist tenets.16 What may be identified as an early rendition of Prajñāpāramitā with a book on lotus is carved outside Cave 10, standing side by side with Bodhisattva Avalokiteśvara (Fig. 5.6). The robust standing female deity is as tall as Avalokiteśvara, and their relationship is that of equal partners. The status of female deities that appear in intrusive panels at Ellora is not subordinate to their male counterparts. For example, another unusual arrangement in Cave 3A at Ellora shows a female deity holding a fully blown lotus (i.e. padma, perhaps representing an early form of the quintessential Buddhist Goddess, Tārā before the iconographic codification, or Bhṛkuṭῑ,

if the object in her right hand is a rosary and if her damaged diadem held a stūpa) standing tall next to a standing Buddha (Fig. 5.7).17 Looking eastwards to the next stage of iconographic development, we note that the book on top of an utpala became a ubiquitous marker of the Wisdom related deities in Pāla Buddhist art, including Mañjuśrῑ, Cundā, and Prajñāpāramitā. A number of examples from Magadhan sites depict the goddess Prajñāpāramitā holding an utpala with a book on top as seen in a bronze sculpture from Nālandā (Fig. 5.8). The book on a lotus became established as a visual code to signal wisdom in Buddhist iconography, including that of Mañjuśrῑ as bodhisattva of wisdom. In a

16. Ellora is where Linrothe (1999) finds early visual records signalling ‘Phase One’ Esoteric Buddhism. Along with the krodha-vighnāntaka imagery that Linrothe uses to measure the stages of historical development of Esoteric Buddhism, images of goddesses can be another yardstick that can help investigate the ritual and material history of Indian Esoteric Buddhism.

17. The designation of Cave 3A follows Malandra’s (1993: 39). Although the image appears in Cave 4 iconographic plan in her book, it is located in the connecting cell between Caves 3 and 4 called Cave 3A. Malandra (ibid.: 40) describes this configuration as ‘carved, but without apparent connection among them, an indication of the disrupted nature of these subsidiary areas in the Cave 2 through 4 group.’

05_CreSouthV1_9P_11Feb22.indd 174

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

175

Fig. 5.8: Prajñāpāramitā, c. 9th century. Bronze. Nālandā. ASI Site Museum, Nālandā. Photo © Jinah Kim.

way, this feature distinguished Buddhist wisdom deities from those of other religious traditions. A book held in hand is a more ancient visual strategy that was used to signal the connection with knowledge, as the earliest surviving image of the goddess Sarasvatī found in the Jain context at Mathura, India, demonstrates.18 Sarasvatī’s iconography consistently depicts a book in the Goddess’ hand in mediaeval Jain and Hindu contexts, as seen in a circa 10th-century copper alloy sculpture of Sarasvatī found in eastern Maharashtra (Rajnapur Khinkini), now in the Nagpur Central Museum (CMN/C/148/11) (Fig. 5.9).19 A 7th-century bronze 18. This is the well-known image in the State Museum in Lucknow, Acc. No. J.24. Most recently published in Kim 2015, fig. 24. 19. It is interesting to note that Jina in meditation floats above her, signalling an important connection between wisdom and meditation, which, as we will see, is also emphasized in Angkorian Prajñāpāramitā images.

05_CreSouthV1_9P_11Feb22.indd 175

Fig. 5.9: Sarasvatī, c. 10th century. Rajnapur Khinkini, Maharashtra. Bronze (copper alloy). 23.2 cm. Nagpur Central Museum CMN/C/148/11. Image: Regents of the University of Michigan, Department of the History of Art, Visual Resources Collections.

image of Prajñāpāramitā from Gilgit, possibly the earliest surviving image that can be confirmed as the goddess Prajñāpāramitā through inscription also holds a book in her left hand (Fig. 5.10).20 In Pāla period Buddhist images, this mode is followed in some multi-armed images of Prajñāpāramitā and Cundā, shown as holding a book in one of their hands. A two-armed image of Prajñāpāramitā with a book in her hand as seen in Jayavarman VII’s ABP triad, however, is rare among Pāla period examples. A unique stone sculpture found at Nālandā provides us a glimpse into an early formulation of the goddess Prajñāpāramitā’s iconography that 20. See the discussion of the inscription and the date in von Hinüber 2007.

11/2/22 3:34 PM

176

Jinah Kim

Fig. 5.11: Prajñāpāramitā (?) with four bodhisattvas, ca. 8th century. Site No. 3, Nālandā. ASI Site Museum, Nālandā. Photo © Jinah Kim.

Fig. 5.10: Prajñāpāramitā, c. 7th or 8th century. Gilgit, Pakistan. Copper alloy with copper and silver inlay. H: 41 cm. Private collection. Photo © Christian Luczanits.

adopted this book-in-hand feature to signal her motherly identity as articulated in the text of the Aṣṭasāhasrikā Prajñāpāramitā [AsP]. A circa 8th-century stone sculpture of a goddess surrounded by a retinue of male figures (Fig. 5.11) was found in the southeast corner of site No. 3 at Nālandā during the early years of excavation (Sastri 1986: 38). This image has been variously identified as Koṭiśrῑ, Mahāsarasvatῑ, and Vajraśāradā (Bhattacharyya 1968: 351; see Asher 1980: 82–83). Here, I would like to suggest that this image is an early attempt to capture plastically the image of the Prajñāpāramitā as the ‘mother of bodhisattvas, the great beings (mātā bhagavan bodhisatvānām mahāsatvānām prajñāpāramitā)’, as sung by Śāriputra in Chapter VII of the AsP (Conze 1973: 135). A

05_CreSouthV1_9P_11Feb22.indd 176

benign, two-armed goddess bedecked in jewels and ornaments sits on a lotus with her legs crossed at the ankle. Surrounding her are four male figures, also seated on lotuses that sprout from the same mother stalk as the goddess. The four are differentiated in their hair styles and attires, two of them featuring antelope skin around their chests and loftily matted hair, similar to those seen in contemporary images of Avalokiteśvara and Maitreya. Interestingly, these male figures all hold a book in their left hands.21 Given the positioning of her broken left arm, it is very likely that the goddess also held a book in her left hand like the spiritual ‘sons’ that surround her (Asher 1980: 82). The two snakes (one is damaged) that peak out from underneath her lotus seat also signals a connection between this plump goddess and the Prajñāpāramitā, especially given the oft-cited story of nāgas guarding the Prajñāpāramitā until Nāgārjuna rescued it for the hu21. The top right corner image is damaged and we cannot see the left hand, but it is most likely that this image also held a book in his left hand like his brothers.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor mankind.22 This form does not seem to have been a successful model for the subsequent development of the iconography of Prajñāpāramitā, perhaps owing to the fact that she does not appear so different from other goddesses, including contemporary images of Saptamātṛkās (Seven Mothers), like those in Ellora’s Hindu caves.23 If this image represents an early effort to render the goddess Prajñāpāramitā as the mother to all bodhisattvas, this two-armed form with a book in hand certainly emphasized her motherly quality surrounded by her spiritual sons, in this case, bodhisattvas. This is worth remembering as we consider Prajñāpāramitā images in Angkor, especially those made during Jayavarman VII’s reign. Angkorian examples typically identified as the goddess Prajñāpāramitā look quite different from the iconographic types known in visual records from Indian monastic sites and those recorded in sādhana texts of the late 11th century. For example, the Prajñāpāramitā image in the ABP triad of Jayavarman VII’s reign differs considerably from most typical images of the goddess Prajñāpāramitā known in India. Two-armed and standing tall, she holds a book firmly in her right hand and a lotus bud in her left. Both hands are held low along the hip, projecting slightly forward. In contrast, the gesture of preaching (dharmacakrapravartanamudrā, developed from the iconography of the first sermon) is the most typical mudrā shown in the images of Prajñāpāramitā from the 9th century onwards in India as seen in the stone stele of Prajñāpāramitā now in the Asian Art Museum, San Francisco (Fig. 5.12). Although the found spot of the image is unknown, it is likely to have been prepared at an important monastic site in Magadha 22. The image was found with a larger Nāgarāja image. Asher (1980: 82) suggests that these two images were strategically placed as guardians of the transcendent wisdom. I suggest that they were not just guardians but rather the very representation of the transcendent wisdom as the mother with her protector/promulgator. This image of the goddess presaged a vision of Prajñāpāramitā that became important in subsequent centuries while the Nāgarāja image planted a seed for the image of Nāgārjuna in later Tibetan Buddhist iconography. 23. In fact, the Nālandā image appears remarkably similar to the mātṛkās images in Cave 21 at Ellora.

05_CreSouthV1_9P_11Feb22.indd 177

177

Fig. 5.12: Goddess Prajñāpāramitā, Bihar (Nālandā?), India, c. 825–875 CE. Asian Art Museum, San Francisco B62S32.

(today’s Bihar, India). This iconography travelled far beyond the Magadhan Buddhist centres. From the 9th-century metal sculpture of Prajñāpāramitā of Sri Lankan origin reportedly found in southern Thailand (Fig. 5.13) to the famous royally sponsored Siṅhasāri Prajñāpāramitā of ca. 1300 CE from Candi Singhosari, East Java (Museum Nasional, Indonesia, Jakarta), the two-armed goddess is shown displaying the gesture of preaching. Yet, this gesture is absent in all surviving Angkorian examples, whether two-armed or multi-armed.24 Even in ten-armed or twenty-two armed forms, 24. We should also note that this gesture of preaching is conspicuously absent in Khmer Buddhist images. In fact, the earth-touching gesture that becomes ubiquitous in Pāla period Buddha images does not seem to have a huge impact on the production of Buddha images in Angkor. The absolute preference especially of the Jayavarman VII’s era is the gesture of meditation, signalling both cultural confidence and conservatism of image makers and knowledge experts behind iconographic choices.

11/2/22 3:34 PM

178

Jinah Kim

Fig. 5.13: Prajñāpāramitā, Sri Lanka, probably Polunaruwa, ca. 9th century, reportedly found in southern Thailand. Copper alloy with traces of gilding, 15 × 13 × 7 cm, V&A, IS.21-1988. Image © Victoria and Albert Museum, London.

the main hands are held apart, low by the hip, each holding a book and a lotus bud. In Conze’s survey of Prajñāpāramitā images, only the Angkorian examples feature as many as twenty-two arms, the highest number of arms known in the corpus of surviving Prajñāpāramitā images, and they are not recorded in any known textual source (Conze 1949–50: 51). The twenty-two armed, eleven headed (1+4+6) standing image of the goddess in bronze that hails from 12th or 13th century Angkor does not look anything like the preaching goddess of wisdom represented through the Prajñāpāramitā iconography codified and circulating widely in other parts of the Buddhist world by this time (Fig. 5.14). We would have been left puzzling over the identity of this goddess, if it was not for the inscription on the pedestal that reads, ‘Vraḥ rūpa vraḥ Prajñāpāramitā’, literally an ‘holy image (form) of holy Prajñāpāramitā’ (Cœdès 1920: 7–8). Khmer Prajñāpāramitā images may be ‘outside the mainstream’ as Conze characterizes them, but at the same time, their stylistic and

05_CreSouthV1_9P_11Feb22.indd 178

Fig. 5.14: Twenty-two armed Prajnāpāramitā, Angkor, late 12th–13th century. After G. Cœdès, ‘Note sur une statuette Cambodgienne de la Prajñāpāramitā’, BEFEO 20 (4) (1920): 7–8.

iconographic uniqueness indicates the level of confidence on the part of Khmer knowledge experts and image makers that produced these images following their own, local visual language. More importantly, the unique iconographic formulation in Angkorian Prajñāpāramitā images can help unpack the little understood process behind the formation of a divine iconography in a doctrinally complex environment.

meditating buddha on the headdress One of the common features in Angkorian Prajñāpāramitā images is the presence of the meditating Buddha on the headdress. Even in a mul-

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

179

Fig. 5.15: Twenty-two armed Prajñāpāramitā, c. late 12th–early 13th century. National Museum, Phnom Penh Ga5333.

ti-armed, multi-headed example, like the inscribed standing example discussed by Cœdès, an image of a Buddha in meditation appears on the apex of the stacked heads. As seen in a seated, twenty-twoarmed, eleven headed metal image of the goddess now in the National Museum Phnom Penh, the Buddha in meditation appears in front of the conical shaped hairdo, as if affixed on a diadem or a crown (Fig. 5.15). Who is this Buddha and why is he on the headdress of Prajñāpāramitā? What is the meaning of this small insignia? The Buddha on the headdress is adopted to signal and codify a spiritual lineage system in Esoteric Buddhist iconography.25 In Sanskrit sādhana texts that record the forms of Prajñāpāramitā invoked in visionary practices in the 11th-century Indic Buddhist world, Prajñāpāramitā is known with two types of headgear: bearing either the five tathāgatas or the 25. The earliest instance of this formulation may be the meditating Buddha Amitābha on the headdress of Avalokiteśvara images.

05_CreSouthV1_9P_11Feb22.indd 179

image of Akṣobhya who holds the earth-touching gesture.26 Yet, even among Indian examples, this particular insignia is absent in most surviving artistic evidence representing the goddess.27 The 26. Woodward (2015: 248) glosses over this discrepancy of the jina on the crown and does not explain why it is a meditating Buddha rather than Akṣobhya when he characterizes that Jayavarman VII’s reign’s two-armed Prajñāpāramitā images developed in accord ‘with the Indian textual description’. Indian sādhana texts are discussed below. 27. Here is another good example to warn against wholesale adoption of sādhana texts for explaining surviving images from different periods and contexts. Prajñāpāramitā’s connection with Akṣobhya in these sādhanas has more to do with her essential connection with the enlightenment, as the progenitor of enlightenment, which is most powerfully represented through the image of Buddha touching the earth to call to witness this victorious moment. The Buddha’s awakening experience at dawn under the bodhi tree that is encapsulated in this iconography informed the iconographic formulation of Akṣobhya: his association with eastern direction and the colour blue suggest this narrative connection. See Kim 2021: 232–233.

11/2/22 3:34 PM

180

Jinah Kim

meditating Buddha in the headdress signals Ami- tacharyya (1968: 222). Although Multzer o’Naghten tābha especially in the five tathāgata system of (2016: 43) mentions this text as one of the twen‘Phase Two’ Esoteric Buddhist texts like the Sar- ty-nine Tārā sādhanas in the Sādhanamālā, strictly vatathāgatatattvasaṅgraha. The confident icono- speaking, it is not a sādhana but a vidhi (instrucgraphic expression seen in Khmer Prajñāpāramitā tion for a rite—in this case, stutividhi for adulaimages urges us to explore further the potential tion of the goddess), based on one Sragdharāstuti, connection between Amitābha and Prajñāpāramitā more precisely the Āryatārāsragdharāstotra (lit. and its significance. I propose two possible expla- ‘Hymn to the noble Tārā, in sragdharā metre’) by nations, one historical and the other more liturgical a Kashmiri master (mahāpaṇḍita) Sarvajñāmitra and ritual-based. (Bhattacharyya 1968: 222–224). The date of the If the Buddha in the headdress is indeed Ami- author, Sarvajñāmitra, has been suggested as the tābha, it underscores the indebtedness of Pra- 8th century based on the appearance of his name in jñāpāramitā iconography to Mahāyāna cultic the 12th-century Kashmiri chronicle Rājataraṅgiṇῑ deities, especially the bodhisattva of compassion, (Vidyabhusana 1908: xxx). If the 8th-century date Avalokiteśvara. Amitābha on the headdress is a of Sarvajñāmitra can be entertained, stories about typical iconographic feature of Avalokiteśvara, es- him remembered in the later Tibetan sources, like tablished from at least the 4th century onwards, as Tāranātha’s History of Indian Buddhism and the suggested by both textual and art historical evi- Pagsam-jon-zang, provides one window to condence (Bautze-Picron 2004: 228; Studholme 2002: sider trans-regional movement of visions of the 14–15). Early images of Prajñāpāramitā were most goddess. Tibetan stories about Sarvajñāmitra recall likely modelled after images of ‘mother goddesses’ that he was a Buddhist monk from Kashmir who known in the Brahmanical context, such as the resided at Nālandā in the 8th century whose deSaptamātṛkās (seven mothers), just as images of votion to Tārā saved many lives. His vidhi also the archetype Buddhist goddess Tārā may have mentions Prajñāpāramitā as an epithet of a twobeen modelled after Hindu goddesses (Ghosh armed form of Tārā. In a way, a Kashmiri Buddhist 1980: 26; Shaw 2006: 313). Whatever her icono- monk Sarvajñāmitra provides a tantalizing thread graphic model, Tārā is unquestionably tied to to connect rather different early articulations of Avalokiteśvara, either as an attendant or a female the goddess Prajñāpāramitā, one in the Kashmir articulation of the same principle, i.e. compassion region (Gilgit) and the other at Nālandā. Although (as personification of Avalokiteśvara’s tears of the relationship between Tārā and Prajñāpāramitā compassion). The iconography of Aṣṭamahābhaya especially in terms of their respective historical Tārā (Tārā as a saviouress from eight great perils), origin and subsequent iconographic development for example, is compositionally nearly identical is out of the scope of the present chapter, it is imwith the iconography of Avalokiteśvara in the portant to note continued overlaps between the same role.28 In later Indian Buddhist sādhanas, two goddesses (such as the common iconographic Tārā is categorized under Amoghasiddhi’s family, attribute of utpala) even after the introduction of but various origin stories of Tārā recorded in the a more defined iconographic form of the goddess Tibetan cannon recall her deep connection to Ava- Prajñāpāramitā at Magadhan Buddhist sites and beyond. As Christian Luczanits (2004: 216) notes, lokiteśvara (Shaw 2006: 307). In the Sādhanamālā, a late 11th-century com- in the Sumtseg at Alchi, a western Himalayan Budpilation of Sanskrit sādhana texts, there is a single dhist site with a contemporaneous construction occasion in which Tārā is envisioned as holding date to Jayavarman VII’s reign, Prajñāpāramitā and Buddha Amitābha on her hair (amitābhabuddha- Tārā are not completely distinct from each other madhyām jaṭikām), which is number 109 in Bhat- with a lot of overlaps and shared features. As seen in Fig. 5.6, an early image of Prajñāpāramitā is paired with an image of Ava28. Tārā effectively replaces him in this role at various lokiteśvara outside Ellora’s Cave 10, hinting at a medieval Buddhist sites in eastern India (e.g., Odisha) and devotional milieu in which images of the goddess also in the Himalayas.

05_CreSouthV1_9P_11Feb22.indd 180

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor Prajñāpāramitā may have been formulated.29 In a way, Amitābha on the headdress of Angkorian Prajñāpāramitā images may be taken as a point of reference to the undocumented devotional cultic milieu in which early Prajñāpāramitā images were shaped along with images of Tārā and Avalokiteśvara. If this hypothesis can be entertained, we may suggest that the unique Angkorian Prajñāpāramitā iconography remembers an older layer of iconographic development before codification of iconographic types that ushered in at Buddhist monastic sites in eastern India during the 11th century.

amitābha and prajñāpāramitā Angkorian Prajñāpāramitā images record a relationship between Amitābha and Prajñāpāramitā that no surviving text seems to testify. An intriguing clue to understand this relationship that seems to have been well understood in 12th- and 13th-century Angkor is found in manuscripts of the Prajñāpāramitā texts, especially of the AsP, prepared in 12th-century eastern India. As discussed elsewhere, there are at least eight dated 12th-century painted manuscripts of the AsP that feature Amitābha and Prajñāpāramitā facing each other on the first two folios (Kim 2021: 124, Table 3.1). For example, in a late 12th-century AsP manuscript sponsored by a lay donor now in the Asiatic Society, Mumbai (Kim 2013, Web 5-4), the central panel of folio 1 verso houses an image of the Buddha Amitābha, clearly identifiable by his meditation gesture and his red colour, being adored by four monks.30 The facing panel of folio 2 recto is a two-armed seated image of the goddess Prajñāpāramitā, holding her hands in front of her chest in preaching gesture with two lotus blossoms on either side of her shoulder, each topped by a manuscript. This pairing, which is not attested in any known text, seems rather unexpect29. In fact, outcropping of goddess images and inconsistency and variation in iconographic details suggest a strongly experiential nature of goddess worship at Ellora. 30. For published images of some of these examples, also see Kim (2013), figs. 5-7, Web 2-1, and Web 5-6. The ‘Web’ figures are available online at https://www.ucpress.edu/ book/9780520273863/receptacle-of-the-sacred.

05_CreSouthV1_9P_11Feb22.indd 181

181

ed given the close proximity between the Sanskrit sādhanas and the images of Buddhist deities that appear in painted AsP manuscripts from eastern India (Kim 2013, 2021).31 Eastern Indian manuscripts that feature the prominent pairing of Amitābha and Prajñāpāramitā may not directly explain the enigmatic relationship between Amitābha and Prajñāpāramitā in Angkorian Prajñāpāramitā images, but further consideration of the ritual context in which these manuscripts were used help shed light on it. As I have suggested elsewhere (Kim 2013, 2015), these manuscripts were prepared in ‘deluxe’ editions with painted panels to serve ritual needs of Esoteric Buddhist communities in eastern India. As seen in a stele of the goddess Prajñāpāramitā from the eastern coastal state of Odisha in India, a manuscript, represented as a horizontally shaped bar on a single legged pedestal in the centre of the bottom register, is a central object of a ritual attended by a ritual master who holds a vajra and a ghaṇṭa—a set of quintessential Esoteric Buddhist ritual implements—in his hands with offerings next to him while a group of lay devotees pay homage to the book from the opposite side (Fig. 5.16). This stele most likely dates to the late 11th century or the 12th century, given the exceptionally deep carving and other stylistic features, such as the delicately carved treatment of the tri-lobed architectural framing device around the goddess, as well as the clear articulation of a book-centred ritual scene.32 The manuscript placed in the centre of the ritual scene depicted with flower offerings on top is most likely a beautifully prepared manuscript of the Prajñāpāramitā. In addition to a clear reference to a book as a central object of a ritual, other iconographic elements suggest that the image 31. Pairings of other deities in this manuscript are predicated by their consort relationships known in texts for tantric visionary practices, like Saṃvara with Vajrayoginī, and Heruka with Nairātmyā. 32. Images from Odisha are difficult to date with certainty as very few examples survive with a secure date. Thomas Donaldson’s survey of Buddhist sculptures of Odisha is extremely useful, but his dates are not based on any serious stylistic analysis or art historical investigation. The appearance of the book-centred ritual scene with a ritual master suggests a post-11th century date as observed by Claudine Bautze-Picron (1995).

11/2/22 3:34 PM

182

Jinah Kim

Fig. 5.16: Stele of the Goddess Prajñāpāramitā with a ritual scene on the bottom register, Mangalpur (Balasore district), Odisha, 11th century, Odisha State Museum. Photo © Sonali Dhingra, 2016.

was made within a community that also understood a close relationship between Amitābha and Prajñāpāramitā. Surrounding the goddess above the tri-lobed arch are five tathāgatas arranged as if floating around the architecture in which Prajñāpāramitā sits, and the apex, right above the crown, is occupied by none other than Amitābha. If the human world below depicts devotees with an organized ritual scene, the heavenly realm above Prajñāpāramitā is filled with auspicious devotional fervour. Interspersed between the five

05_CreSouthV1_9P_11Feb22.indd 182

tathāgatas, two standing figures appearing amidst wavy pattern (perhaps representing water) offer garlands with their bodies slightly bent respectfully towards the goddess. Two standing female figures positioned on either side of Amitābha hold an offering plate and an incense burner, respectively. The top of the stele is finished with two celestial beings (often identified as vidyādharas) flying in towards the centre amidst clouds from either side, carrying offerings for the goddess. These figures celebrate the goddess while her spiritual sons reverently hover around her. Two monk-like figures seated on either side of the goddess looking up in veneration express the awe felt by the human world witnessing this auspicious moment of the goddess’ preaching that generates her sons. In understanding the ritual context in which these manuscripts of the AsP were used, we are helped by the Vajrāvalīnāmamaṇḍalopāyikā [henceforth VA], a ritual manual compiled by the famed Indian Buddhist master Abhayākaragupta, who was active at the Vikramaśīla monastery during the late 11th–early 12th century. In the VA, Abhayākaragupta describes the process by which a manuscript is prepared for use in consecration within an installation ritual, during which Amitābha is invoked and visualized within the manuscript (Mori 1999: 84–94; 2005). As Mori (2005: 211) summarizes, in the visualization stage for an installation ceremony according to the VA, ‘the deities visualized as the pledge-being (samaya-sattva) are Amitābha for sacred texts [pustaka] and Vairocana for buildings [vihāra], and their symbols, a red lotus [for Amitābha] or an eight-spoked wheel [for Vairocana], which is visualized as manifesting its form within either the sacred text or the building.’33 Whether this instruction in the VA inspired manuscript makers to include Amitābha in the design of manuscripts is difficult to confirm since it is equally possible that Abhayākaragupta may have recorded what he saw as a common practice in manuscript production that followed another set of instructions. Interestingly, at least as far as can 33. The term Mori translates as ‘sacred text’ is pustaka. To indicate the material quality of particular items, I take this term more straightforwardly as ‘manuscript’ or ‘book’.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor be determined from extant examples, this feature was adopted only in the manuscripts of the AsP, and only in those prepared somewhere in eastern India. The surviving manuscript evidence, then, suggests a distinct possibility that the VA’s pustaka used in tantric Buddhist ritual practices that Abhayākaragupta compiled may well have been that of the AsP. The advanced tantric iconography that appears in late 12th-century AsP manuscripts that are heralded by Amitābha in the opening page also supports this hypothesis (Kim 2013, 2021). The pairing of Amitābha and Prajñāpāramitā seen in 12th-century AsP manuscripts from eastern India helps locate Angkor in the Buddhist network that seemed to have shared a similar understanding of Prajñāpāramitā’s importance for tantric Buddhist rituals in the 12th century. Concurrence of this understanding in Angkor and at eastern Indian Buddhist sites is remarkable: out of twenty-six dated painted AsP manuscripts known from Eastern India, eight manuscripts feature this pairing of Amitābha and Prajñāpāramitā, and most date to the mid-12th to the early 13th centuries (Kim 2021: 122–126). Angkorian Prajñāpāramitā images remain iconographically and stylistically unique, but the appearance of Amitābha on the headdress of Prajñāpāramitā images in Angkor and the pairing of Amitābha and Prajñāpāramitā in painted manuscripts from eastern India offers a room to consider Angkor’s trans-regional connections to eastern Indian Buddhist centres, especially since palm-leaf manuscripts are portable objects.34 Although no such instance of painted Sanskrit palm-leaf manuscripts is known from Angkor, palm-leaf manuscripts with miniature paintings embedded in them could certainly have informed the knowledge experts in Jayavarman VII’s ambit, who, in turn, would have instructed these new aspects to be translated into art and architecture. While this last statement is a complete conjecture, it is worthwhile to keep this model as a possibility in our exploration of modes of trans34. We may also add here that Cambodia also has a tradition of making palm-leaf manuscripts in pothi format. How ancient this tradition is it is difficult to ascertain, and most surviving manuscripts are similar to Sinhalese manuscripts that are inscribed with stylus rather than written with ink.

05_CreSouthV1_9P_11Feb22.indd 183

183

fer and transmission of esoteric visual knowledge, which must have relied heavily on oral transmission within an initiation lineage. A manuscript could confer spiritual authority to its holder, i.e. a Buddhist master, while miniatures could provide visual intimations of iconographic information that can be adapted by local artists.

prajñāpāramitā, the begetter of the transcendence An impressive bronze image of a sixteen-armed, multi-headed image of dancing Hevajra (known as Kapāladhara Hevajra) signals artistic caliber, mastery over bronze casting technique, and the impact of tantric Buddhism in Jayavarman VII’s empire (Fig. 5.17).35 This apotheosis of the Hevajratantra, an important text of Yoginītantras (also known as the Anuttarayogatantras in later Tibetan classification system) also appears in monumental stone images and on Jayavarman VII’s temple at Banteay Chhmar (Woodward 1981; Sharrock 2006).36 These images depicting the complex iconography of Hevajra seem to stand at odds with images of Prajñāpāramitā, understood to be a quintessential Mahāyāna Buddhist goddess, and the importance of the ABP triadic conception during Jayavarman VII’s reign. Here, too, the AsP manuscripts from eastern India offer a fresh perspective to understand the complex doctrinal context of Esoteric Buddhism in Jayavarman VII’s Angkor and the second/third waves of ‘Vajrayāna Buddhism’ in general. A number of late 12th- and early 13th-century manuscripts of the AsP from eastern India feature 35. The authenticity of this sculpture was once questioned due to several unusual features on the metal’s surface. Unwavering efforts to understand the object better by the museum’s curator, Sonya Rhie Mace, led to multiple rounds of scientific examination and deep provenance research, which ultimately solved the mystery of the metal’s unique composition. Personal communication with the curator, 2017. Also see the museum’s object page for further information: http://clevelandmuseumofart.art/art/2011.143. 36. Philip Green (2014) has recently challenged the Hevajra identification of Banteay Chhmar images based on a more thorough investigation of the iconographic programme.

11/2/22 3:34 PM

184

Jinah Kim

Fig. 5.17: Kapāladhara Hevajra (sixteen-armed dancing Hevajra), Jayavarman VII’s reign (c. 1181–1220 CE). Bronze; overall: 46 × 23.9 cm. The Cleveland Museum of Art, Gift of Maxeen and John Flower in honour of Dr Stanislaw Czuma 2011.143.

painted panels that depict images of the Mahāyoga- and Yoginī-tantras (Anuttarayogatantra in the Tibetan classification system). We find images of Hevajra and Cakrasaṃvara, and in two rare examples, Kālacakra, with their respective consorts on the facing folios, which come together in embrace when the manuscript is closed (Kim 2013: 157–166). A complex image of sixteen armed Kapāladhara Hevajra in his fantastic four-legged dancing form is perfectly positioned facing his consort Nairātmyā in a late 12th- or early 13th-century manuscript of the AsP now in the Varendra Research Museum (Acc. No. 851) (Fig. 5.18).37 Like all other AsP manuscripts bearing tantric iconography from eastern India, this manuscript, too, has its opening pages marked by Amitābha and Prajñāpāramitā (Kim 37. See Kim (2021), Chapter 4, for more information about this manuscript.

05_CreSouthV1_9P_11Feb22.indd 184

2021: fig. 4.4). It is as if the text of the AsP, personified and preached by the goddess Prajñāpāramitā provides a substance, a physical and symbolic body that generates these images. The AsP’s text is baffling and esoteric to begin with, which may have held a special appeal to the Esoteric Buddhist practitioners (Seton 2017).38 Just as Amitābha and Prajñāpāramitā together herald and mark the context in which tantric deities arise in eastern Indian manuscripts, Angkorian Prajñāpāramitā bearing the Buddha in meditation may also be taken as generating powerful tantric Buddhist deities like Hevajra. The generative power of the goddess Prajñāpāramitā may seem to be an expected quality of the goddess known as the mother of all tathāgatas. The idea of Prajñāpāramitā as the mother to all Buddhas, or the ‘Buddha mother’, has been frequently assumed ever since Edward Conze’s seminal study (1949). Yet, Prajñāpāramitā, or the perfection of insight (prajñā), is not an earthly mother goddess, but one of the fundamental theological principles of Mahāyāna and Vajrayāna (Esoteric) Buddhism. Relying on an analogy in the earliest redaction of the Prajñāpāramitāsūtra, such as the AsP, Conze spawned the misconception of the Prajñāpāramitā as an age-old mother goddess.39 The passage in the AsP is an analogy designed to emphasize the role of Prajñāpāramitā, or the perfection of insight, in the process of achieving enlightenment. In Chapter 12 of the AsP, Prajñāpāramitā is likened to a mother whose many sons eagerly care for her because she is their begetter. The terms of relationship are explained through what the sons/Buddhas should do for their mother/Prajñāpāramitā. Prajñāpāramitā is a mother (mātā)/a begetter (janayitrῑ). But she is not a tender, loving mother but rather a rational one who instructs and shows the way of the world (lokasya saṃdarśa-

38. The text’s core conception of śūnyatā and paradoxical approach provided a philosophical foundation for the Buddhist Tantras to flourish. 39. Conze (1967: 243–244) suggests that the study of the images of Prajñāpāramitā affords a ‘contribution to the long and varied history of the Mother-goddess who, from the Paleolithic onwards, has occupied so great a place in human affection.’

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

185

Fig. 5.18: Kapāladhara Hevajra (sixteen armed and four legged). Fol. 44v. AsP manuscript, Bengal, late 12th–early 13th century. Pigment and ink on palm-leaf, folio size: 31.6 × 6 cm. Varendra Research Museum, University of Rajshahi (VRM) Acc. No. 851. Photo by Jinah Kim.

yitrῑ).40 It is only after the 11th century that we see the conception of Prajñāpāramitā as the mother of the Buddhas appearing in artistic records in India. In the Prajñāpāramitā stele from Odisha discussed above (see Fig. 5.17), the arrangement of hovering five tathāgatas surrounding the goddess references the generative power of the goddess Prajñāpāramitā as the begetter of the transcendental tathāgatas. A late 12th-century manuscript cover from the Nālandā monastery (in today’s Bihar, India) depicts the goddess Prajñāpāramitā in the centre attended by the Buddhas of the past, the present, and the future (Kim 2013: 169). This articulation of Prajñāpāramitā in 12th-century India suggests that Jayavarman VII (and his advisors)’s emphasis on Prajñāpāramitā’s motherly quality was a political move that embraced crystallization of Prajñāpāramitā’s role in Esoteric Buddhism in familial terms, which in turn allowed Jayavarman VII to situate himself in the apex of power. 40. Conze 1973: 173; AsP, Vaidya edition, p. 125.

05_CreSouthV1_9P_11Feb22.indd 185

epilogue In Musée Guimet’s renowned Khmer collection, we find a rare image of nearly life size stone sculpture of a female figure from Jayavarman VII’s reign, at one point identified as ‘Seated Tārā’ (Jessup and Zéphir 1997: 304–305). The image is carved fully in the round (Fig. 5.19). Her upper body with compact chest is left bare while delicately patterned skirt wraps around her lower body as she kneels tall. Her eyes are closed under thick, well defined eyebrows, and she holds a broad, subtle smile on her lips, evoking familial affinity with the well-known portrait sculptures of Jayavarman VII. With her head ever so slightly tilted downwards, the unnamed sculptor masterfully created an image that epitomizes serenity in flesh. Perhaps it is this humanity or evocation of a person along with the unique posture of kneeling that prompts the identification of the statue as an image of Jayarājadevī, one of the two principal queens of Jayavarman VII (Baptiste and Zéphir 2008). According to the inscription left by her

11/2/22 3:34 PM

186

Jinah Kim

Fig. 5.19: Prajñāpāramitā/Jayarājadevī/?, Jayavarman VII’s reign. Stone. H: 130 cm. Musée Guimet, Paris. MG 18043. Photo © Jinah Kim.

sister, Indradevī, who was the second principal queen of Jayavarman VII, she had images of the royal couple, Jayavarman VII and her late sister, Jayarājadevī, installed across Jayavarman VII’s territory (K 485; Multzer o’Naghten 2016: 44).41 Her arms are unfortunately missing from just below the armpit, making it impossible to suggest how she held her hands, adding to the enigma of the image.42 In front of a tightly pulled chignon on the very top of her head is a seated Buddha in meditation, which, as we have seen, is a common icono41. According to Multzer o’Naghten (2016), so far, seven such statues are known. 42. If her hands were held together in front of her chest holding a lotus, it would be a clear indication of adoration, as Woodward (1994–95) discusses in other Jayavarman era images.

05_CreSouthV1_9P_11Feb22.indd 186

graphic indicator of the goddess Prajñāpāramitā in Jayavarman VII’s Angkor. If this is Jayarājadevī represented in the mould of Prajñāpāramitā, in a way, this image participates in an interesting trope of projecting a human queen on the image of goddess Prajñāpāramitā in different parts of the Buddhist world. In addition to the famous Prajñāpāramitā stone statue of 1300 CE from Candi Singhosari, East Java, popularly identified with the Queen Ken Dedes of the Siṅhasāri dynasty (Reichle 2007; Multzer o’Naghten 2016), the 7th-century Gilgit bronze image of Prajñāpāramitā mentioned earlier may also have a connection with the queen, Maṅgalahāṃsikā named in the inscription on the pedestal of the image (see Fig. 5.10).43 In compar43. Along with many unique elements only seen in this image, the unusual headdress of bird feathers that are held by two haṃsas (swans, i.e. auspicious bird) on either side, may in fact be a reference to the queen, Maṅgalahaṃsika, whose name means auspicious swan (Seton 2008; Luczanits 2016). I thank Greg Seton for sharing his paper and insight on the sculpture.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor

187

Fig. 5.20: Angkor gallery, Musée Guimet. Installation shot taken in August 2009. Photo © Jinah Kim.

ison to the Singhosari Prajñāpāramitā image that follows the iconographic code established in Magadhan sites quite closely, perhaps the formula here is not the queen as the goddess, but the goddess represented like a human queen whose images protect the territory like the Khmer kings whose posthumous titles locate them in the realm of gods (Kim 2010: 114; Woodward 1994–95: 108). Multzer o’Naghten (2016: 48) suggests a possibility of a devotional practice of wearing insignia of one’s tutelary deity (or one’s chosen deity in a tantric lineage tradition) during Jayavarman VII’s reign.44 The miniature image of Viṣṇu affixed on 44. See Multzer o’Naghten (2016) for illustration (fig. 21). She suggests that the miniature image on Jayavarman VII’s

05_CreSouthV1_9P_11Feb22.indd 187

Jayavarman VII’s hairdo carved on the walls at Banteay Chhmar suggests a practice akin to other such practices of affixing sectarian insignia in various faith communities, such as those known among Śaiva groups in India.45 Following Multzer o’Naghten’s suggestion, we may hypothesize that the Buddha in meditation may have been such a devotional insignia of the queen although this remains a highly imaginative speculation.

hairdo is of Avalokiteśvara. According to Peter Sharrock (2015), this image is of Viṣṇu as the Khmer god of war. 45. For example, Lingayats are known to carry the sectarian marks like śivaliṅgas on their body. Jangams in today’s Punjab wear liṅga and feather on their turbans. I thank Swati Chemburkar for this reference.

11/2/22 3:34 PM

188

Jinah Kim

As seen in the photograph taken in 2009 (Fig. 5.20), the image is displayed together with the ‘Radiant’ Avalokiteśvara, the multi-armed Avalokiteśvara images that Hiram Woodward (1994– 95) identifies as a special type of Avalokiteśvara image distributed by Jayavarman VII across his territory, Jayabuddhamahānātha. Above them in the middle on a higher pedestal is a seated image of the Buddha in meditation under a nāga hood. Although the image is slightly damaged, the scale and the installation position make the hierarchy among the three certain.46 In this Guimet-generated context that simulates Jayavarman VII’s ABP triad, the evocation of the goddess Prajñāpāramitā is clear. With the Buddha in meditation under the nāga hood hovering over her, whether the Buddha in meditation on her head is Amitābha or not may not matter as much.47 The most important focus in the image is on deep meditation and sincere humility, which transcends any particular sectarian association. If anything, the Buddha that she bears most dearly may in fact be the very Buddha in the centre of Bayon who her consort Jayavarman VII installed, encapsulating the pious, deeply personal dimension of the political use of religious iconography.

References

Acri, Andrea, ed. 2016. Esoteric Buddhism in Mediaeval Maritime Asia: Networks of Masters, Texts, Icons. Singapore: ISEAS – Yusof Ishak Institute. Akhtar, Naseem. 2001. Patna Museum Catalogue: Terracottas and Metal Images. Patna: Patna Museum. Asher, Frederick M. 1980. The Art of Eastern India, 300-800. Oxford: Oxford University Press. Baptiste, Pierre and Thierry Zéphir. 2008. L’art Khmer dans les collections du Musée Guimet. Paris: Réunion des musées nationaux. Bautze-Picron, Claudine. 1995. ‘Between Men and Gods: Small Motifs in the Buddhist Art of 46. The Prajñāpāramitā is from Preah Khan of Angkor Thom and the others from Preah Khan of Kompong Svay. 47. Her pulled chignon in fact forms a mandorla around the Buddha and the unnamed artist may have carved the detail of braided (?) hair to evoke the shape of the nāga hood with its multiple heads and scale.

05_CreSouthV1_9P_11Feb22.indd 188

Eastern India, an Interpretation’, in K.R. van Kooij and H. van der Veere (eds.), Function and Meaning in Buddhist Art: Proceedings of a Seminar Held at Leiden University, 21–24 October 1991, pp. 59–79. Groningen: Egbert Forsten. . 2004. ‘The Universal Compassionate Bodhisattva, Miscellaneous Aspects of Avalokitasvara/Avalokiteshvara in India’, Silk Road Art and Archaeology 10: 225–290. . 2015. The Forgotten Place: Stone Images from Kurkihar, Bihar. 1st ed. New Delhi: Archæological Survey of India. Bhattacharyya, Benoytosh. 1925 [reprint 1968]. Sādhanamālā. Gaekwad’s Oriental Series, 2 vols. Vol. 1. Baroda: Oriental Institute of Baroda. . 1968. The Indian Buddhist Iconography, Mainly Based on the Sādhanamālā and Cognate Tantric Texts of Rituals. 2nd revised and enlarged ed. Calcutta: Firma K.L. Mukhopadhyay. [1st ed. 1924] Bianchini, Francesco. 2020. The Tradition and Innovation in late South Asian Buddhism: The Impact of Spell Practices on the Recasting of Prajñāpāramitā Scriptures. PhD thesis, University of Oxford. Bourgarit, David, Benoît Mille, Thierry Borel, Pierre Baptiste, and Thierry Zéphir. 2003. ‘A Millennium of Khmer Bronze Metallurgy: A Technical Study of Seventy-five Bronze Artefacts from the Musee Guimet and the Phnom Penh National Museum’, in Paul Jett (with Janet G. Douglas, Blythe McCarthy, and John Winter, eds.), Scientific Research in the Field of Asian Art: Proceedings of the 1st Forbes Symposium at the Freer Gallery of Art, pp. 103–126. London: Archetype. Bunker, Emma C. 2015. ‘Pala-inspired Vajrayana Buddhist Bronzes from the Khorat Plateau: The Khmer-Pala Link’, Marg 67/2: 86–95. Cœdès, George. 1906. ‘La Stèle de Ta-Prohm’, Bulletin de l’École française d’Extrême-Orient 6: 44–86. . 1920. ‘Note sur une statuette cambodgienne de la Prajnâ Pâramita’, Bulletin de l’École française d’Extrême-Orient 20: 7–8. . 1937. Inscriptions du Cambodge. 8 vols. Hanoi: Imprimerie d’Extrême-Orient.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor . 1941. ‘La Stèle du Práh Khằn d’Ankor’, Bulletin de l’École française d’Extrême-Orient 41: 255–302. . 1989. Articles sur le pays Khmer. Paris: École Française d’Extrême-Orient. Conze, Edward. 1948. ‘Text, Sources, and Bibliography of the Prajñāpāramitā-hṛdaya’, Journal of the Royal Asiatic Society 80/1–2: 33–51. . 1949–50. ‘The Iconography of the Prajñāpāramitā—Part 1’, Oriental Art 1: 47–52. . 1951. Buddhism: Its Essence and Development. New York: Philosophical Library, . 1956. ‘Tantric Prajñāpāramitā Texts’, Sino-Indian Studies V/2: 100–122. . 1960. The Prajnaparamita Literature. Indo-Iranian Monograph. Gravenhage: Mouton. . 1967. ‘The Iconography of the Prajñāpāramitā’, in Edward Conze (ed.), Thirty Years of Buddhist Studies: Selected Essays by Edward Conze. Oxford: Bruno Cassirer. . 1973. The Perfection of Wisdom in Eight Thousand Lines and Its Verse Summary. Bolinas, CA: Four Seasons Foundation. . 1978. The Prajñāpāramitā Literature. 2nd revised and enlarged ed. Tokyo: Reiyukai. Davidson, Ronald M. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New York: Columbia Uniersity Press. de Mallmann, Marie-Thérèse. 1975. Introduction à l’iconographie du tântrisme bouddhique. Paris: Centre de recherches sur l’Asie centrale et la Haute Asie. (Bibliothèque du Centre de recherches sur l’Asie centrale et la Haute Asie, 1). Dowling, Nancy. 1996. ‘Honolulu Academy of Art Tenth-Century Khmer Buddhist Trinity’, Artibus Asiae 56/3–4: 325–341. Estève, Julia and Brice Vincent. 2010. ‘L’about inscrit du musée national du Cambodge (K. 943): Nouveaux éléments sur le bouddhisme tantrique à l’époque angkorienne’, Arts Asiatiques 65: 133–158. Falk, Harry and Seishi Karashima. 2013. ‘A First-Century Prajñāpāramitā Manuscript from Gandhāra—Parivarta 5 (Texts from the Split Collection 2)’, Annual Report of the International Research Institute for Advanced Buddhology at Soka University XVI, pp. 97–169.

05_CreSouthV1_9P_11Feb22.indd 189

189

Ghosh, Mallar. 1980. Development of Buddhist Iconography in Eastern India: A Study of Tārā, Prajñās of Five Tathāgatas and Bhṛikuṭī. New Delhi: Munshiram Manoharlal. Green, Phillip S.E. 2014. ‘The Many Faces of Lokeśvara: Tantric Connections in Cambodia and Campā between the Tenth and Thirteenth Centuries’, History of Religions 54/1: 69–93. Harrison, Paul (ed.). 2018. ‘Early Mahāyāna: Laying out the Field’, in Paul Harrison (ed.), Setting Out on the Great Way: Essays on Early Mahāyāna Buddhism. Sheffield, South Yorkshire; Bristol, CT: Equinox Publishing. Harrison, Paul and Jens-Uwe Hartmann (eds.). 2014. From Birch Bark to Digital Data: Recent Advances in Buddhist Manuscript Research: Papers Presented at the Conference Indic Buddhist Manuscripts: The State of the Field, Stanford, June 15-19, 2009. Wien: Verlag der Österreichischen Akademie der Wissenschaften. von Hinüber, Oskar. 2007. ‘Three New Bronzes from Gilgit’. Annual Report of the International Research Institute for Advanced Buddhology at Soka University for the Academic Year 2006 (Vol. 10), pp. 39–43. Huntington, John and Dina Bangdel. 2003. The Circle of Bliss: Buddhist Meditational Art. Columbus, Ohio: Columbus Museum of Art; Chicago: Serindia Publications. Jessup, Helen I. and Thierry Zephir (eds.). 1997. Sculpture of Angkor and Ancient Cambodia: Millennium of Glory. Washington, D.C.: National Gallery of Art. Kim, Jinah. 2010. ‘Unfinished Business: Buddhist Reuse of Angkor Wat and Its Historical and Political Significance’, Artibus Asiae 70/1: 77–122. . 2013. Receptacle of the Sacred: Illustrated Manuscripts and the Buddhist Book-Cult in South Asia. Berkeley, CA: University of California Press. . 2015. ‘Painted Palm-Leaf Manuscripts and the Art of the Book in Medieval South Asia’, Archives of Asian Art 65/1–2: 57–86. . 2019. ‘Performing Texts, Engendering Merits: Manuscripts and Paubhas as Ritual Objects’, in Jinah Kim and Todd Lewis (eds.),

11/2/22 3:34 PM

190

Jinah Kim

Dharma and Puṇya: Buddhist Ritual Art of Nepal, pp. 88–91. Leiden: Hotei Publishing. . 2021. Garland of Visions: Color, Tantra, and a Material History of Indian Painting. Oakland, CA: University of California Press. Kinnard, Jacob N. 1999. Imaging Wisdom: Seeing and Knowing in the Art of Indian Buddhism. Richmond, Surrey: Curzon. Linrothe, Robert N. 1999. Ruthless Compassion: Wrathful Deities in Early Indo-Tibetan Esoteric Buddhist Art. Boston; New York: Shambhala. Lobo, Wibke. 1997. ‘L’image de Hevajra et le bouddhisme tantrique’, in Helen I. Jessup and Thierry Zéphir (eds.), Angkor et dix Siècles d’art Khmer, pp. 70–78. Paris: Réunion des musées nationaux. Luczanits, Christian. 2004. Buddhist Sculpture in Clay: Early Western Himalayan Art, Late 10th to Early 13th Centuries. Chicago: Serindia Publications. . 2016. ‘Prajñāpāramitā, Alchi and Kashmir: On the Cultural Background of a Unique Bronze’, in An Exceptional and Magnificent Bronze Alloy Figure of Prajnaparamita. Beijing: Poly Auction. Malandra, Geri. 1993. Unfolding a Maṇḍala: The Buddhist Cave Temples at Ellora. Albany: State University of New York Press. Mori, Masahide. 2005. ‘The Installation Ceremony in Tantric Buddhism’, in Shingo Einoo and Jun Takashima (eds.), From Material to Deity: Indian Rituals of Consecration, pp. 199–240. New Delhi: Manohar Publishers. . 1999. ‘The Vajrāvalī of Abhayākaragupta: A Critical Study, Sanskrit Edition of Selected Chapters and Complete Tibetan Version’. PhD Dissertation, SOAS, University of London. Multzer o’Naghten, Hedwige. 2016. ‘Prajñāpāramitā dans le bouddhisme du Cambodge ancien’, Arts Asiatiques 71: 31–54. Owen, Lisa. 2013. ‘Relationships between Art, Architecture, and Devotional Practices at Ellora’, in Pia Brancaccio (ed.), Living Rock: Buddhist, Hindu, and Jain Cave Temples in the Western Deccan, pp. 126–137. Mumbai: Marg Publications. Pal, Pratapaditya. 1988. ‘A Forgotten Monastery of Ancient Bihar’, South Asian Studies 4: 83–88.

05_CreSouthV1_9P_11Feb22.indd 190

Polkinghorne, Martin, Brice Vincent, Nicolas Thomas, and David Bourgarit. 2014. ‘Casting for the King: The Royal Palace Bronze Workshop of Angkor Thom’, Bulletin de l’École française d’Extrême-Orient 100: 327–358. Reichle, Natasha. 2007. Violence and Serenity: Late Buddhist Sculpture from Indonesia. Honolulu: University of Hawai‘i Press. Sastri, Hirananda. 1966. Nalanda and Its Epigraphic Material. New Delhi: Archaeological Survey of India. Seton, Gregory Max. 2008. ‘The Mother Prajñāpāramitā and the Development of Buddhist Yoginī Tantra’. A paper delivered at the University of California, Santa Barbara. . 2015. ‘Defining Wisdom: Ratnākaraśānti’s Sāratamā’. DPhil dissertation, Wolfson College, Oxford University. . 2017. ‘Integrating Non-Tantric and Tantric Doctrines through Prajñāpāramitā at Vikramaśīla During the Mid Eleventh Century’, paper presented at the XVIIIth Congress of International Association of Buddhist Studies, University of Toronto. Shah, Umakant Premanand. 1987. Jaina-rūpamaṇḍana = Jaina Iconography. New Delhi: Abhinav Publications. Sharrock, Peter D. 2006. ‘Hevajra at Bantéay Chmàr’, The Journal of the Walters Art Museum 64/65: 49–64. . 2007. ‘The Mystery of the Bayon Face Towers’, in J. Clark (ed.), Bayon: New Perspectives, pp. 230–281. Bangkok: River Books. . 2009. ‘Garuḍa, Vajrapāṇi and Religious Change in Jayavarman VII’s Angkor’, Journal of Southeast Asian Studies 40/1: 111–151. . 2012. ‘Kīrtipaṇḍita and the Tantras’, Udaya: Journal of Khmer Studies 10: 203–237. . 2013. ‘The Tantric Roots of the Buddhist Pantheon of Jayavarman VII’, in Marijke J. Klokke and Véronique Degroot (eds.), Materializing Southeast Asia’s Past: Selected Papers of the 12th International Conference of the European Association of Southeast Asian Archaeologists, Vol. 2, pp. 41–55. Singapore: NUS Press. . 2015. ‘Serpent-enthroned Buddha of Angkor’, Marg 67/2: 21–31.

11/2/22 3:34 PM

Goddess Prajñāpāramitā and Esoteric Buddhism in Jayavarman VII’s Angkor Sharrock, Peter D., Claude Jacques, Olivier Cunin and Thierry Zephir. 2015. Banteay Chhmar: Garrison-temple of the Khmer empire. Bangkok: River Books. Shaw, Miranda. 2006. Buddhist Goddesses of India. Princeton: Princeton University Press. Snellgrove, David L. 1987. Indo-Tibetan Buddhism: Indian Buddhists and Their Tibetan Successors. 2 vols. Boston: Shambhala. Studholme, Alexander. 2002. The Origins of Oṃ Maṇipadme Hūṃ: A Study of Kāraṇḍavyūha Sūtra. Albany: State University of New York Press. Vidyabhusana, Satis Chandra (ed.). 1908. Bauddha-Stotra-Saṁgrahaḥ or a Collection of Buddhist Hymns. Volume I: Sragdharā-Stotram or a Hymn to Tārā in Sragdharā Metre by Bhikṣu Sarvajña Mitra of Kaśmīra. With the Sanskrit Commentary of Jina Rakṣita, together with two Tibetan Versions. Calcutta: The Asiatic Society of Bengal. Vincent, Brice, David Bourgarit and Paul Jett. 2012. ‘Khmer Bronze Metallurgy during the Angkorian Period (Twelfth to Thirteenth Centuries): Technical Investigation of a New Selected Corpus of Artifacts from the National Museum of Cambodia, Phnom Penh’, in Paul Jett, Blythe McCarthy, and Janet G. Douglas (eds.), Scientific Research on Ancient Asian Metallurgy: Proceedings of the Fifth Forbes Symposium at the Freer Gallery of Art, pp. 124–153. London:

05_CreSouthV1_9P_11Feb22.indd 191

191

Archetype Publications/The Freer Gallery of Art, Smithsonian Institution. Wayman, Alex. 1973. The Buddhist Tantras: Light on Indo-Tibetan Esotericism. New York: S. Weiser. . 1977. Yoga of the Guhyasamājatantra: The Arcane Lore of Forty Verses; a Buddhist Tantra Commentary. 1st ed. Delhi: Motilal Banarsidass. Woodward, Hiram W. 1979. ‘The Bàyon-Period Buddha Image in the Kimbell Art Museum’, Archives of Asian Art 32: 72–83. . 1980. ‘Some Buddha Images and the Cultural Developments of the Late Angkorian Period’, Artibus Asiae 42/2–3: 155–174. . 1981. ‘Tantric Buddhism at Angkor Thom’, Ars Orientalis 12: 57–67. . 1994–95. ‘The Jayabuddhamahānātha Images of Cambodia’, Journal of the Walters Art Gallery 52/53: 105–111. . 2004. ‘Esoteric Buddhism in Southeast Asia in the Light of Recent Scholarship’, Journal of Southeast Asian Studies 35/2: 329–354. . 2007. ‘The Karandavyuha Sutra and Buddhist Art in 10th Century Cambodia’, in Pratapaditya Pal (ed.), Buddhist Art: Form & Meaning, pp. 70–83. Mumbai: Marg Publications. . 2015. ‘Aspects of Buddhism in Tenth-Century Cambodia’, in D. Christian Lammerts (ed.), Buddhist Dynamics in Premodern and Early Modern Southeast Asia, pp. 218–260. Singapore: Institute of Southeast Asian Studies.

11/2/22 3:34 PM

Chapter 6

Dancers, Musicians, Ascetics, and Priests: Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor1 S wa t i C h e m b u r k a r

Ś

introduction

aivism was the dominant Indic religious tradition in the ancient Khmer domains in the early mediaeval period (ca. 7th–13th century CE). The ideologies and ritual practices of its various currents, from the Atimārga or Pāśupata Śaivism (flourished ca. 3rd–8th century) to the Mantramārga or tantric Śaivism (flourished ca. 7th–13th century),2 inspired the extant corpus of Sanskrit and Old Khmer inscriptions as well as the regions’ temple architecture.

1. While this chapter specifically focuses on performance-related practices in Khmer temples, some parts regarding Pāśupata practices in the ancient Khmer domains have been previously published in Chemburkar and Kapoor 2018; see also Chemburkar 2015 for references to dancers in Khmer inscriptions and temple iconography, especially under Jayavarman VII, and Chapter 7 in this volume for a possible Pāśupata origin of Khmer temples’ ‘annex buildings’. I owe special thanks to Andrea Acri for the helpful discussions on the subject of performers and their religious affiliations, for his inspiring work on the subject, and for being a most attentive and helpful editor. I also thank Olivier Cunin for generously sharing his breadth of knowledge on the Khmer temples, and Shivani Kapoor and Peter Sharrock for tirelessly reading and commenting on previous drafts of this chapter. All the images reproduced here are mine except otherwise specified. 2. The Mantramārga was less severe in terms of asceticism and more oriented towards (state-sponsored) ritualism than the Atimārga; the two movements also differed with regard to their conception of initiation, being a fundamentally transformative and purifying act for the former, and a mere rite de passage for the latter. A succinct and authoritative characterization of these traditions is the seminal article by Sanderson (1988); on their respective textual corpora, see Sanderson 2012–2013.

06_CreSouthV1_8P_11Feb22.indd 192

The Pāśupata movement was one of the earliest organized and widely distributed ascetic orders of Śaivism. The Brahmins of this sect left significant traces of their presence across the subcontinent and also secured a prominent position in the ancient Khmer, Cam, and Javanese domains.3 One of their most characteristic traits was the incorporation of song and dance into observance (vrata), both within and without temples. The key text Pāśupatasūtra indicates that the initial systematic inclusion of song and dance in ascetic observance can be traced to the Pāñcārthika tradition, whose earliest textual sources date back to ca. 4th or 5th centuries CE.4 An association with performance may also be detected among the various sub-traditions that sprung up within the Pāśupata fold, including the Lākulas and the Kāpālikas/ Kālamukhas,5 as well as the elusive ‘Siddhas’, who

3. Early scholarship on the history of the Pāśupatas is that of D.R. Bhandarkar (1908) and R.G. Bhandarkar (1913), Banerjea (1941, 1951), Pathak (1960), and Hara (1966); recent scholarship is that of Bakker (2000, 2014), Bisschop (2006, 2010), and Acharya (2006, 2007, 2018). Relevant contributions on Pāśupata Śaivism in the ancient Khmer domains are those by Bhattacharya (1955), Sanderson (2003–2004), and Goodall (2015). 4. On the basis of the exposition of the Gaṇakārikā and its commentary Ratnaṭīkā, we can infer that the Pāñcārthika tradition recognized five principles (pañcārtha), namely effect (kārya), cause (kāraṇa), a prescribed set of rules (vidhi), attainment of union with the Lord (yoga), and the end of suffering (duḥkhānta). 5. See Acri 2018. On the various sub-sects and historical developments of Pāśupatism, see Jash 1974: 61–79 and Bakker 2000.

11/2/22 3:36 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor appear to have taken up some of the antinomian behaviours of the Pāśupatas.6 While the Pāñcārthika Pāśupata system was characterized by an ascetic character, early references to this sect in the epigraphical records document endowments for temples, and ‘refer to Pāśupatas as recipients for the performance of worship in the temples’ (Bisschop 2010: 485).7 It has, thus, become increasingly clear that, fairly soon in the history of the order, some Pāśupatas started to perform temple rituals, thereby obtaining royal support. The performing arts, including singing and dancing, seem to have played a role in their modes of worship. These may then have been carried forward and further developed by the Mantramārga throughout the mediaeval period in Śaiva temples and monasteries in both the Indian subcontinent and Southeast Asia. This chapter documents and re-examines the Śaiva ritual practices enjoining performance in the mediaeval Khmer domains. It combines data mined from the Śaiva textual archive in Sanskrit and vernacular languages with hitherto neglected art historical material from the ancient Khmer domains, as well as insights from Campā, Java, and South India.8 Following on from Bhattacharya’s (1955) pioneering work on Pāśupatism in the ancient Khmer domains, this chapter will first discuss the presence of this movement in the early Angkorian period and its survival as an elite corp of state-sponsored ritualists through the Angkorian period. Then, starting from the reflections by Ronald Davidson (2002), who accords a major historical role to the Pāśupa6. On Siddhas, see Davidson 2002 and Samuel 2006. On the trope of Siddhas portrayed as evil ascetics or sorcerers fond of singing, dancing, and consuming alcohol in the company of women in mediaeval Indic Sanskrit literature, see Bloomfield 1924. 7. For instance, the 4th-century Bāgh copper plates (nos. III, V, VI, IX, X, XII and XIV) refer to Pāśupatas as recipients of grants for the performance of temple worship (Ramesh and Tewari 1990); the Mallar copper plate grants of Śivagupta Bālārjuna (ca. 7th century) mention the involvement of Pāśupata saint Aghorajyotis in the proceedings of the Śiva Bāleśvara shrine (Shastri 1995, vol. 2: 380–382). 8. For the iconography of Pāśupata ascetics in the Khmer domains, see Kapoor 2019: 122–131 and Chemburkar and Kapoor 2018: 45–57.

06_CreSouthV1_8P_11Feb22.indd 193

193

tas as the ritual agents who first developed music, song, and drama in the Indic religious landscape,9 it will focus on the evidence of performance-based worship by Pāśupata agents in the Khmer domains, and finally describe the development of performance in Śaiva and Buddhist temples and festivals in Śaiva Mantramārga and even Buddhist (e.g., in the case of Jayavarman VII) milieus in the late Angkorian period. The aim is to assemble the multifarious and fragmentary evidence scattered across both South and Southeast Asia that could help us to form a more authentic appreciation of the religious landscape of the premodern societies under study, so as to reach a richer perception of how the sect functioned beyond the dimension of prescriptive texts in its various geographical and socio-cultural contexts.

the pāśupatas: from the indian subcontinent to the khmer domains Pāśupata adepts worshipped Śiva under the name Paśupati, Lord (pati) of cattle (paśu, also in the sense of ‘sentient beings/bound souls’). The ascetic path of the sect followed five stages, in which the practitioners were to live in five different places: in a temple, among people, in an empty cave, in a cremation ground, and ultimately becoming one with Rudra (Śiva) (Acharya 2018). The first stage of practice included bathing and lying in ash, and worshipping Śiva—presumably the liṅga—with singing, dancing, laughing, reciting mantras, and so on. During the second stage, while living amongst people, Pāśupatas exhibited hedonistic and debased behaviour such as pretending to be asleep, being paralysed, walking as if crippled, consuming alcohol or exhibiting signs of lust at the

9. See Davidson 2002: 223: ‘the Pāśupatas are the probable source for the employment of song and dance in the Buddhist forms of worship […]. The Pāśupatas particularly enjoyed the use of song and dramatic forms in the worship of Śiva, and this emphasis occurs from the earliest documents right through the life of the order. […] Their virtuosity in vocal song and structured forms of dance were perhaps an extension of their involvement in court life and missionary activity.’

11/2/22 3:36 PM

194

Swati Chemburkar

sight of woman.10 These unorthodox and outlandish ascetic practices might explain why references to Pāśupatas in Indian mediaeval Sanskrit literature are often hostile. Several Sanskrit inscriptions testify to the fact that Pāśupata ascetics spread from North India to Nepal, Bengal, Karnataka, Tamil Nadu, and overseas to Southeast Asia. Inscriptions from the Khmer and Cam domains mention Pāśupata masters or refer to typically Pāśupata forms of Śiva, doctrines, and practices,11 which are in harmony with those described in the seminal Pāśupatasūtra (with the commentary Pañcārthabhāṣya) and Gaṇakārikā (with the commentary Ratnaṭīkā), as well as other transmitted literature in Sanskrit.12 Alexis Sanderson (2003–2004: 403) notes that the re-creation of a Śaiva landscape of Indic sacred sites in Kambujadeśa is seen as early as 6th century CE, and many names of Śaiva sites and sanctuaries recorded in pre-Angkorian inscriptions are borrowed from Indian holy places that were associated with Pāśupatism.13 Vṛddheśvara in 10. Some of these behaviours became (proverbially) associated with the Kāpālikas, a sect that probably developed from Pāśupata Śaivsim. The farcical drama Mattavilāsaprahasana, composed in Kāñcī in the 7th century, portrays three characters—a Pāśupata by the name of Babhrukalpa, the Kāpālika Satyasoma (addressed as a Mahāpāśupata), and his partner, a dancing Kāpālinī called Devasomā. It describes how the hedonist couple was engaged in wild dances, consumed meat and alcohol, and enjoyed sexual intercourse (Lorenzen 2000: 82). 11. From the Khmer domains, see inscriptions K. 604, v. 12 (IC, IV: 17–19), K. 733, v. 4 (IC, I: 3), K. 80 (IC, VI: 3), K. 258, face CII, st. 21 (IC, IV: 175), K. 13 (Barth 1885: 13 and Bhattacharya 2009: 22), K. 701 (Cœdès 1932: 84–112), K. 364 (Finot 1912: 1–28), K. 1049, st. III (Goodall 2015: 12), K. 1155 (Estève 2009: 518–523). From the Cam domains, see the Mỹ Sơn stele inscription of Vikrāntavarman (no. 17, v. IV, line 12, Finot 1904: 930; Majumdar 1927: 31–35), the Bang An stele inscription of Bhadravarman III (no. 42, v. VI, line 24; Huber 1911: 9; Majumdar 1927: 128), 9th-century Po-Nagar inscription of Vikrāntavarman II (no. 30, v. III and IV, Majumdar 1927: 71–74), and the stele of Glai Lomov (Bergaigne 1893: 226, 393B line 2; cf. Bhattacharya 1955: 481), mentioning white ashes, yoga, muttering, huṃkāra, and becoming one with Rudra (sitabhasmaprabhāvayogādijapahuṅkāranirmmalataraśarīrapradeśa). 12. For a brief account of Pāśupata and Pāśupata-influenced literature, see Chapter 7 of this volume, fn. 35. 13. Sanderson (2003–2004: 405–406) traces at least nine-

06_CreSouthV1_8P_11Feb22.indd 194

Kārohaṇa (Karvan), the supposed place of Rudra’s incarnation as Lakulīśa—the originator of the Pāśupata teaching—is also found in Khmer epigraphy.14 Bhadreśvara, the benevolent form of Śiva well known in Indian sources, became the state God of the Khmers at Wat Phu and Sambor Prei Kuk, but also of the Cams across the mountains at Mỹ Sơn, playing a similar role to Śiva Naṭarāja of Cidambaram for the Tamils, and Paśupati of Deopatan for the people of Kathmandu. In fact, an inscription from Khmer king Dharaṇīndravarman (r. 1107–1113) equates Bhadreśvara with Paśupati.15 Particularly noted for their accumulated knowledge in a vast array of subjects such as debating and teaching skills, and maybe also magical powers aiming at securing success in battle for their kings and patrons, Pāśupata masters must have been revered in the Khmer domains. Some were placed in charge of temples, and many received land grants.16 By the 7th century, Pāśupata teachers had secured a place at the Khmer court and were also involved in the worship activities that were carried out at the temples of their patrons, as evident from the early Sambor Prei Kuk inscription of 627 CE:17 dvijaf pāśupato rājñādhikṛto devatārccane idan devakulaṃ bhoktum arhaty ābhūtasaṃplavam A Pāśupata Brahmin appointed by the king for the service of the god must enjoy the benefits of this temple till the end of the world.

teen doubles of the Śivas of which the first eight are Pāśupatas, and included Amareśvara, Prabhāsa and Siddheśvara. Vickery’s analysis (1998: 140–141) of some two-hundred pre-Angkorian inscriptions observes fifty names—most ending in Īśvara—that seem to be Śaiva, fourteen Vaiṣṇava, and eight Śaiva and Vaiṣṇava combined. 14. Vṛddheśvara as a Śaiva god/shrine occurs in K. 374, vv. 3, 8 (IC, VI: 251), K. 33, v. 24 (IC, III: 148), and K. 284, v. 4 (IC, III: 107). The word occurs in K. 388B, v. 11 and K. 550, v. 2, possibly in Buddhist context or as an ‘ancient’, pre-Indic god (see Vickery 1998: 146–148). 15. K. 258C, v. 2 (IC, IV: 175). 16. K. 81, v. 33 (Barth 1885: 19); K. 437, vv. 5–7 (IC, IV: 32); K. 44, vv. 4–5 (IC, II: 12) 17. K. 604 (IC, IV: 17); for the Sanskrit text, see Finot 1928: 43–46.

11/2/22 3:36 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

pāśupata brahmins: an elite class? There is no doubt that Pāśupatas became key agents in Khmer royal courts, obtaining the confidence of the kings. The Sambor Prei Kuk inscriptions of Īśānavarman I (r. 617–635) entrust Pāśupata Vidyāviśeṣa with the care of a temple,18 and those of Bhāvavarman II (r.?–c. 657) mention the employment of a Śaiva ascetic named Vidyāpuṣpa as a poet.19 K. 523 mentions an endowment to Śiva by the tapasvin Vidyāspada. These priestly names beginning with Vidyā are typically Pāśupata in both India and the Khmer domains.20 Khmer epigraphy also attests the names commencing with Bhā- or Bhāva- such as Bhāvaruci, Bhāpuṣpa, Bhāvajñāna, which were possibly Pāśupatas.21 Vickery’s study (1998: 103–111) on the regional groups of officials discusses Poñ and Mrātan as elite class people. These titles are always accompanied by Pāśupata names such as (Poñ) Bhākumāra, (Poñ) Bhāpuṣpa or (Mrātan) Bhāprasanna with few exceptions, and always with respect to the donations of personnel to Śiva or a liṅga. This suggests that Pāśupata agents formed an elite circle within Khmer courts at least up to the 7th century. The names ending in -rāśi, such as Tejorāśi and Śivarāśi, might also be suggestive of the continuing of the Pāśupata tradition in the Khmer domains as Bakker (2014: 175) argues that these were Pāśupata post-initiatory names. For instance, one Brah18. K. 604, v. 12 (IC, IV: 17–19). 19. K. 733 (IC, I: 3). It mentions Pāśupata ascetic Vidyāpuṣpa, a learned poet consecrated a silver praṇālika and practices of austerities (tapas) in conformity with Śaiva doctrine. Cf. Bhattacharya 1955: 480–487. 20. See Estève 2009: 475 (cf. 547–551), relating an information by Goodall. Other relevant names are Vidyākīrti (K. 127, IC, II: 89), Vidyādeva (K. 80, IC, VI: 3), Vidyāvindu (K. 13, ISC, no V, p. 31), Vidyāpuṣpa (K. 733, IC, I: 3), and Vidyāvinaya (K. 54, IC, III: 157). However, Zakharov (2019) has argued that vidyā was popular in ancient Cambodia and does not necessarily denote a Pāśupata. 21. On the basis of his study of Śaiva cult in northern India, Pathak (1960: 19, fn. 2) argues that ascetics of Praṇāma and Ananta gotras of the Pāśupata sect bore names beginning with Bhāva-, such as Bhāvavālmīki, Bhāvatejas, Bhāvodyota. Estéve (2009: 476–480) and Goodall (2015: 26–27) support this view. Vickery (1998: 152, fn. 53) says: ‘probably the element bha should be bhā, occurring in several elite titles but not yet understood’.

06_CreSouthV1_8P_11Feb22.indd 195

195

maṇarāśika had occupied an important place in Jayavarman III’s court (r. 835–877).22 Many ascetics with names ending in -rāśi lived in Varanasi, a city whose connection with the Pāśupata religion is found until the 12th century (ibid.: 30). The pre-Angkorian tradition of names beginning with Bhā-/Bhāva-/Vidyā- or ending in -rāśi continued during the reign of Indravarman I (r. 877–889), and Yaśovarman I (r. 889–ca. 910). Later, especially during the 10th century, names ending in -śiva such as Jyotiśiva, Varmaśiva, and Sadāśiva became more common, and Goodall (2015: 25) notes that these are indicative of Śaiva tantric Mantramārga traditions;23 at the same time, he suggests that ‘this does not necessarily mean that we can conclude with certainty that the Mantramārga only reached the Khmers in the Angkorian period’ (ibid.: 25–26). Along similar lines, Sanderson (2003–2004: 444) points out that certain iconographical elements found in Mantramārgic sources from the subcontinent are already attested among the Khmers in the 7th century, ‘well before the Mantramārga reached their shores’. The Prasat Prei inscription of Yaśovarman I mentions Śaiva and Pāśupata ācāryas side by side, and differentiate them from Brahmins: Immediately after the Brahmin, one honours the Śaiva master and the Pāśupata master, and whichever of those is instructed in grammar, he will be honoured more than the others. Preferred among those who know in depth the doctrine of the Śaivas and the Pāśupatas, science and grammar, is the master who teaches them, who will be honoured to the highest degree in this excellent hermitage.24

22. K. 449, st. 24–25 (Cœdès 1913: 29, 34); compare Estève 2009: 486. Ācārya Tejorāśi is mentioned in 9th-century K. 233B, stanza 16 (Cœdès 1954: 63), and in early 10th-century K. 262S, stanza 9; K. 263B, stanza 44; and K. 263D, stanza 37 (IC, IV: 108). Śivarāśi is mentioned in 11th-century K. 258B, stanza 60; K. 258C, stanza 50 (IC, IV: 175). 23. For a discussion of what appear to be Pāśupata/Śaiva Saiddhāntika names in Cambodian inscriptions, see Estève (2009: 475–488) and Goodall (2015: 20–25). 24. K. 279 (Bergaigne 1893, nos. LVI–LX, C1, vv. 6, 7, 10, pp. 243, 248, 423, 428; Cœdès 1932: 97–103).

11/2/22 3:36 PM

196

Swati Chemburkar

After Yaśovarman, the Pāśupatas seem to have remained relatively influential as 11th-century Khmer epigraphy mentions Paśupati (as visible on mount Bhadreśvara)25 and the influential priest Divākara making offerings to Paśupati (a liṅga?) as part of the royal coronation of Sūryavarman II at Phnom Sandak (K. 194) and at Preah Vihear (K. 383).26 Sanderson has convincingly argued that Mantramārga Śaivism took over in the public/royal worship, but it is also possible that Pāśupata beliefs, rituals, and modes of worship were absorbed into the Mantramārga and, perhaps, adopted by subsequent generations of local practitioners and reconfigured as inherently ‘Khmer’.27 Furthermore, Pāśupata cultic agents and their subgroups would have remained integrated into the Khmer court irrespective of the primary religious affiliations of the kings. The temple epigraphy of the late Angkorian period does not expressly mention Pāśupata ascetics, but it still mentions Paśupati and offerings being made to it. Zhou Daguan, who spent a year (1296–1297) in Angkor as an envoy of Temur Khan of the Yuan dynasty, mentions banjie (as learned men or paṇḍitas/Brahmins), zhugu (as Buddhist monks) and Śaiva ascetics as Basiwei or Pa-sseu-wei (Pāśupata? Tapasvins?) wearing a tall headdress (jaṭā?), making offerings to stones (liṅgas?) as opposed to icons, and wearing red or white cloth on their head (Harris 2007: 52, 53, 104). Some of the ascetics in the historical parade scene depicted on the southern gallery of Angkor Wat might be Pāśupatas, as suggested by Cœdès (1911b: 201), in the light of Daguan’s description.28

25. K. 364, v. 49 (Finot 1912: 27); K. 258C, v. 21 (IC, IV: 186). 26. Cœdès and Dupont 1943: 134. 27. It is often difficult to know to what degree the Śaiva beliefs were Indian and to what degree they were specific to Khmers. Maxwell (2008: 74) argues that ‘we should not expect to find Indian Hinduism, or Indian Buddhism, in Cambodia of Jayavarman VII—or even, perhaps in Funan.’ 28. There are two types of ascetics depicted on this wall: the one wearing a very high, beehive-like bun, and the other wearing a fabric cap ending with a knot enclosing the bun. The fabric cap would be indicative of Pāśupatas, who used to wear red or white cloth on their head (see Harris 2007: 53).

06_CreSouthV1_8P_11Feb22.indd 196

pāśupata textual accounts of performance-based worship The Pāśupatasūtra with the Pañcārthabhāṣỵa (ed. Sastri 1940) and the Gaṇakārikā with the Ratnaṭīkā (ed. Dalal 1966) mention the process (vidhi) for the attainment of liberation through five main sacred acts in the first stage of the adept’s ascetic career, when he lives in the temple. From the Pāśupatasūtra (verses 1.7 to 1.9) we learn that his observances are as follows: purifying himself by the besmearing with ashes three times a day (bhasmanā snānam),29 sleeping in ashes, supplementary bath with ash, wearing a single garment, and finally wearing flowers that had been offered to the deity (nirmālya). He resides in the temple (āyatana), the holy place of Maheśvara, and serves the Lord with the offerings of laughing, singing, dancing, making the sound huḍu(k), making internal salutations, and mantra recitations (hasita-gīta-nṛtya-ḍuṃḍuṃkāra-namaskāra-japyopahārenopatiṣṭhet, Pāśupatasūtra 1.8). The Ratnaṭīkā (Dalal 1966: 17–19) illustrates the procedure: the aspirant walks to the temple after the ash bath, chanting the Sadyojāta mantra, and bowing to Śiva. He then proceeds to the interior, where he kneels to the right of the image (dakṣiṇāmūrti), laughing loudly (hasita), singing (gīta), dancing (nṛtya), producing the auspicious sound of huḍukkāra or ḍuṃḍuṃkāra like the bellowing of a bull, produced by the contact of the tongue-tip with the palate. In this way, he circles around the deity three times and completes his inner worship (namaskāra) with single-minded concentration on the meaning of each successive syllable. Songs are played according to the Brahmanical rules of Gāndharvaśāstra, in which stories relating to Śiva are sung in public, while dance is performed according to the Nāṭyaśāstra, which comprises all possible gestures of the hands and feet.30 29. On bhasmasnāna, see Chapter 7 of this volume, pp. 240–244. 30. If we apply the reading adopted by Hara (1966: 182) in Pañcārthabhāṣya on Pāśupatasūtra 1.8 (Sastri 1940: 13, lines 14–16). For the view that the music and dance to be performed not in strict accordance with Gāndharvaśāstra and Nāṭyaśāstra, see Törzsök 2016: 460; Acri 2018: 4. Given the ambiguous nature of these rare references, it is

11/2/22 3:36 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor We may compare the above with the account of early Śaiva practice found in the Niśvāsamukha, the introductory book of the early Mantramārga scripture Niśvāsatattvasaṃhitā, large sections of which show traces of Pāśupata Śaivism. To the sages who ask how liṅga worship is to be performed, and question the benefits of partaking in the same through dancing, singing, playing the lute (tantrīvādya), and making the sound huḍu(k), Śiva reveals that the sages will obtain the same infinite pleasures as the Gaṇas, the members of the retinue of the Lord, who reside with him on Mount Kailāsa.31 Commenting on Niśvāsamukha 1.169, Acri (2018: 3) observes that those performance-based acts of worship seem to have been the preserve of both lay people and initiated into the Śaiva (i. e. Pāśupata?) system, and that the latter received an even greater benefit from them—i.e., liberation (apavārga). References to the musical and dramatic accomplishments of Pāśupatas and Kālamukhas, often carried out in a context of public religion and temple-based worship, are found in some Indian inscriptions.32 Furthermore, there are frequent difficult to determine how exactly the dance and music were performed. 31. Niśvāsamukha 1.73–76, 1.166–167, 4.72 (Kafle 2020: 255–257, 276, 353). 32. The important role of the Pāśupata community is confirmed by the Kālañjar stone inscription of Kīrtivarman’s reign (1090 CE) found in the Nīlakaṇṭha temple in Karnataka, mentioning the kings’ Śaiva guru Vāsudeva as the builder of a maṇḍapa that was possibly meant to house the dance and music worship of the order (Trivedi 1978: 370). Dancers and musicians are mentioned in the grants of Śaiva temples associated with Kālamukha priests: see, e.g., the grant to Kedāreśvara temple by Hoysala king Vīra-Ballāla (Rice 1902: 76, inscription ŚK 105). A 12th-century inscription in the same temple praises its ācārya as being pre-eminent in drama and the science of music, while another 11th-century inscription attributes the same qualities to Śrīkaṇṭhapaṇḍita, disciple of Kedāraśaktimunipati (Rice 1902: 59–61, inscriptions ŚK 92 and 94; cf. Lorenzen 1991: 127). Many more inscriptions mention priests of possible Kālamukha affiliation: see, e.g., the 1059–1060 CE Nāgeśvara temple inscription from Someśvara I’s reign at Sudi (Karnataka), mentioning the ‘ones acting for the god’s enjoyment and dancers graced the four pillars’ (vv. 6, 7, 13), and the beneficiary assigned to the performance rituals in the temple, which include Someśvara I, a flute player and dancing women (Barnett 1925); and an inscription dated

06_CreSouthV1_8P_11Feb22.indd 197

197

mentions of Pāśupatas in Sanskrit literature in association with a colourful landscape of dancers, musicians, ascetics, hermits, and courtesans. For instance, already in the Meghadūta by Kālidāsa (ca. 5th century), we find a mention of certain veśyās or dancing-girls/courtesans at the famous Mahākāla temple of Ujjain, one of the major and most ancient Pāśupata sacred sites (verse 35); the word veśyā is interpreted by 10th-century Kashmirian commentator Vallabhadeva as bhagavadgaṇikāḥ, ‘the courtesans of the Lord’ (Goodall 2018: 93; cf. infra). The 8th-century novel Kuṭṭanīmata by Kashmirian Dāmodaragupta exposes the whole craft of courtesans in the form of a dialogue between a novice and an experienced one. Ascetics figure predominantly in the colourful world of courtesans. A verse (539, Goodall and Dezső 2012: 210–211) makes fun of the fondness of Bhāvaśuddha (a Pāśupata ācārya from Varanasi) for the dancing girl Anaṅgadevī, for whom he has built a mansion. Another one (753, ibid.: 282–283) calls a courtesan ācāryanī, as a result of her closeness to Pāśupatācārya Bhāśuddha—presumably, the same Pāśupata ācārya mentioned in verse 539. The text (v. 743, ibid.: 278–279) also mentions a devadāsī at the Gambhīreśvara temple at Varanasi. Kṣemendra’s 11th-century satire, Samayamātṛkā (The Courtesan’s Keeper), refers to an ascetic with piled-up hair (jaṭādhārī) named Līlāśiva, who visits a courtesan (gaṇikā) at night and in the morning avoiding the main street to reach his maṭha.33 Merutuṅga’s 10th-century Prabandhacintāmaṇi (Part 1, Mūlarājaprabandha, v. 15; trans. Tawney 1901: 26) narrates an episode where Solanki King Mūlarāja searches for a tapasvin to take care of the Śiva temple; the tapasvin Vayajalladeva agrees on the condition that Mūlarāja supplied him daily with saffron, musk, camphor 1055 CE from Kādambeśvara temple at Bantapur, Dharwar, Karnataka, listing the grants made to temple girls, dance master, the keeper of the temple girls, musicians and drummers (Barnett 1915–1916: 175). 33. Samayamātṛkā of Kṣemendra (part 6, verse 9), see Pandit and Parab (1925: 40). Names ending in -śiva usually denote a Mantramārga rather than Atimārga (i.e. Pāśupata) affiliation, but the relevant point here is that the character in question is supposed to be an ascetic, as the term jaṭadhārī and the reference to his residing in a maṭha would seem to suggest.

11/2/22 3:36 PM

198

06_CreSouthV1_8P_11Feb22.indd 198

Swati Chemburkar

11/2/22 3:36 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

199

Fig. 6.1: (a) The liṅga pedestal of Mỹ Sơn E1 temple on the top as displayed at the Da Nang museum, depicting the activities of various ascetics (photo: Tran Ky Phuong). Particulars of the pedestal: top row from left to right: ascetics in the forest or hermitage setting and conversing or teaching (b–d); second row: ascetic in the forest or hermitage setting, meditating with a rosary in hand and playing a musical instrument (e–g); third row: bearded ascetic kneels before a yoni and lustrates a liṅga while his disciple is carrying an offering tray (h) + ascetic with a disciple (i); fourth row: ascetic holding a book (l), ascetic (m), ascetics playing a flute (n) and a three-stringed instrument (o); bottom row: ascetics playing flute, drums, and dancing (p) (photos: Paisarn Piammattwat).

and thirty-two women, as well as a white umbrella and a land grant, clearly indicating the lax moral atmosphere of the temple and the luxurious life led by ascetics.

art historical evidence of pāśupata temple-worship from southeast asia Visual and epigraphic evidence suggests that Pāśupata Brahmins in the Khmer courts may have participated in dance- and music-based temple worship. The Mỹ Sơn E1 temple pedestal from early 8th-century Campā (Fig. 6.1) is significant as its relief carvings show ascetics—probably Pāśupatas—performing meditation, preaching, conversing, dancing, playing musical instruments, and holding a book. It gives us a unique glimpse of the daily

06_CreSouthV1_8P_11Feb22.indd 199

practices of Śaiva ascetics,34 and since these activities are depicted on the pedestal that hosted a liṅga, it can be assumed that they have an implied association with the liṅga and its worship.35 The book carried by an ascetic figure on the pedestal, and other such ascetic figures on Khmer temple reliefs (Fig. 6.2), has been recently interpreted by Chemburkar and Kapoor (2018: 51) as an emblem of the Pāśupatas—possibly a copy of the Pāśupatasūtra itself, which might have represented a seal of their missionary nature and their role in the transmission of Śaiva knowledge.36 34. For a detailed account of the art historical evidence of Pāśupata practices as depicted in Khmer temple reliefs, such as Pre Rup and East Mebon, see Chemburkar and Kapoor 2018; cf. Haendel 2005: 213–252. 35. On the importance of liṅga for the Pāśupatas, see R.G. Bhandarkar 1913: 119–124; Chakraborti 1970: 194–203. 36. The existence of the Pāśupatasūtra in Cambodia is suggested by some echoes of that text found in Sanskrit inscriptions from Cambodia, e.g. K. 1049 (see Goodall 2015).

11/2/22 3:36 PM

Swati Chemburkar

200

Fig. 6.2: Ascetic figures seen holding a book at 11th-century Baphuon temple, Enclosure III, southern gopura-inner side facing north (a); 12th-century Angkor Wat, South western corner pavilion (b); late 12th–13th century Bayon, inner gallery, south wall between BY25 and BY26 (c), and Banteay Chhmar, east gallery, north BC 162 temple (d).

Long-haired, bearded holy men are depicted on a sandstone lintel from the ‘S’ and ‘N’ group of Śaiva temples of Sambor Prei Kuk. One such lintel depicts six ascetics with tied up or braided hair and beards, wearing simple dhotis and sacred threads, playing instruments, and dancing ecstatically in celebration (Fig. 6.3). Given the importance for the Pāśupatas of the observance of bathing in ashes, as well as the cleaning of the hair-bun with ashes, it does not seem far-fetched to identify the ascetics depicted on the Sambor Prei Kuk lintel, whose hair is tied up in

06_CreSouthV1_8P_11Feb22.indd 200

buns, as Pāśupatas. This view is also supported by inscription K. 279, mentioning the quantity of ash needed for ācāryas and elderly ascetics of a brāhmaṇāśrama, to be identified as Pāśupatas, to clean their buns (jaṭāśuddhi).37 37. See Cœdès’ table (1932: 97–108) comparing the inscriptions issued by Yaśovarman I (r. 889–c. 910) mentioning these āśramas (at, respectively, Prasat ong Mong [K. 290], Prāsāt Komnap [K. 701], Prei Prāsāt [K. 279]). Estève (2009: 339–345) has added another Śaiva āśrama inscription (K. 1228) to the table, and argued that both the brāhmaṇāśrama and the śaivāśrama hosted Pāśupata ascetics.

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

201

Fig. 6.3: Lintel depicting Śaiva ascetic celebrants, found at Sambor Prei Kuk- S1, 7th century, National Museum of Cambodia, Phnom Penh (Ka. 1748).

An early image of Khmer kings and their Brahmin priests is found on a 7th-century lintel from Wat Eng Khna (Fig. 6.4), where a group of men identifiable as ‘Brahmins’—insofar that they are bearded and wear a sacred thread—is performing what may be a consecration ritual for a king, seated directly below a śivaliṅga. A group of musicians and male dancers accompany these Brahmins. On top of the lintel, we see the liṅgodbhavamūrti, which narrates the supremacy of the liṅga. The intended message seems clear: ‘the newly consecrated king ruled with the authority of Śiva, invoked according to a ritual enacted for his benefit by the ṛṣis’ (Guy 2014: 166). Given the importance of liṅga worship with dance and music rituals for the Pāśupatas, we can assume these ascetics to be Pāśupatas. Furthermore, it must be pointed out that both the communal liṅga-worship and the liṅgodbhava myth are described in early, Pāśupata-influenced Śaiva sources like the Niśvāsamukha and the Śivadharma (see Kafle 2013; cf. Acri 2018: 3).38 Pre-Angkorian inscriptions mention donations of singers and dancers to Śaiva temples by priests

38. Sanderson (2003–2004: 422) claims that the Niśvāsatattvasaṃhitā was known in ancient Cambodia: cf. the depiction of thirty-two hells on the walls of Angkor Wat. Further, see Goodall and Golzio 2011: 58–59, fn. 10. K. 532B, v. 36 (Finot 1925: 359) mentions a Śaiva ascetic named Śivācārya mastering two Śaiva treatises, the Niśvāsa and Sarvajñānottara (Chhom 2016: 215).

06_CreSouthV1_8P_11Feb22.indd 201

and kings.39 This tradition continued during Angkorian times, and may reflect cultic and socio-religious evolutions from an originally Pāśupata tradition to a Mantramārga tradition. According to inscription K. 713, Yaśovarman (889–910) offered a great number of beautiful dancers, singers, reciters, musicians and handsome, mature men, skilful in dance and other arts, to the temples of Preah Kô and Lolei.40 He himself was an accomplished dancer.41 Rājendravarman’s guru, Yajñavarāha, was well versed in the art of music and drama.42 Sūryavarman I’s (1011–1050) guru Sadāśiva—a typically Saiddhāntika name—was a skilled musician and donated ‘hundred beautiful women accompanied by the wonderful sound of hundred 39. K. 51 (IC, V: 14); K. 138 (IC, V: 18–19); K. 137 (IC, V: 115–118); K. 155 (IC, V: 64–68); K. 713 (IC, I: 28); K. 270, vv. 8–25 (IC, IV: 69–70); K. 600 (IC, II: 21) and K. 323A, v. 63 (Bergaigne 1893: 391). 40. K. 713, vv. XXXV, XXXVI (Barth 1885: 28) and K. 323A, v. 63 (Bergaigne 1893: 399). 41. K. 323A, v. 51 (Barth 1885: 398) and K. 282C, v. 27 (ibid.: 483). 42. K. 842, v.14 (IC, I: 149). On the basis of this inscription where Dāmodara is mentioned as Yajñavarāha’s father and the name of a certain Dāmodaram engraved on the column of the Virūpākṣa temple in Karnataka, Vasundhara Filliozat (2015: 283) suggests a possible Pāśupata link, namely that the former might have been an Indian Pāśupata immigrant (inscription no. 10 on a north-eastern pillar, western faced, at the eastern entrance of the Lokeśvara temple at Virūpākṣa).

11/2/22 3:37 PM

202

Swati Chemburkar

Fig. 6.4: Lintel depicting the liṅgodbhavamūrti myth and a royal coronation ceremony. Mid-7th century, Wat Eng Khna, Central Cambodia, National Museum of Cambodia, Phnom Penh (Ka. 1774).

female musicians playing flute and stringed in- depicted on an unprecedented scale, notably at the struments, as well as fifty musicians with copper Bayon, his Buddhist state temple.47 cymbals, drums and other percussion instruments’ When one looks at these inscriptional records, 43 to the temple. One of the most celebrated Khmer the offerings of dancers and musicians by the early priests, Divākarapaṇḍita, purohita of Jayavarman (possibly Pāśupata) ācāryas are intriguing, while VI (r.1080–c. 1107), Dhāraṇīndravarman (r.1107–c. the donations by later Śaiva rājapurohitas or rāja1113), and Sūryavarman II (r.1113–c. 1150), was gurus reflect analogous developments in the Indian known for installing dancers, singers, musicians subcontinent, and especially South India. The fact and jesters as a part of a daily worship at Wat Phu, that priests received donations of dancers, singers Preah Vihear, and Phnom Sandak as well as ren- and musicians for the temple opens up a series ovating a dance hall. At Preah Vihear, he erected of interesting questions. Who were these ācāryas? a gold image of dancing Śiva.44 Dance- and mu- What was the nature of the relationship—i.e. as sic-based worship remained a significant aspect of owner or teachers—between the ācāryas and the Khmer temples, as 13th-century Jayavarman VII dancing girls? And, since all the records mention offered two gold dancing Śivas to the śivaliṅga ‘dancers’ and not just women, female slaves or girls, installed in the (primarily Buddhist) temple of were they already trained or maintained somePreah Khan.45 The dancing Śiva remained the most where as dancers? common image of Śiva in Khmer reliefs, along with Śiva as an ascetic (Roveda 2005: 163). Jayavarman Javanese and Khmer art historical material VII installed 615 female dancers in Ta Prohm (a Buddhist temple dedicated to his mother),46 1,000 While it will not be possible to fully answer those dancers in the Preah Khan (dedicated to his father), questions here, an analysis of relevant reliefs on and 1,622 dancers in other temples throughout his Khmer temples and their comparison with epkingdom. Under him, female dancing figures were igraphical data is in order. But before doing so, it is worth looking at comparative evidence from another strongly Śaivized region of Southeast Asia, 43. K. 235, vv. 72, 112–113 (Finot 1915: 83–88; Cœdès and i.e. Java. Enigmatic ascetic figures depicted on the Dupont 1943: 56; Bhattacharya 2009: 123). walls of some Central Javanese temples as a part of 44. K. 235, K. 194, K. 383 (Cœdès and Dupont 1943). dancing scenes featuring female dancers were first 45. ‘The two Lords of the dance, made of gold, placed by the king before the Serpent’ (vv. A 59–60; see Maxwell 2008: 21 for translation). 46. K. 273, vv. 64–67, 87 (Cœdès 1906: 77–78).

06_CreSouthV1_8P_11Feb22.indd 202

47. For Preah Khan (v. 76) and Bayon (v. D24), see Maxwell 2008: 51, 69.

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

203

Fig. 6.5: Borobudur relief—Third Gallery, north side, west end, III 65.

analysed by Willem Stutterheim (1956 [1935]). He discussed the many ascetic/Brahmanical figures at Borobudur and asked ‘what are we to think of this holy man, who, judging by his beard, moustache, top-knot, [and sacred thread] should be in a hermitage rather than in a dancing scene?’ (Stutterheim 1956: 94) (Fig. 6.5). Ascetic figures also appear on the reliefs of Candi Sari and Candi Śiva at Prambanan. Following the textual and visual material discussed by Stutterheim, Becker (2004: 176–177) argued: In addition to the firm textual evidence of their presence in Java, there is less-firm but suggestive evidence of the involvement of Pāsupata monks in performance traditions. The reliefs of a Śaivite priest dancing and singing or reciting in the company of dancing women on temple reliefs at Borobudur and Prambanan (Stutterheim 1956: 93) may indicate Pāsupata monks in the “marked” or first stage of spiritual practice.

Acri (2014) notes that some characters directly involved in dance, like those depicted on Candi Sari,

06_CreSouthV1_8P_11Feb22.indd 203

evidently represent religious figures whose iconography corresponds to the Ṛṣi/Pāśupata or Siddha typology.48 At the same time, he argues that other figures, especially those represented on Borobudur, may represent a more general Brahminical type associated with dancing girls: The Brahman appearing in the dancing scenes discussed by Stutterheim calls to mind the master (svāmī), leader (nāyaka), or stage director (sūtradhāra) known in classical Indian theatre and dance. This figure, who was the actual ‘proprietor’ of troupes, took part in the performances as lead singer and keeper of the rhythm for the dancers ... and may have fulfilled both functions, i.e. that of the master of the troupe/dancing girls and religious practitioner. (ibid.: 32)

Later ‘survivals’ of religious practitioners of the Kāpālika or Siddha type may be the 19th–20th-cen48. Pāśupatas were undoubtedly an important reality in the socio-religious landscape of premodern Java, as they are mentioned in many Old Javanese texts; see Acri 2014, 2018.

11/2/22 3:37 PM

204

Swati Chemburkar

Fig. 6.6: Śaiva ascetics are depicted in a hermitage setting in a company of women/dancers on the southwest corner pavilion of Angkor Wat.

tury officials called chaṇṭang balungs (‘rattling bones’), who performed antinomian farces, involving the drinking of alcohol, and supervised dancing girls-cum-prostitutes in the Islamic princely courts of Surakarta (Acri 2014; Becker 2004: 178; Stutterheim 1956). Returning to the Khmer reliefs, there are many parallels with the Javanese material mentioned above. Relevant depictions of ascetics or Brahmins in the company of women or involved in performance scenes are found at 11th-century Baphuon, 12th-century Angkor Wat, Phnom Rung, and the 13th-century Bayon temple. Three reliefs, one from Angkor Wat and the other two from the Bayon, possibly depict ṛṣis in a hermitage setting along with women (Figs. 6.6, 6.7 and 6.8). We may compare this visual evidence with the epigraphic record, namely an inscription from Yaśovarman’s reign mentioning musicians, singers, and instrument players assigned to a Śaiva hermitage (āśrama), and specifying that the functionary who acted as the head of Śiva’s worship was accompanied by actors,49 presumably for the 49. See 10th-century Bhadrāśrama inscription K. 450, vv. 32–42 (IC, III: 114).

06_CreSouthV1_8P_11Feb22.indd 204

staging of a performance as a form ritual. Early 12th-century inscription K. 364 mentions the constant sound of recitation of sacred texts and harmony of instrumental music from an āśrama, and frequent visits by ṛṣis and apsarases (Finot 1912: 24, 27). Cœdès’ comparative table on religious āśramas in Angkor mentioned above refers to the fact that relationship with a woman, including the legitimate wife, even outside the hermitage is strictly prohibited for Viṣṇuite hermits of the āśramas, and similar prescriptions applied to Buddhist monks;50 however, no such prohibitions are prescribed for Śaiva hermits. Does this reflect an ideal conduct by Śaiva priests, whereby female dancers were regarded as a part of ritual worship as the accepted norm? The Khmer inscriptional records make it clear that these āśramas were far from being simple hermitages. In fact, they functioned as monastic schools with a significant number of resident Brahmins/ascetics and lay people. Did dance and music form a part of the curriculum? In order to answer this question, it may be worth to draw a comparison with India. Śaiva monasteries of Tamil Nadu, 50. K. 701, v. 84; K. 290, v. 71 (Cœdès 1932: 84–112).

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

205

Fig. 6.7: Śaiva ascetic in a hermitage setting in a company of woman/dancer on the inner gallery-north wall (side wall of the BY33) of the Bayon.

06_CreSouthV1_8P_11Feb22.indd 205

11/2/22 3:37 PM

206

Swati Chemburkar

Fig. 6.8: Śaiva ascetic on the far right is seen in a company of woman/dancer, possibly in a hermitage setting (deer on the lower register is the indication) on the inner gallery-west wall (between BY29 and BY30) of the Bayon.

Southern India, did include singing and dancing along with the teaching of Vedas and Śāstras in the late 11th–early 12th century (see Inscriptions nos. 390 and 565 in Annual Report on Indian Epigraphy 1916: 134). In 13th-century Madurai, Southern India, ascetics (kovaṇvar) were charged with the performances of Tiruvempāvai and temple women were given the responsibility of singing hymns indicating the involvement of ascetics along with women in performance-based worship.51 Similarly, in 12th–13th-century Andhra Pradesh, Devi (1960: 61) has noted the close proximity of the large settlements of temple women and Brahmin priests in certain towns, possibly indicating their interde51. See the transcriptions of inscriptions ARE 192 of 1943–1944 and ARE 143, 144, 149 and 161 of 1940–1941 in Davis and Orr 2007: 80.

06_CreSouthV1_8P_11Feb22.indd 206

pendence and involvement in temple worship.52 It is therefore not impossible that analogous activities were carried out in temple settings in the Khmer domains around the same time, and especially during the reign of Jayavarman VII between the 12th and 13th century. The entrances of Jayavarman VII’s temples, such as Ta Prohm, Banteay Kdei, Bayon and Preah Khan of Angkor, house a ‘hall with dancers’ structure that has finely carved reliefs of several hundred female dancers. Just after crossing this structure at Preah Khan, we see several dancing Śivas or Śaiva 52. These large settlements were found at Draksharama, Palakol, Srikakulam, Chebrolu and Velpur. For inscriptions on temple women and priests, see no. 989 (South Indian Inscriptions, vol. 5: 375) and nos. 107, 110, 115 (South Indian Inscriptions, vol. 10: 48, 50, 52).

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

207

Fig. 6.9: (Left and centre) Dancers’ reliefs on the columns of ‘hall with dancers’ structure at Preah Khan of Angkor; (right) depictions of dancing Śiva or ascetics on the columns of the structure PK 54 placed after the hall with dancers.

ascetics carved on the pillars (Fig. 6.9). The entire inner gallery of the Bayon temple is filled with the depictions of Śaiva ascetics, indicating their importance in the Buddhist court of Jayavarman VII (Fig. 6.10). Khmer Buddhism, especially during Jayavarman VII’s time, did not at all exclude Śaivism, especially for certain royal or temple rituals.53 The 53. See Stern 1927: 31–32. Inscription K. 288, v. 79 (IC, IV: 216) of Prasat Chrung, Angkor Thom, compares Jayavarman VII to a Śaiva guru in his role of saviour of souls through initiation, and the benefit of his governance is said to be similar to the benefit bestowed on his disciples by a Śaiva guru: cf. Sanderson 2003–2004: 430, who argues that Jayavarman VII too was initiated into Śaivism in order to remove the stain (mala) that prevents the soul from experiencing the deployment of Śiva-ness. As Kamaleswar Bhattacharya (1997: 49) has observed: ‘The old Hindu cults continued under Jayavarman VII. In fact all the local cults, whether Hindu or Buddhist, were reunited in the Bayon… The king and his dignitaries erected numerous Hindu images at Preah Khan of Angkor and Banteay Chhmar… The old Hindu families retained their traditional privileges.’ He further adds that ‘Hearing that Cambodia was still full of excellent scholars of Vedas, a Brahman from Narapatideśa (Burma) went there to display his knowledge. King Jayavarman VII made him his Chaplain.’ See 13th-century inscription K. 567, v. 23 (Finot 1925: 393).

06_CreSouthV1_8P_11Feb22.indd 207

Brahmins’ roles included worship of deities with dance and music. They would have been active in a large wooden ‘hall with dancers’ at the main entrance of the Bayon.54 Under him, the female dancing figures occur on an unprecedented scale, reproduced in thousands of times in the reliefs of the Bayon.55 Davidson (2002: 223) maintains that the Pāśupatas may be the origin for the diffusion of singing and dancing in Buddhism, especially in reference to esoteric Yoginītantras.56 Pāśupatas also seem to have exerted some influence on Mahāyāna Buddhists, in whose ritual music played an important

54. For a detailed discussion on this structure and its possible ritual function, see Chemburkar 2015. 55. Sharrock (2007: 260) calls this final phase of Jayavarman VII’s architectural decoration with 6,250 dancing female figures carved into the entrance pillars and gopura friezes of the Bayon ‘yoginification’. 56. Pāśupatas were so influential for early Esoteric Buddhist circles that Avalokiteśvara is depicted wearing the dress and attributes of Śiva Paśupati. Vajrayāna Siddhas were, for all the appearances, the first Buddhists to employ singing (not chanting) and dancing (not simple hand gestures) in the acts of offering before images.

11/2/22 3:37 PM

Swati Chemburkar

208

Fig. 6.10: Sandstone fragment of the Bayon pediment depicting ascetics and dancers. National Museum of Cambodia, Phnom Penh, Accession no. 1692.

role.57 It is not impossible that these originally Śaiva modes of worship were carried forward in Buddhist milieus of the ancient Khmer domains.

57. The Saddharmapuṇḍarīkasūtra states that the worship is done with songs, instrumental music, dance, cymbals, and singing: see Schopen (2005: 38, 134) and Hara (1966: 136–137), who has noted that some passages inspired by Pāśupata doctrines also appear in that Buddhist scripture. For a mention of different forms of music and dance employed in a Buddhist temple, see the 12th-century Gaya inscription (Indraji 1881).

06_CreSouthV1_8P_11Feb22.indd 208

Dancers, singers, musicians, and other entertainers In addition to the dancers who carried out rituals and recitations on a daily basis in the temples, many more performed during the festivals. Temple inscriptions in Old Khmer often mention the Sanskrit word utsava or festival along with the descriptions of various spectacles inside the temple complexes. Words such as pañcotsava (five religious festivals), mahotsava (the great festival), and mahānavamī (the Hindu festival of Dussehra) occur as royal celebrations throughout the Ang-

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

209

mentions the commencement of the phālguna festival every year when 122 deities were displayed. These would be festival replicas brought in on palanquins from other temples in Angkor and received at the entrance of the temple with much fanfare. We know that images of dancing Śiva were indeed carried in palanquins across Angkor. On the basis of the inscriptional record (e.g., K. 155, IC V: 65), Pou (1997: 246) has suggested that the dancers of a ballet group performed for effigies of Śiva during ceremonies held in the temple precincts. The early 11th-century Old Khmer inscription K. 276 from Prasat Ta Keo lists among the temple donations of a scholar named Yogīśvara ‘a palanquin in which is placed the ten-armed Lord Nāṭakeśvara with all his ornaments’.58 The Preah Khan inscription stele mentions how Śiva Bhadreśvara and Pṛthuśaileśvara were carried to Preah Khan every year for the Phālguṇa festival (Maxwell 2008: 72, 73). These records remind us of the grand festivities and processions held during the Cōḻa in Tamil Nadu, who were in contact with Angkor.59 Aghoraśiva’s mid-12th-century ritual guidebook Mahotsavavidhi, composed during the reign of Rājarāja Cōḻa II, describes the temple festivals and the procedures to be followed.60 Of all the performers, the ones who play the most prominent role in the festival are the rudragaṇikās (Śiva’s female retinue/courtesans), the female temple dancers. In the procession, rudragaṇikās accompany the ācāryas (Davis 2014: 159). The Ajitāgama (27.157) speaks of rudragaṇikās as lovely bodied women of Śiva’s abode who wear garlands and ornaments.61 The Fig. 6.11: A sword swallower along with a drummer and cymbal player. Leper king terrace, north wall.

korian period. A relief from the walls of the ‘Leper King Terrace’ in Angkor’s Royal Plaza depicts a sword swallower accompanied by a cymbalist and a drummer (Fig. 6.11). During Jayavarman VII’s reign, the annual rice festival, according to Zhou Daguan’s account, was the greatest annual celebration of earth’s fecundity, where rice was burned as an offering to all the Buddhas in Angkor. The Preah Khan inscription

06_CreSouthV1_8P_11Feb22.indd 209

58. K. 276 (IC, IV: 154–155). The festival images were no doubt carried from their home temples in the provinces and elsewhere in Angkor on palanquins, as depicted on the walls of Banteay Chhmar. 59. For a discussion on parallels in the domain of temple architecture and performance-based worship between the Cōḻa court and Angkor from the 10th to the 13th century, see Chemburkar 2015. 60. The Mahotsavavidhi is a Śaiva ritual manual composed by the ācārya Aghoraśiva during the reign of Cōḻa King Rājarāja II in 1157 CE to guide a priest in the performance of a nine-day temple festival. A translation may be found in Davis 2014. 61.  Quoted in Davis 2014: 26.

11/2/22 3:37 PM

210

Swati Chemburkar

Fig. 6.12: Historic procession on the south gallery wall of Angkor Wat depicting dancers, flutist, singers and drummers leading the procession (top two images); Śaiva priests are depicted (lower two images) holding a book or a bell, and some are seated with crossed arms.

Rauravāgama (ch. 19) adds that they should be pure, calm, beautiful, young and well trained in dance (ibid.). According to the Mahotsavavidhi, the rudragaṇikās perform throughout the festival and often carry out key ritual practices (ibid.: 26). The question to explore is whether we can regard the Khmer dancers as counterparts of the Indian gaṇikās/rudragaṇikās, and whether these dancers are regular members of the temple staff or professionals drawn from the surrounding community.62 62. The Śivadharmottara (12.168) refers to free women who ‘could be forced as punishment for unspecified offences to become Rudragaṇikās (also called Rudrakanyās or Rudranārīs), female temple slaves of a superior class whose duty was to gratify the deity with dancing… They are superior to other persons who are the god’s property’ (Sanderson 2003–2004: 399, fn. 181). Goodall’s (2018) work on Rudragaṇikās discusses the 16th-century Āśaucadīpikā,

06_CreSouthV1_8P_11Feb22.indd 210

Indian literature conventionally considers names ending in Senā, such as Vasantasenā (heroine of the classical Sanskrit play Mṛcchakaṭikā), Madanasenā and Mādhavasenā (in the 11th-century Sanskrit play Ubhayābhisārikā of Vararuci) to be courtesans of high social status.63 Similar names occur a text on ritual impurity based on a collation of four rather corrupt sources. V. 9c states: ṛṣibhāryā varārohā mama saṃparkataḥ punaḥ, tais tyaktā mama gehe tu tāḥ smṛtā rudrakanyakāḥ, ‘Rudrakanyās are held to be fine-hipped wives of the sages whom those [sages] abandoned in my temple, because of their contact with me’ (quoted from Goodall 2018: 117). 63. The Ubhayābhisārikā contains multiple mentions of courtesans, such as Nārāyaṇadattā, Devadattā and Anaṅgadattā (Venkatacharya and Warder 1967). Nāṭyaśāstra (ch. XIX, v. 33) explains that courtesans’ names should end in -dattā, -mitrā and -senā (for a translation of the passage, see Ghosh 1950: 345; cf. Chandra 1973: 97). Goodall (2018:

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor in Khmer epigraphy such as Priyasenā (a dancer), Samarasenā (dancer), Madhurasenā (musician who plays an instrument).64 The Mahotsavavidhi account of the festival, describing rudragaṇikās, matches well a relief of Angkor Wat (Fig. 6.12): In front of the god are dancers and flutists, singers and drummers. On both sides are the Rudragaṇikās and others engaged in dance. In front come the Mahāśaivas, and behind come brahmins intoning the sounds of the Veda and other sacred hymns. After them come Maheśvaras belonging to all four Classes … Pleasing god this way, on the first day in the morning, Śiva rides on a cart.... (trans. Davis 2014: 89)

In the Bayon reliefs, dancers and musicians perform on nāga balustrade terraces, in floating pavilions, or pillared halls. Pou’s study of dance and music-related inscriptions categorizes proper names based on the specific performances, achievements, technical competence, and physical qualities: see e.g., Tīttaru (one who dances like partridge), Vṛttavalī (one who moves in row, in circle), Raṅgaśrī (pearl of theatre), Carumatī (pretty as parrot), Payodharī (having beautiful breasts) or Suvibhrama (displaying beautiful coquetry). It’s an illuminating account of the sociological connotations of these dancers and musicians (Fig. 6.13). The Sanskrit Śaivāgamas talk about the qualifications expected by these performers as well as their ritual ‘masters’. Uttarakāmikāgama specifies the masters of music should be knowledgeable about musical instruments, compositions, songs, and recitations as well as nine moods of drama (Davis and Orr 2007: 81, fn. 6). 10th-century Old Khmer inscription K. 181 mentions an elite man upādhyāya Thṃon, or a master of percussion music.65 Pou (1997: 236) observes a harp-like instrument being played by court ladies, celestial beings and 102–103) discusses the names ending in -senā as typical courtesan names found in Indian literature. 64. K. 600E (IC, II: 22); cf. Zakharov 2019. See Saveros Pou (1997: 244) for the names of singers and dancers in Khmer epigraphy. 65. K. 181A, v. 8 (IC, VI: 140); cf. Pou 1997: 243.

06_CreSouthV1_8P_11Feb22.indd 211

211

learned men or paṇḍita, who were apparently accompanying reciters and singers (Fig. 6.14). Vīṇa appears in many Khmer inscriptions as an offering, an instrument, and also as a proper name Kaṃsteṅ or Sir Vīṇa.66 Another term guru laṃpeṅ or master performer that appears in the Khmer epigraphy has a strong ritual connotation, as argued by Pou.67 After the 10th century, we come across Vāditrapura or the ‘town of music’ in the Khmer epigraphy.68 What exactly did the town contribute? Music and dance had acquired an institutional status in Angkor as we see from the term Varṇa rmāṃ or corporation of dancers. One such corporation belonged to a lady Tāñ Pañ, who as the text adds, was in charge of the concubines (khloñ snaṃ) of the wealthy high-ranking author of the inscription.69 The head of these corporations are all attested with honorific titles such as Vāp or Sir, for e.g., Vāp Ānanda, Vāp Rājadāsa, and held high positions in the court.70 Inscriptional records mention a king or an official offering rpam or ballet group to the gods. These records mention dancers, musicians and instrument players by Sanskrit names with complimentary meanings rather than the derogatory Khmer names.71 K. 831 mentions dancers and other temple personnel who were in active service for half a month at a time, with music performed at least three times daily or even, as we shall see, ‘without cease’.72 66. K. 205, v. 14; IC, III: 5. 67. K. 254B, vv. 19, 23 (IC, III: 185). It mentions guru laṃpeṅ, which Cœdès translates as unpaid or voluntary guru, but Pou (1997: 234) argues for him being a game performer with a ritual connotation. 68. K. 554, v. 5 (IC, II: 14); K. 532B, v. 16 (Finot 1925: 354). 69. The Khmer word pedā (deriving from Skt. peṭaka) occurs in the inscription K. 155 (IC, V: 64), which Pou (1997: 246) translates as ‘company’ or ‘troop’, and thus pedānātaka or pedānatta would be a ‘group or company of dancers’. 70. K. 831, vv. 6–8 (IC, V: 147); Pou 1997: 240. 71. For the names such as Vasantamālikā, Tanvaṅgī, Gandharvagītā or Sakhīpriyā, see inscriptions K. 137, v. 55 (IC, II: 115–118); K. 557, vv. 33–34 (IC, VIII: 166) and K. 155 (IC, V: 64). See Pou (1997: 244); cf. Jacob (1993: 129–130). Vickery (1998: 231) argues that ‘the examples are too few to decide the status of the musicians and dancers and questions whether it differs from one place to another.’ 72. K. 831 (IC, V: 147).

11/2/22 3:37 PM

212

Swati Chemburkar

Fig. 6.13: Bayon reliefs depicting dance and music in different settings: (top) inner gallery, east wall between BY22 and BY37; (middle) inner gallery, south wall between BY27 and BY28; (bottom) inner gallery-east between wall BY36 and BY37.

06_CreSouthV1_8P_11Feb22.indd 212

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor

213

Fig. 6.14: (Top) An ascetic is seen playing a harp like instrument and accompanying a dance at the Bayon temple, inner gallery, north wall between BY33 and BY34; (bottom) a sage is reciting from the Rāmāyaṇa with Śiva in the centre and Hanumān, Sītā, and Rāma with a bow. Banteay Chhmar, pediment in the ‘hall with dancers’ (BC 54).

06_CreSouthV1_8P_11Feb22.indd 213

11/2/22 3:37 PM

214

Swati Chemburkar

The Ta Prohm inscription details how Jayavarman VII organized the performances by dancers attached to the temple, possibly referring to dramatized Buddhist stories.73 Epigraphy mentions words such as vāca, kinnara, and gandharva, indicating reciters performing divine services, a class of performing artists and singers in the temples.74 These records refer to singers by proper names such as Sugītā (one who sings beautifully), Haṅsavādī (one with a goose voice), Kaṇṭhagītā (one with a guttural voice) and Gāndhārasvanā (producing the third primary note-gāndhāra of music). Based on her analysis of relevant Khmer inscriptions, Pou argues for the gandharvas of ancient Cambodia as being human rather than semi-divine figures.75 These are well known in Indian culture as celestial musicians playing string instruments. Gāndharvavidyā or science of music was known to the Khmers, as can be evinced from the Banteay Srei inscription.76 Along with gandharva and kinnara, Pou has analysed the Khmer expression caṃryaṅ Śikharā and argued that caṃryāṅ were entertainers of the small temple-community, of the bard type.77 The most popular character in the Khmer Rāmakerti version of the Rāmāyaṇa theatre is a 73. The king paired the Buddhist practice of dāna (charity) and śīla (proper conduct) with the performances of dancers. See K. 273, v. 87 (Cœdès 1906: 77–78). 74. Based on the inscription K. 356 (Cœdès 1911a: 400– 404), Pou (1997: 242) translates the word vāca as a reciter performing a divine service. For Kinnara, see K. 327A, v. 29; K. 327B; K. 331B (Bergaigne 1893: 319) and Bhattacharya 2009: 79, 81. For gandharva, see K. 842, sv. 20 (IC, I: 147); K. 129 (IC, II: 83); K. 155 (IC, V: 64). 75. Inscription No. 264 at the Darasuram temple mentions a musician Perundanattu-Gandharva; see http://www. whatisindia.com/inscriptions/south_indian_inscriptions/ darasuram/introduction.html. Another inscription (no. 306) from the same place gives us an idea of the patronage accorded to the performing arts by royalty. It refers to the appointment of an abhinaya-nattuvanar in addition to others in the temple for performing what was called the ahamarggam, an expression style of dance as distinct from action nṛtya. 76. K. 842A, v. 20 (IC, I: 147). See note 30 above for the remark that Pāśupatas were prescribed to perform singing according to (or against) the rules of the Gāndharvaśāstra. 77. In this case it would be a group of people playing Śikharā or string instrument (Pou 1997: 242). Another example is caṃryyāṅ stuti, or a group of singers.

06_CreSouthV1_8P_11Feb22.indd 214

Fig. 6.15: Possible depiction of a buffoon in the victory parade at the Bayon (inner gallery, north wall, between BY32 and BY33).

buffoon, which traces its origins to the Sanskrit word bhaṇḍa (buffoon or vidūṣaka). Groslier (1921: 100, 125, 128) has detected the presence of such figures in parades and processions depicted in the bas-reliefs of Khmer temples, but they also played a role in the ritual offerings for the daily worship78 (Fig. 6.15). 78. At Preah Vihear for Śikhareśvara and at Phnom Sandak for Śivapura Danden, priest Divākara had offered jesters along with the dancers and musicians for the daily worship: see Cœdès and Dupont 1943: 148–149. In spite of pointing out the presence of such figures, Groslier never

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor This character may hark back to the buffoon of the classical Indian theatre, who was a Brahman yet at the same time was ugly and ridiculous. It may be relevant to mention here that critical references to Pāśupatas and, especially, Kāpālikas found in mainstream literary texts often portray them as buffoons and performers holding drums and trumpets (Lorenzen 2000: 83). As noted by Acri (2014: 41, fn. 89), a figure mixing the characteristics of the vidūṣaka with those of Śaiva tantric antinomian practitioners may be the buyut mentioned in the Middle Javanese chronicle Pararaton, carrying the honorific title hyaṅ, acting as a buffoon, master or partner of dancing girls, as well as a magician. A connection between buffoonery and supernatural/ magical power is also found in the character embodying a comedian and black magician known as bondres in sacred Balinese temple-dance and performance (ibid.). It is still difficult to determine whether these dancers, musicians, singers, and entertainers were members of the working population with special skills, or belonged to the elite class of ascetics or as a part of a laity that belonged to a particular sect. Nevertheless, they held an important position in the royal and sacred sphere.

temple dancers in 13th-century cambodia That temple worship activities in which female dancers played a role were thriving in the 13th century in both South and Southeast Asia—for instance, at Angkor after the reign of Jayavarman VII—is suggested by two contemporary historical accounts. One is the Description of the World by famous Venetian traveller Marco Polo (1256–1323), who visited Malabar in Southern India at the end of the 13th century. Polo makes interesting observations about the temples, priests, and dancers: And again I tell you that they have many idols in their monasteries [temple] men and women, identifies any particular figure in the bas-reliefs. My hypothesis regarding Fig. 6.15 is based on the depiction of the same figure on the west wall (BY29 and BY28) of the inner gallery, where people are making offerings to him.

06_CreSouthV1_8P_11Feb22.indd 215

215

to serve which idols many girls are offered in this way. For their mother and their father offer them to the idols … And when they have offered them, every time that the [Kālamukha?] monks of the monastery [temple] of the idol require those girls who have been offered to the idol to come to the monastery [temple] to make amusement for the idol, they come their immediately and sing and sound music and dance and make great festival…. and there are great numbers of these girls, for they make great company.79

During his Malabar visit, Marco Polo had spent some time at the port of Muṭṭupallī80 in Andhra, where Kākatiyas were ruling. Among the several schools of Śaivism, it was the Kālamukha, an offshoot of the Pāśupatas, that gained prominence during their reign.81 It appears that dance- and music-based temple rituals were performed by this current of Śaivism. It is likely that the account here is alluding to the institution of rudragaṇikā/ devadāsī in Śaiva temples. The other source of interest is the 1225 CE Chinese chronicle Zhufanzhi, by Zhao Rugua, the superintendent of maritime trade in Canton. Even though the chronicle is based on hearsay from traders and merchants rather than on visits to the countries engaged in maritime trade, it remains an important source of information on geography and ethnography of mediaeval mainland Southeast Asia. It narrates how twenty thousand dancing girls sang songs twice daily in honour of Buddha at four 79. Marco Polo, trans. Moule and Pelliot 1938: 393 (additions between brackets are my own). 80. Ibid.: 397. Nilakantha Sastri (1939: 174–175) argues that Mutifili mentioned by Marco Polo is the famous port of Muṭṭupallī indeed. 81. See Hanmkomḍa epigraph (Venkataramanayya 1974, nos. 15 and 35), mentioning a Rāmeśvara Paṇḍita who was proficient in the lakuleśvarāgamamahāsiddhānta; Saṅigaram inscription (Parabrahma Sastry 1974, nos. 19, 22, 24), mentioning Rudradeva as Parama-Maheśvara; Ablur inscription (Fleet 1898–1899: 220), mentioning the Kālamukha ācārya Someśvara Paṇḍita as being proficient in the lākulasiddhānta; Mallakapuram inscription (Pantulu 1930: 160), describing the activities of the Goḷaki maṭha in which Pāśupatas resided, which gained popularity among the masses while Kālamukhas lost their hold at the court; cf. Yazdani 1960: 700–708.

11/2/22 3:37 PM

Swati Chemburkar

216

thousand Buddhist establishments in Gujarat, and how Cōḻa kings retained ‘10,000 dancing girls’ in their temples during the 12th–13th century. Speaking of Angkor, Zhao refers to the devout Buddhists and 300 Khmer women called a-nan [Skt. Ānanda, bliss] offering dance and food to the Buddha in Chen-la [i.e. Cambodia] (Hirth and Rockhill 1911: 95, 100, 153). Were these dancing women from Vāp Ānanda’s troupe some local counterparts of the Indic gaṇikās?

conclusion On the basis of textual, literary, and inscriptional evidence from South and Southeast Asia, I have highlighted the importance of dance and music making in Śaiva devotional and ritual settings in the ancient Khmer domains. I have argued that ascetics belonging to the Pāśupata tradition of Śaivism and its further subgroups may have been involved in the performance-based ritual and worship in Angkor from the 7th century onwards, and that these modes of worship were further developed during the Angkorian period in Śaiva Mantramārgic milieus in a way akin to roughly contemporary South Indian realities. The involvement of Śaiva ascetics and/or priests in temple worship may have been strictly connected to the political sphere— for instance, royal coronations. While the textual evidence from the Sanskrit Śaiva corpus studied thus far by philologists is often characterized by an esoteric and prescriptive nature, which only gives us a partial view,82 a wider-ranging exploration of other sources of data suggest that the community on the ground was more diverse than we imagine.83 For instance, Bisschop (2010: 485) argues that not all Pāśupata adepts abided by the 82. Though Kauṇḍinya’s Pāśupatasūtra outlines and emphasizes a lifelong asceticism, manuals discovered by Diwakar Acharya (such as the Saṃskāravidhi: see Acharya 2007) demonstrate a more diverse religious landscape, involving both ascetics and the lay community. 83. A comprehensive study of the corpus of early Śaiva texts catering to the lay community of Śaiva devotees, including the Śivadharmottara and the Śivadharmaśāstra, will no doubt yield precious data to reconstruct the actual practices of the community of religious specialists and lay people.

06_CreSouthV1_8P_11Feb22.indd 216

strict ascetic practice prescribed in the texts, for there were some who served the needs of a larger Śaiva community, such as priests or ācāryas. The presence in Khmer art and epigraphy of these religious specialists, as well as of a community of lay devotees, is evident. The depictions of liṅga veneration on the temple walls and the involvement of ‘Brahmanical’ religious types in performance scenes, and again their association with female dancers, show varied groups of agents involved in Śaiva ritual practice—even in a Buddhist setting, such as the one of Jayavarman VII. While it is impossible to draw a clear timeline—for the line has twists, spirals, and vanishing points—, my chapter has made a preliminary effort to understand the specific Śaiva tradition of temple-centred, performance-based worship by the practitioners hailing originally from the Indian subcontinent, and has argued for its continuing existence in Angkor for more than 600 years. More importantly, given the presence of significant visual and epigraphic evidence, it has highlighted the creative role of Southeast Asia in the historical development of Śaivism—for it is precisely in this region that new forms of political patronage, temple ritual, and service to the lay community might have arisen and spread further across the Indic world.

References

Acharya, Diwakar. 2006. ‘The Role of Caṇḍa in the Early History of the Pāśupata Cult and the Image on the Mathurā Pillar dated Gupta Year 61’, Indo-Iranian Journal 48/3–4: 207–222. . 2007. ‘The Saṃskāravidhi: A Manual on the Transformatory Rite of the Lakulīśa Pāśupatas’, in Dominic Goodall and André Padoux (eds.), Mélanges tantriques à la mémoire d’Hélène Brunner, pp. 27–48. Pondicherry: Institut français de Pondichéry/École française d’Extrême-Orient. . 2018. ‘Pāśupatas’, in Knut A. Jacobsen, Helene Basu, Angelika Malinar, and Vasudha Narayanan (eds.), Brill’s Encyclopaedia of Hinduism. [Online publication: http: //dx.doi.org/10.1163/2212-5019_beh_COM_ 9000000056] Acri, Andrea. 2014. ‘Birds, Bards, Buffoons and Brahmans: (Re-)Tracing the Indic Roots of

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor Some Ancient and Modern Performing Characters from Java and Bali’, Archipel 88: 13–70. . 2018. ‘Performance as Religious Observance in Some Śaiva Ascetic Traditions from South and Southeast Asia’, Cracow Indological Studies 20/1: 1–30. Annual Report on Indian Epigraphy. 1916. Madras: Government of Madras, Public Department. Bakker, Hans. 2000. ‘Somaśarman, Somavaṃśa and Somasiddhānta: A Pāśupata tradition in seventh-century Dakṣiṇa Kosala: Studies in the Skandapurāṇa III’, in Ryutaro Tsuchida and Albrecht Wezler (eds.), Harānandalaharī: Volume in Honour of Professor Minoru Hara on his Seventieth Birthday, pp. 1–20. Reinbek: Dr. Inge Wezler Verlag für Orientalistische Fachpublikationen. . 2014. The World of the Skandapurāṇa: Northern India in the Sixth and Seventh Centuries. Leiden/Boston: Brill. Banerjea, Jitendra Nath. 1941. The Development of Hindu Iconography. Calcutta: University of Calcutta. . 1951. ‘Lakulīśa—The Founder or the Systematiser of the Pāśupata Order’, in The Proceedings of the Indian History Congress 14: 32–36. Jaipur. Barnett, Lionel D. 1915–1916. ‘Bankapur Inscription of the time of Someśvara I and the Kaḍamba Harikeśarin: A.D. 1055’, in Sten Konow and F.W. Thomas (eds.), Epigraphia Indica, vol. 13, pp. 168–197. Calcutta/Bombay: Superintendent Government Printing/British India Press. . 1925. ‘Inscription of Sudi F., of the Reign of Someśvara I: Śaka 981’, in F.W. Thomas (ed.), Epigraphia Indica, vol. 15, pp. 85–93. Calcutta/ Bombay: Government of India Central Publication Branch/British India Press. Barth, Auguste. 1885. Inscriptions sanscrites du Cambodge (extraits des notices et extraits des manuscrits de la Bibliothèque nationale 27, 1re partie, 1er fascicule), pp. 1–181. Paris: Imprimerie Nationale. Becker, Judith. 2004. Gamelan Stories: Tantrism, Islam, and Aesthetics in Central Java. 2nd ed. Arizona State University: Program for Southeast Asian Studies.

06_CreSouthV1_8P_11Feb22.indd 217

217

Bergaigne, Abel. 1893. Inscriptions sanscrites de Campā et du Cambodge (extraits des notices et extraits des manuscrits de la Bibliothèque nationale 27, 2e partie, 2e fascicule), pp. 181–632. Paris: Imprimerie Nationale. Bhandarkar, D.R. 1908. ‘An Êkliṅgjî Stone Inscription and the Origin and History of the Lakulîśa Sect’, Journal of the Bombay Branch of the Royal Asiatic Society 22: 151–165. Bhandarkar, R.G. 1913. Vaiṣṇavism, Śaivism and Minor Religious Systems. Grundriss der Indoarischen Philologie und Altertumskunde Bd. 3, Hft. 6. Strassburg: Trübner. Bhattacharya, Kamaleswar. 1955. ‘La secte des Pāçupata dans l’ancien Cambodge’, Journal Asiatique 243/4: 480–487. . 1997. ‘The Religions of Ancient Cambodia’, in Helen Jessup and Thierry Zéphir (eds.), Sculpture of Angkor and Ancient Cambodia: Millennium Glory, pp. 34–53. Washington/Paris: Thames and Hudson. . 2009. A Selection of Sanskrit Inscriptions from Cambodia. Phnom Penh: Centre for Khmer Studies. Bisschop, Peter. 2006. Early Śaivism and the Skandapurāṇa: Sects and Centres. Groningen Oriental Studies XXI. Groningen: Egbert Forsten. . 2010. ‘Śaivism in the Gupta-Vākāṭaka Age’, Journal of the Royal Asiatic Society 20/4: 477– 488. Bloomfield, Maurice. 1924. ‘On False Ascetics and Nuns in Hindu Fiction’, Journal of the American Oriental Society 44: 202–242. Chakraborti, Haripada. 1970. ‘Appendix I: On Liṅga Worship and the Pāśupatas’, in Pāśupata Sūtram with Pañchārtha-Bhāṣya of Kauṇḍnya, pp. 194–203. Calcutta: Academic Publishers. Chandra, Moti. 1973. The World of Courtesans. Delhi: Vikas Publications. Chemburkar, Swati. 2015. ‘Dancing Architecture: “Hall with Dancers” in Jayavarman VII’s Temples’, Journal of Southeast Asian Studies 46/3: 514–536. Chemburkar, Swati and Shivani Kapoor. 2018. ‘Pāśupata Sect in Ancient Cambodia and Champa: Art Historical Evidence’, in Võ Văn

11/2/22 3:37 PM

218

Swati Chemburkar

Thắng and Peter D. Sharrock (eds.), Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 45–57. Bangkok: River Books. Chhom, Kunthea. 2016. Le rôle du Sanskrit dans le développement de la langue khmère: Une étude épigraphique du VI au XIV siècle. PhD dissertation, École Pratique des Hautes Études. Cœdès, George. 1906. ‘La stèle de Ta Prohm’, Bulletin de l’École française d’Extrême-Orient 6/1: 44–86. . 1911a. ‘Études cambodgiennes V. Une inscription d’Udayādityavarman I’, Bulletin de l’École française d’Extrême-Orient 11/3: 400– 404. . 1911b. ‘Les bas-reliefs d’Angkor-Vat’, Bulletin de la Commission archéologique de l’Indochine. Paris: Imprimerie Nationale. . 1913. ‘Ètudes cambodgiennes VII–XI’, Bulletin de l’École française d’Extrême-Orient 13/6: 1–36. . 1932. ‘Études cambodgiennes XXX. À la recherché du Yaçodharāçrama’, Bulletin de l’École française d’Extrême-Orient 32/1: 84–112. . [IC] 1937–1966. Inscriptions du Cambodge, éditées et traduites, vols. I–VIII. Hanoi and Paris: École française d’Extrême-Orient. . 1954. ‘La stèle du Tûol Room Tim. Essai d’interprétation par la langue bahnar d’un texte juridique khmer du Xe siècle’, Journal Asiatique 242/1: 49–67. Cœdès, George and Pierre Dupont. 1943. ‘Les Stèles de Sdôk Kâk Thom, Phnom Sandak et Práah Vihār’, Bulletin de l’École française d’ExtrêmeOrient 43: 56–134. Dalal, Chimanlal D. 1966. Gaṇakārikā of Ācārya Bhāsarvajña with the Ratnaṭikā and four appendices including the Kāravaṇa-Māhātmya. Baroda: Oriental Institute. Davidson, Ronald. 2002. Indian Esoteric Buddhism: A Social History of the Tantric Movement. New York: Columbia University Press. Davis, Richard. 2014. A Priest’s Guide for the Great Festival: Aghorasiva’s Mahotsavavidhi (South Asia Research). New York: Oxford University Press.

06_CreSouthV1_8P_11Feb22.indd 218

Davis, Richard and Leslie Orr. 2007. ‘People of the Festival’, in Dominic Goodall and André Padoux (eds.), Mélanges tantriques à la mémoire d’Hélène Brunner, pp. 73–97. Pondicherry: Institut français de Pondichéry. Devi, Yasoda V. 1960. ‘The History of Andhra Country, Subsidiary Dynasties (A.D. 10001500)’, Journal of the Andhra Historical Research Society 26: 1–102. Estève, Julia. 2009. Étude critique des phénomènes de syncrétisme religieux dans le Cambodge angkorien. PhD dissertation, École Pratique des Hautes Études. Filliozat, Vasundhara. 2015. Śaiva Monuments at Paṭṭadakkal. New Delhi: ASI, e-publication. Finot, Louis. 1904. ‘Notes d’épigraphie: 11. Les inscription de Mī-sõn’, Bulletin de l’École française d’Extrême-Orient 4/1: 897–977. . 1912. ‘Notes d’épigraphie: 13. L’inscription de Ban That’, Bulletin de l’École française d’Extrême-Orient 12/1: 1–28. . 1915. ‘Notes d’épigraphie: 16. L’inscription de Sdok Kak Thom’, Bulletin de l’École française d’Extrême-Orient 15/2: 53–106. . 1925. ‘Inscriptions d’Angkor’, Bulletin de l’École française d’Extrême-Orient 25/3–4: 297–407. . 1928. ‘Nouvelle Inscriptions de Sambor’, Bulletin de l’École française d’Extrême-Orient 28/1: 43–46. Fleet, F.J. 1898–1899. ‘No 25, Inscriptions at Ablur’, Epigraphia Indica, vol. 5, pp. 213–264. Calcutta: Government of India Central Printing Office. Ghosh, Manomohan. 1950. The Nāṭyaśāstra: A Treatise on Ancient Indian Dramaturgy and Histrionics, Ascribed to Bharata Muni. Calcutta: Asiatic Society. Goodall, Dominic. 2015. ‘On K. 1049, a Tenth-Century Cave Inscription from Battambang and on the Sectarian Obedience of the Śaiva Ascetics of Non-Royal Cave Inscriptions of Cambodia’, UDAYA: Journal of Khmer Studies 13: 3–34. . 2018. ‘Rudragaṇikās: Courtesans in Śiva’s Temple? Some Hitherto Neglected Sanskrit Sources’, Cracow Indological Studies 20/1: 91–143.

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor Goodall, Dominic and Csaba Dezső. 2012. Dāmodaraguptaviracitaṃ Kuṭṭanīmatam. The Bawd’s Counsel, being an Eighth-Century Verse Novel in Sanskrit, by Dāmodaragupta. Groningen: Egbert Forsten. Goodall, Dominic and Karl-Heinz Golzio. 2011. ‘Book Review of Kamaleswar Bhattacharya (ed.), A Selection of Sanskrit Inscriptions from Cambodia’, Indo-Iranian Journal 54: 49–60. Groslier, George. 1921. Recherches sur les Cambodgiens d’après les textes et les monuments depuis les premiers siècles de notre ère; ouvrage illustré de 200 photographies et de 1153 dessins et plans inédits de l’auteur (voyages et missions du Cambodge de 1909 à 1904 et de 1917à 1920). Paris: A. Challamel. Guy, John. 2014. Lost Kingdoms: Hindu Buddhist Sculpture of Early Southeast Asia. New York: The Metropolitan Museum of Art. Haendel, Alexandra. 2005. ‘The Divine in the Human World: Sculpture at Two Tenth-Century Temples at Angkor’, Artibus Asiae 65/2: 213–252. Hara, Minoru. 1966. Materials for the Study of Pāśupata Philosophy. PhD dissertation, Harvard University. Harris, Peter (ed. and trans.). 2007. A Record of Cambodia: The Land and its People, by Zhou Daguan. Chiang Mai: Silkworm Books. Hirth, Friedrich and W.W. Rockhill (trans.). 1911. Chau Ju-Kua, His Work on the Chinese and Arab Trade in the Twelfth and Thirteenth Centuries, Entitled Chu-Fan-Chï. St. Petersburg: Imperial Academy of Sciences. Huber, Edouard. 1911. ‘Études Indochinoises’, Bulletin de l’École française d’Extrême-Orient 11/1: 1–22. IC, see Cœdès, George. Indraji, Bhagvanlal. 1881. ‘An Inscription of Gaya dated in the Year 1813 of Buddha’s Nirvana, with Two Others of the Same Period’, Indian Antiquary 10: 341–347. Jacob, Judith. 1993. ‘Pre-Angkor Cambodia: Evidence from the Inscriptions in Khmer Concerning the Common People and their Environment’, in D.A. Smyth (ed.), Cambodian

06_CreSouthV1_8P_11Feb22.indd 219

219

Linguistics, Literature and History, pp. 299–318. London: School of Oriental and African Studies, University of London. Jash, Pranabananda. 1974. History of Śaivism. Calcutta: Roy and Chaudhuri. Kafle, Nirajan. 2013. ‘The Liṅgodbhava Myth in Early Śaiva Sources’, in Nina Mirnig, PéterDániel Szántó and Michael Williams (eds.), Puṣpikā: Tracing Ancient India Through Texts and Traditions. Contributions to Current Research in Indology. Vol. 1. Proceedings of the International Indology Graduate Research Symposium, pp. 241–262. Oxford: Oxbow. . 2020. Niśvāsamukhatattvasaṃhitā: A Preface to the Earliest Surviving Śaiva Tantra (on Non-Tantric Śaivism at the Dawn of the Mantramārga); Critical Edition, with Introduction & Annotated Translation and an Appendix Containing Śivadharmasaṅgraha 5–9. Pondicherry: Institut français de Pondichéry/École française d’Extrême-Orient/Asien-Afrika Institut, Universität Hamburg. Kapoor, Shivani. 2019. ‘Iconography of Pashupata Ascetics in Cambodia’, Orientations 50/1: 122–131. Lorenzen, David. 1991. The Kāpālikas and Kālāmukhas: Two Lost Śaivite Sects. 2nd ed. Delhi: Motilal Banarsidass. [1st ed. 1972] . 2000. ‘A Parody of the Kāpālikas in the Mattavilāsa’, in David Gordon White (ed.), Tantra in Practice, pp. 81–97. Princeton/Oxford: Princeton University Press. Majumdar, R.C. 1927. Ancient Indian Colonies in the Far East (Vol. I–Champa). Lahore: The Punjab Sanskrit Book Depot. Maxwell, Thomas. 2008. ‘The Stele Inscription of Preah Khan, Angkor, Text with Translation and Commentary’, UDAYA: Journal of Khmer Studies 8: 1–85. Moule, A.C. and Paul Pelliot (trans. and annotated). 1938. The Description of the World by Marco Polo, Vol. 1. London: Routledge. Paṇḍit Durgāprasād and Kāśīnāth Pāṇḍurang Parab. 1925. Samayamātṛkā of Kṣemendra (Sanskrit verses in Devanagari script). Bombay: Nirnaya Sagar Press.

11/2/22 3:37 PM

220

Swati Chemburkar

Pantulu, Ramavya J. 1930. ‘Malkapuram Stone-Pillar Inscription of Rudradeva (Rudramba)’, Journal of the Andhra Historical Research Society 4/1–2: 148–162. Parabrahma Sastry, P.V. 1974. Inscriptions of Andhra Pradesh, Karimnagar District. Hyderabad: Government of Andhra Pradesh. Pathak, V.S. 1960. History of Śaiva Cults in Northern India from Inscriptions (700 A.D. to 1200 A.D.). Varanasi: Ram Naresh Sharma. Pou, Saveros. 1997. ‘Music and Dance in Ancient Cambodia as Evidenced by Old Khmer Epigraphy’, East and West 47/1–4: 229–248. Ramesh, K.V. and S.P. Tewari. 1990. A Copper-Plate Hoard of the Gupta Period from Bagh, Madhya Pradesh. New Delhi: Archaeological Survey of India. Rice, Benjamin L. 1902. Epigraphia Carnatica, Vol. 7: Inscriptions in the Shimoga District (Part I). Mysore: Archaeological Survey of India. Roveda, Vittorio. 2005. Images of the Gods: Khmer Mythology in Cambodia, Laos and Thailand. Bangkok: River Books. Samuel, Geoffrey. 2006. ‘The Siddha as a Cultural Category’, in Robert N. Linrothe (ed.), Holy Madness: Portraits of Tantric Siddhas, pp. 22–33. New York/Chicago: Rubin Museum of Art/Serindia. Sanderson, Alexis G.J.S. 1988. ‘Śaivism and the Tantric Traditions’, in Stewart Sutherland, Leslie Houlden, Peter Clarke and Friedhelm Hardy (eds.), The World’s Religions, pp. 660– 704. London: Routledge and Kegan Paul. . 2003–2004. ‘The Śaiva Religion Among the Khmers, Part I’, Bulletin de l’École française d’Extrême-Orient 90–91: 352–464. . 2012–2013. ‘The Śaiva Literature’, Journal of Indological Studies 24–25: 1–113. Sastri, Nilakantha K.A. 1939. Foreign Notices of South India: From Megasthenes to Ma Huan. Madras: University of Madras. Sastri, R. Anantakrishna. 1940. Pasupata Sutras with Pancarthabhashya of Kaundinya. Trivandrum: The Oriental Manuscript Library of the University of Travancore.

06_CreSouthV1_8P_11Feb22.indd 220

Schopen, Gregory. 2005. Figments and Fragments of Mahāyāna Buddhism in India. Honolulu: University of Hawai’i Press. Sharrock, Peter D. 2007. ‘The Mystery of the Bayon Face Towers’, in Joyce Clark (ed.), Bayon: New Perspectives, pp. 230–282. Bangkok: River Books. Shastri, Ajay Mitra. 1995. Inscriptions of the Śarabhapurīyas, Pāṇḍuvaṃśins and Somavaṃśins. 2 vols. Delhi. South Indian Inscriptions (Texts), Vol. 5. 1925. Miscellaneous Inscriptions from the Tamil, Malayalam, Telugu and Kannada Countries, edited by Rao B.H. Krishna Sastri. Madras: The Superintendent Government Press. South Indian Inscriptions (Texts), Vol. 10. 1948. Telugu Inscriptions from the Madras Presidency, edited by J. Ramayya Pantulu. Delhi: The Manager of Publications. Stern, Philippe. 1927. Le Bayon d’Angkor et l’évolution de l’art khmer: étude et discussion de la chronologie des monuments khmers. Paris: Geuthner. Stutterheim, Willem F. 1956 . ‘A Thousand Years Old Profession in the Princely Courts on Java’, in id., Studies in Indonesian Archaeology, pp. 91–103. The Hague: Martinus Nijhoff. Tawney, C.H. 1901. The Prabandhachintāmaṇi or Wishing-Stone of Narratives. Translated from the Original Sanskrit. Calcutta: The Asiatic Society. Törzsök, Judith. 2016. ‘Theatre, Acting and the Image of the Actor in Abhinavagupta’s Tantric Sources’, in Eli Franco and Isabelle Ratié (eds.), Around Abhinavagupta: Aspects of the Intellectual History of Kashmir from Ninth to Eleventh Century, pp. 451–495. Berlin: LIT Verlag. Trivedi, H.V. 1978. ‘Kalanjar Stone Inscription of the Time of Kirtivarman’, Corpus Inscriptionum Indicarum VII/3. Venkatacharya, T. and A.K. Warder. 1967. The Ubhayābhisārika or Both Go to Meet: A Satirical Monologue or Bhāna by Vararuchi. Madras: V. Sambamurthy. Venkataramanayya, N. 1974. Inscriptions of Andhra Pradesh: Warangal District. Hyderabad: Government of Andhra Pradesh.

11/2/22 3:37 PM

Performance-based Śaiva Worship and its Development in the Temple Cults of Angkor Vickery, Michael. 1998. Society, Economics and Politics in pre-Angkor Cambodia: The 7th–8th Centuries. Tokyo: Centre for East Asian Cultural Studies for UNESCO. Yazdani, G. (ed.). 1960. The Early History of the Deccan, Vol. VII-XI. London/Bombay/New York: Oxford University Press.

06_CreSouthV1_8P_11Feb22.indd 221

221

Zakharov, Anton O. 2019. ‘The Earliest Dated Cambodian Inscription K. 557/600 from Angkor Borei, Cambodia: An English Translation and Commentary’, BOCTOK (Oriens) 1: 66–80.

11/2/22 3:37 PM

Chapter 7

Libraries or Fire Shrines? Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples from the Prism of Early Śaivism1 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

A

introduction

unique feature of Cambodian temple complexes dating from the early to the late Angkorian periods was a small annex building (of brick and sometimes stone or laterite) with ventilation holes, built in the southeastern quarter of the main temple, which scholars have tentatively called either a ‘library’ or a ‘fire shrine’. These enigmatic structures and their function call for further investigation. In what follows we tentatively seek to establish the original function of those structures, and its possible evolution from the early Angkorian through the Angkorian period, by analysing their peculiar architectural features against the background of the Sanskrit inscriptional record from the Khmer domains and Sanskrit transmitted texts from South Asia describing the observances and ritual practices of (Pāśupata or Pāśupata-influenced) early Śaivism. We do not deny the possibility that (some of) these annex buildings might have functioned as manuscript repositories, but rather argue that they served predominantly (or, perhaps, even simultaneously) as sacred spaces for specific Śaiva rituals, including ash-related rituals of the Pāśupata sect, as well as initiation- and homa-rituals. Having first analysed the architectural features of the annex buildings and similar struc1. We wish to express our sincere thanks to Peter Sharrock and Hiram Woodward for reading an early draft of this chapter and making useful suggestions. Any mistakes are ours.

07_CreSouthV1_8P_11Feb22.indd 222

tures, for which we propose a threefold taxonomy, we review the interpretations thus far advanced by scholars on the basis of both archaeological data and Sanskrit inscriptions from the ancient Khmer domains, and start to elaborate our own in terms of a possible dual function for the buildings. Then we look for traces of Pāśupata practices and their underlying ideology in the Khmer temple architecture, in particular in the annex buildings under scrutiny, by matching the archaeological data with textual data, including Pāśupata texts and early Śaiva sources like the Śivadharmottara. We conclude by elaborating on the importance of sacred fire and the bestowal of the Śaiva gnosis by the ācārya in the Pāśupata initiation ritual, which may have provided one of the raisons d’être for the annex buildings, and by giving a brief survey of the archaeological finds recovered from some of those buildings, which would seem to support their use for homa rituals.

architectural features and functions of the annex buildings The small annex sanctuaries built within Khmer temple complexes, which have been variously referred to in secondary literature as ‘library buildings’ or ‘fire shrines’, are usually placed in the front of the main entrance on the right side of the central shrine(s) of the monument (i.e., at the visitor’s left), and when there are two such buildings, they are located on the right and left side of the central shrine(s). See Fig. 7.1 for the location and typical plan of this building at Preah Ko, and Appendix

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

223

Fig. 7.1: Annex building at 9th-century Preah Ko temple placed in the southeastern corner of the second enclosure of the temple. (Drawing: Olivier Cunin)

7.2 (‘Brief Survey of the Annex Buildings in Khmer Temples’) at the end of this chapter.2 While proposing a taxonomical categorization of these annex buildings into three types—viz. A (mainly or distinctively early Angkorian, yet possibly pre-Angkorian in origin), B (mainly Angkorian), and C (late Angkorian, and not so widespread)—here we only examine some key examples of ‘Type A’ and ‘Type B’ buildings with or without ascetic imagery on their exterior, variation in the design of the ventilation hole, and modifications in the architecture (see Appendix 7.1, ‘Photographic Database of Annex Buildings in Khmer Temples’, and Appendix 7.2). For the sake of developing our argument, we also briefly discuss the buildings called ‘lodging stations’ (gîte d’étape) or ‘resthouse temples’ (chapelles de gîte d’étape). The ‘Type A’ annex buildings that are seen in Khmer temples have several unusual architectur2. To avoid any confusion, it is uniquely to these structures that we refer to as ‘annex buildings’ in this chapter. For an in-depth discussion on the architecture of Bayon style annex buildings, see Cunin 2004; for Banteay Kdei temple specifically, see Arahi 2000.

07_CreSouthV1_8P_11Feb22.indd 223

al features. The building is located in association with the central sanctuary in the first enclosure. The structure has one western entrance, facing the central shrine. Generally, no false doors are seen on the walls of this structure. The walls are usually much thicker than the other buildings of the temple complex (up to 1.4 m). The building has square or diamond shaped holes in the walls as well as in the simple corbelled vaulted-roof on all sides,3 and a chimney-like interior (Fig. 7.2b). The only images carved on the ‘Type A’ building are a row of ascetics sculpted into the brickwork of the external walls and the superstructure. All the ascetics are seated in cross-legged position, supported by a yogapaṭṭa (meditation support strap to hold the posture for long period of time) and their hands are held in añjalīmudrā or prayer posture. Sometimes a stone piece (Fig. 7.3) is found with only ascetic imagery that would have been placed in front of the ventilation holes. Ascetic figures, usually referred to as ṛṣis in modern scholarly lit3. The only exceptions are 13th-century Banteay Kdei and tower 64 of Banteay Chhmar, where the roof is a crowned tower (Cunin 2004: 58).

11/2/22 3:44 PM

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

224

Fig. 7.2: ‘Type A’ building at 9th-century Preah Ko (a) with an interior-like chimney (b), displaying a row of ascetics at the top level (c) and possibly at the lower level (d), square shaped holes on the wall (e) and 1.4 m thick wall (f). (Photos: Shivani Kapoor)

Fig. 7.3: Ventilation piece in stone with ascetic imagery from Prasat Tram Neak (IK 578.03) at the east of Siem Reap. Possibly 10th century based on the yogapaṭṭa. Angkor National Museum, Siem Reap. CA no. 365, N. 1768. (Photo: Swati Chemburkar)

07_CreSouthV1_8P_11Feb22.indd 224

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples erature, are found in profusion on reliefs, cornice, and pediments of Khmer temples throughout the Angkorian period, and are usually accompanied by various other images of deities or merely regarded as a decorative motif. On these structures this simple band of ascetic figures stands out as a distinguishing mark (Fig. 7.2a). The development of the ascetic figure with a yogapaṭṭa seemingly coincides with the early phase of Śaivism in Cambodia, when the Pāśupata tradition still held a prominent place in the socioreligious ladscape. To our knowledge, the earliest depiction of an ascetic figure with a yogapaṭṭa and possibly holding a club is on N11 temple at Sambor Prei Kuk (Fig. 7.4). Though the club is traditionally associated with the iconography of Lakulīśa as a teacher, this identification has thus far not been proposed for the Khmer figures (Kapoor 2019). Although, as we stated above, the annex buildings are distinctively Angkorian, their construction possibly began during the pre-Angkorian period, as excavations at the site of Sambor Prei Kuk (ca. 7th–8th century) seem to suggest (see Shimoda and Shimamoto 2012: 35). Excavations conducted on a mound located west of shrine N10 have revealed a brick structure. The upper part of the structure is collapsed but its base and lower part are still in good condition. The building consists of a main chamber and an antechamber attached to its western side. Shimoda, who is working on the excavations, thinks that the construction of this building is relatively early even though it was built after the construction of the main sanctuary, and underwent modifications in the course of time (ibid.: 35; Fig. 7.18). He argues: This building introduces an asymmetrical dimension in the southeast quadrant of the temple layout and it faces to west. Such characteristics are common to the so-called ‘library’ structures [annex buildings] in Khmer temples. These library buildings usually also have openings for ventilation, especially on their left sides, a characteristic shared with initial stage of this building with its open-air pillared structure. (ibid.)

This would be the first known annex building in Khmer temple architecture. More excavations are

07_CreSouthV1_8P_11Feb22.indd 225

225

needed to understand the exact number and spread of such buildings during pre-Angkorian times.4 Annex buildings continued to be built during the later period of Khmer temple architecture with some morphological transformations (from Type A to Type B) sometime during the second half of the 10th century, but always in association with the central sanctuary.5 Examples are seen at 10th-century Koh Ker, Banteay Srei, and 11th-century Preah Vihear (Fig. 7.5). The square/diamond shaped holes were replaced by a fenêtre gisante, a rectangular kind of balustrade window at the roof level of the annex buildings6 along with the windows on its walls. A small antechamber and sometimes a false door were added to the building. In spite of the architectural modifications, the iconographic association with ascetic imagery continued. Instead of the row of seated figures, the ascetics are now depicted on the lintel of the main entrance door and on its column bases, as seen at Preah Vihear temple (Fig. 7.6), or above the false windows, as seen at the Bayon (Fig. 7.7). Both Type A and B buildings establish the association with ascetics and denote some functional aspects possibly linked to the use of fire, i.e. a ventilation system made by holes or a particular type of balustrade windows, and ritual dependence on the central sanctuary within the temple complex. Type C building is larger in size and is not as widespread as the first two types. The building is usually built on a high platform accessed by one or two staircases from the main axis and is placed between the second and third enclosure of the temples. This type is seen at Angkor Wat, Beng Mealea, the Bayon, and Banteay Chhmar (see Appendix 7.1). None of these buildings has any ascetic imagery or any openings for ventilation. Balustrade windows have possibly replaced the small openings in these buildings.7 4. We would like to thank Kunthea Chhom for bringing Shimoda and Shimamoto’s paper to our notice. 5. See Dagens 2003: 193, ill. 161; Soutif 2009: 330. 6. The height of the balustrade window is less than its width and is indicative of its ventilator function. The regular windows found in the Khmer monuments are of different proportions. 7. For an in-depth discussion on the architecture of different annex buildings, see Cunin 2004.

11/2/22 3:44 PM

226

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.4: Ascetic wearing a yogapaṭṭa, and possibly holding a club; West Wall, N-11 Sambor Prei Kuk, 6th–9th century, brick and stucco (approx. 50 cm × 50 cm). (Photo: Shivani Kapoor)

07_CreSouthV1_8P_11Feb22.indd 226

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

227

Fig. 7.5: ‘Type A’ annex building at Prasat Pram, Koh Ker (a) and ‘Type B’ annex building at Banteay Srei (b) and Preah Vihear (c). (Photo: Swati Chemburkar)

Fig. 7.6: (a) Row of ascetics seen on the lintel as well as the base of the columns of ‘Type B’ annex building at 11th-century Preah Vihear temple; (b–d) details of the lintel. (Photo: Swati Chemburkar)

07_CreSouthV1_8P_11Feb22.indd 227

11/2/22 3:44 PM

228

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.7: Row of ascetics seen above the false window of the annex building BY50 and BY51 at the 13th-century Bayon temple. (Photo: Swati Chemburkar)

Besides Type A, B, and C buildings, one more specific kind of stone building (of laterite or sandstone) was also developed in the Khmer domains. This building was originally called a ‘palace’ (palais) or a ‘dwelling’ (habitation) by Lunet de Lajonquière (1902: xxviii–xxx), and a dharmaśālā (‘rest house’) by Foucher (1903: 180) and Finot (1925). Cœdès (1940) called it gîte d’étape (‘lodging station’)8 because of its association with the ancient Khmer road network, but also maison avec du feu/maison du feu (‘house with/of fire’). The same expression has been used by Jacques and Lafond (2004: 387).9 A comprehensive study by Hendrickson (2008a) refers to these buildings as ‘“resthouse” temples’. While admitting a possible religious function associated with fire for these buildings, Maxwell makes a clear distinction between dharmaśālās and our annex buildings based on their usage for either portable or permanent fire (Maxwell 2007; 2009: 150; Fig. 7.8). 8. See also Dagens 2003: 70. 9. The English version (2007: 263) has ‘the house of Fire’.

07_CreSouthV1_8P_11Feb22.indd 228

Though Jayavarman VII seems to have commissioned many of them, they were already being mentioned in the reign of Sūryavarman II early in the 11th century (Cunin 2004: 444). These structures are built at regular intervals along the main roads of the Khmer Empire, which presumably led them to being used as resthouses.10 The Sanskrit name for them in the associated inscriptions (such as at Preah Khan) is either ‘house of fire’ (hutabhujaḥ ālayaḥ, vahneḥ ālayaḥ, vahnigṛha) or ‘caravanserai’ (upakāryā) (Maxwell 2007: 32–33). On the basis of the archaeological context and interpretations by previous scholars, Hendrickson (2008a: 73) assumes a dual function for these structures as both fire shrines and (out)posts, postal stations, or inns that played a role within the Angkorian transport system. The architecture of these building is significantly different from the other types and consists of a long, wide vaulted hall with an 10. For instance, those regular-spaced buildings dotted the royal road that linked Angkor to the Phimai sanctuary in modern Thailand. On the Khmer road system, see Hendrickson 2008b.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

229

Fig. 7.8: ‘Resthouse’ at 13th-century Preah Khan temple of Jayavarman VII. (Photo: Swati Chemburkar)

entrance at the front and row of windows along one wall, sometimes with small ventilation grills above, followed by a window chamber at the end with its own entrance in the back wall and surmounted by a temple tower. Since these structures apparently housed both statues (in the sanctum under the crowned tower) and fire (in the long chamber ventilated by large windows), they were used as staging posts or repositories for sacred fires that were carried in procession (Jacques and Lafond 2007: 263; Maxwell and Poncar 2006: 132–137), such as the one depicted on the reliefs of Angkor Wat and the Bayon (Jacques and Lafond 2007: 26; Fig. 7.9), or those mentioned in the Preah Khan inscription (st. 122–126, Maxwell 2007: 84–85). To the best of our knowledge, none of these buildings were inscribed.

07_CreSouthV1_8P_11Feb22.indd 229

identification of annex buildings (type a and b) as ‘library buildings’ The identification of Type A annex building was first proposed at the beginning of the 20th century by Lunet de Lajonquière (1902: xxx–xxxi), who called these structures ‘libraries’ (bibliothèques) with small openings serving as lighting shafts for reading. This identification was not based on any inscriptional evidence. On the basis of two 10th-century inscriptions, Cœdès too identified these structures as ‘libraries’.11 The first inscription, K. 958 (947 CE), comes from Prasat Kôk Čak, just off the road from Phnom Penh to Siem Reap, about 6 km from Siem Reap, 11. IC, VII: 141; Cœdès 1911: 405.

11/2/22 3:44 PM

230

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.9: Depiction of the sacred fire being carried out along the road or between temples. Inner gallery relief on the Eastern side between BY23 and BY24, 13th-century Bayon temple. (Photo: Swati Chemburkar)

and mentions the establishment of a pustakāśrama, literally a ‘resting place for books’, or a ‘building for books’.12 The inscription says that this type of building was erected by Hiraṇyaruci, but does not indicate its exact location: (31) hiraṇyarucinā tena pure rudramahā[laye] / (32) sthāpitaṃ vidhinā liṅgaṃ śrībhadreśvarasaṃjñakam // XVI (33) sa pinākipade śreṣṭhapure rudramahālaye / 12. The term pustakāśrama does not figure among the Sanskrit words for ‘library’ found in the Monier-Williams Sanskrit-English Dictionary (such as e.g. granthakuṭī, pustakāgāra, vidyākośagṛha, etc.). A library is also known as a granthālaya or a pustakālaya: cf. Datta 1960: 3, and De Simini 2016: 143–145.

07_CreSouthV1_8P_11Feb22.indd 230

(34) rudrāśramatribhuvanasthāneśānapurādiṣu // XVII (35) liṅgāny arccāś śivādīnāṃ nyadhāl liṅgapurādiṣu / (36) śrāṇāśraya ca śraminām āśramaṃ pustakāśramam // XVIII

Goodall’s (2017: 138) translation: This Hiraṇyaruci erected, following the [appropriate] rites, in Rudramahālaya a liṅga named Śrī-Bhadreśvara. In Pinākipada, in Śreṣṭhapura, in Rudramahālaya, in Rudrāśrama, Tribhuvanasthāna, Īśānapura and other towns, he set up liṅgas and cult-statues of Śiva and other deities, a place [for distribution] of cooked food, an āśrama for [the repose of] the weary and a library (pustakāśramam).

The second inscription, K. 355 from Prasat Khna (Preah Vihear province, second half of the 10th

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

231

Fig. 7.10: Annex building at Prasat Khna displaying ‘Type A’ architectural features. (Photo: Shivani Kapoor)

century), is important as it is inscribed on the doorjamb of the annex building and identifies that very building as a pustakāśrama (Cœdès 1911: 405; Fig. 7.10): hiraṇyarucinā te(20)[na] **** ~ – ~ – ****** j[e]na kṛto yaṃ pustakāśramaḥ // XXII adhyāpakādhyetṛhitaiḥ (21) **** ~ – ~ – ***** vānāṃ śāstrāṇāṃ śastabuddhinā // XXIII

Goodall’s (2017: 142) translation: That Hiraṇyaruci, of trained intellect (śastabuddhinā), who knew … (…jñena), created this library (kṛto ’yaṃ pustakāśramaḥ) [[filled]] with [[books]] beneficial for teachers and students [and] belonging to … disciplines (…vānāṃ śāstrāṇāṃ).

Cœdès (1911: 406) suggested that the word āśrama in the inscription could be translated as a ‘building’ rather than a monastery or hermitage: In Cambodian epigraphy, the term âçrama seems to have a rather vague meaning and of-

07_CreSouthV1_8P_11Feb22.indd 231

ten designates something other than a monastery or a hermitage. In many cases, the word ‘edifice’ is the one that best translates it. Also, pustakâçramaḥ corresponds quite exactly to ‘library’. On the other hand, the presence of the determinative ayam leaves no doubt as to the location of the monument where the manuscripts were deposited: the small kiosk bearing the inscription was, therefore, surely the library of the temple of Pràsàt Khnà.13

13. French original: ‘Dans l’épigraphie cambodgienne, le terme âçrama semble avoir un sens assez vague et designer souvent autre chose qu’un monastère ou un ermitage. Dans bien des cas, le mot « édifice » est celui qui le traduit le mieux. Aussi pustakâçramaḥ correspond-il assez exactement à « bibliothèque ».  D’autre part, la présence du déterminatif ayam ne laisse pas de doute sur l’emplacement du monument où les manuscrits étaient déposés: le petit édicule qui porte l’inscription était donc sûrement la bibliothèque du temple de Pràsàt Khnà.’ (Compare the word dravyāśrama in K. 366 [IC, V: 294] of Wat Phu, which Cœdès translates as ‘treasure building’).

11/2/22 3:44 PM

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

232

Following Cœdès, Goodall (2017: 144) has hypothesized that these buildings, which he calls ‘temple-libraries’, might have been places for inscribing, copying, and storing books/manuscripts. To account for the peculiar architecture of the small diamond/square shaped openings on the walls of the annex buildings, he says: Returning for a moment to the seemingly windowless building that Hiraṇyaruci’s inscription labels as a library, it might appear that it was not intended as a well-lit space for sitting down and poring over books, but rather as a place of storage, for the small diamond-shaped holes in its sides […] seem, at least to a modern viewer used to the overcast skies of Northern Europe, to be designed for ventilation rather than light. But Lunet de Lajonquière’s account of the features typical of such library buildings across the Khmer world suggests that those small openings were in fact intended for lighting the large vaulted spaces within, and that they are, in other comparable buildings, replaced by windows screened by balustrades (1902, xxx). […] So these may really have been intended as spaces for study, sufficiently lit, given the strong sunshine of the region, by little more than slits in their sides. One other misapprehension should perhaps be touched upon. It has been mentioned to me, but I can unfortunately not remember by whom, that some comparable and similarly positioned (opening to the West on either side of the Eastern approach to the main sanctuary) buildings at other Khmer sites may bear signs of having had fires lit in them, which might seem surprising if they were really places for the storage and study of books. I do not know if this is true, but if any such buildings did regularly have fires lit in them (the SouthEast being after all the direction of Agni and the place of the kitchen in South Indian temples, such as in the great temple at Tanjore), this does not necessarily preclude their having been used for the storage of books, for hanging palmleaf manuscripts above fire places, where smoke and dry warmth would minimise the attacks of insects and fungus, was evidently commonly practised in some parts of South India.

We will return to the dual function of the buildings, i.e. as both ‘libraries’ and fire shrines, as well as to

07_CreSouthV1_8P_11Feb22.indd 232

the evaluation of the idea that literary activities might have been carried out there, later in this chapter. Drawing on Goodall’s analysis of the annex buildings quoted above, De Simini has matched the epigraphic and archaeological evidence of the Khmer structures with the mention of a small building called ‘abode of knowledge’ (vidyālaya)14 in verse 117 of Chapter 2 of the early Śaiva text Śivadharmottara (ca. 7th century). She suggests that the Śivadharmottara ‘alludes to the existence of a small building that is apparently annexed to the compound where the donation of the manuscript is to take place, and which may qualify as a small manuscript repository’ (De Simini 2016: 116). The measurements of this building—whether it applies to breadth, length, or both not being specified—are eight hands or approximately 3.6 metres (ibid.: 142). On this basis, she concludes that: The aforementioned example from Cambodia illustrates that historical evidence has proven from an early date that there were ‘storerooms’ of manuscripts that were attached to religious and educational institutions and financed through private donations. (ibid.: 143)

Further, De Simini (ibid.: 155) resorts to an early Sanskrit inscription (K. 359, prob. 6th century) from the Stung Treng province of Cambodia by the (Pāśupata?) Brahmin Somaśarman (st. IIIa),15 to link the ‘abode of knowledge’ to ‘the ritualistic donation of manuscripts to a temple for storage purposes as well as for their ritual recitation’, and suggests that the building ‘could indeed be a small library in which the donated manuscripts were 14. Compare the synonyms vidyāsiṃhāsana (‘lion-throne of knowledge’, 2.23, 114), vidyāvimāna (‘vehicle of knowledge’, 2.45), vidyāratnakaraṇḍaka, (‘jewel box of knowledge’, 2.112), vidyākośasamāśraya (‘storehouse of knowledge’, 2.109), vidyāvāsagṛha (‘house in which knowledge abides’, 2.111) and vidyākośagṛha (‘treasure-house of knowledge’, 2.113), which ‘while seemingly only indicating the container of a manuscript in this case, could indeed also be employed to designate a “larger” repository for more texts, which in fact we would call a library’ (De Simini 2016: 141). 15. Barth 1885: 28–31; cf. Majumdar 1953: 18–19. The name Somaśarman appears to be associated in the Indian subcontinent with both Lakulīśa, founder of the Pāśupata movement, and a Kāpālika lineage: see Törzsök 2020: 36–38.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples stored (and worshipped)’. The inscription, which might even predate the Śivadharmottara as we know it, seems to reflect the description of ritual activities in the vidyādānādhyāya mentioned in that text, namely the erection of a liṅga as well as ‘the donation of manuscripts to a temple and their preservation in a spot from which, as the intimidating character of the last line suggests, nobody is allowed to remove them’ (ibid.). Besides the Prasat Khna, Prasat Kôk Čak, and Stung Treng inscriptions, several epigraphic documents mention physical books, being copied and stored in a library, but do not specifically indicate any building.16 The 12th-century inscription K. 364 from Ban That (Champasak province, Laos) also refers to books kept in an āśrama (Goodall 2017: 148), but does not indicate any specific location or building: Last face (stanza numeration uncertain): niśśeṣaśāstrair likhitais sanāthā[n] (54) = - ~ - ~ ~ - ~ - ān / sa pustakān adhyayanācchidārthaṃ tatrāśrame ’nekavidhān acaiṣīt // Dans cet āśrama, pour que l’étude y fut poursuivie sans interruption, il réunit un grand nombre de manuscrits traitant de toutes les sciences.... (trans. Finot 1912: 28)

Stanza 19 is striking as it presents—perhaps, deliberately punning?—an association between this āśrama, the study of books, and the performance of sacrifices that cause a ‘fragrant smoke’ to arise above it: vidyāpavarggavihitāpacitiprabandhe / yasyāśrame ’navaratāhutidhūmagandhe / durggāgameṣu matibhedakṛtārthanītyā / vidyārthināṃ vivadatāṃ dhvanir utsasarppa // Above his āśrama, in which there was a constant stream of [donative] acts of honour performed [by graduating students] at the moment of concluding their studies, which was fragrant with the smoke of an uninterrupted sequence of sacrifices, there rose [constantly] the sound of students debating over difficult [passages of] 16. Such as, e.g., Banteay Srei inscription K. 842 of 890 śaka (IC, I: 147).

07_CreSouthV1_8P_11Feb22.indd 233

233

transmitted texts (durgāgameṣu) in a fashion that was successful in accordance with the various schools (matibhedakṛtārthanītyā). (trans. Goodall 2017: 149)

Although we venture to speculate that the compound vidyāpavargga could punningly allude to the ritual of vidyādāna,17 the evidence is too tenuous to associate the inscription, mentioning a particular āśrama linked to the storage and study of texts and manuscripts, with one of the annex buildings discussed here. The 11th-century inscription K. 278 (stanza 23) mentions the donation of a beautiful book by Śivavindu to be kept in the main sanctuary (vimāna) of a Bhadreśa temple (Barth 1885: 116; cf. Goodall 2017: 148). Even though we have several inscriptions associated in some way with the concept of a library or storehouse for books, only one is inscribed on the annex building itself, namely K. 355.

alternative views on the annex buildings In this section we will first discuss the scholarly views advancing an alternative interpretation of the annex buildings as fire shrines, then put forward our own hypothesis, which will be elaborated in detail in the final sections of this chapter. Claude Jacques (Jacques and Lafond 2007: 20) argued that the term ‘bibliothèque’, which dates back to the early 20th century, did not necessarily indicate the ancient function of these buildings, which were rather to be identified as homa structures: In the southeast quadrant of the enclosure there is usually, in temples of the Angkor period, what is mistakenly called a ‘library’ but 17. The compound, translated by Goodall as ‘at the moment of concluding their studies’, could hide a double meaning: cf. Monier-Williams s.v. apavarga, ‘gift, donation’, which may be interpreted as a synonym of vidyā (therefore, vidyāpavarga = vidyādāna?). The alternative translation could thus be: ‘Above his āśrama, in which there was a constant stream of acts of honour performed for the sake of the “gift of knowledge”’.

11/2/22 3:44 PM

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

234

Dominique Soutif in his PhD dissertation devotes a section to fire rituals and fire shrines in ancient Cambodia, and returns to these issues in a 2017 paper (Soutif 2009, 2017). He discusses occurrences of such terms as vahnyāgāra and agnigṛha in the Khmer epigraphical corpus (Soutif 2009: 326– 327),20 noting that the latter may indicate the house of fire associated with a temple, such as in the case of Śaṅkarapāda (K. 485, 944 śaka, st. LX; Jacques 1968: 620). He also mentions inscriptions referring Jacques’ interpretation is not based on any inscrip- to the installation of sacred fires in temple infrational evidence but relies on the architecture of structures, for instance at Śrī Bhadreśvarāśrama the building. (K. 258A, 1001 śaka). Noting that no permanent In his PhD dissertation on the Bayon type of crucibles have ever been recovered from Khmer monuments, Olivier Cunin identified those build- temples, he argues that the Khmer expression kralā ings as fire shrines. However, inscription K. 355 homa attested on inscriptions corresponds to the makes a stronger case for the structure at Prasat Sanskrit agnigṛha and vahnyāgāra in the context of Khna being a library building.18 On the basis of the palace, or to a hall in a temple (ibid.: 328). On the architectural features of the building, Cunin the basis of inscription K. 1065A, he argues that the elaborates on this problem as follows: expression kralā vraḥ vleṅ (‘salle du Feu sacré’ or ‘aire du Feu sacré’) denoted an annex building atD’autres indices semblent cependant contretached to a sanctuary rather than a simple sacrificial dire cette hypothèse et laissent douter que area, and that it corresponded to the expressions l’ensemble de ce type d’édifice était bien le lieu où étaient conservés les manuscrits des monuagni-niketana/-āgāra and homasthāna/yagasthāna ments. En effet, il n’est pas rare de trouver les encountered in Śaiva ritual treatises from India traces d’un piédestal d’une idole accompagné (ibid.: 328). He further analyses the expression des logements d’un dais. Il semble donc que ces padaḥ sthāna jā ’āśrama vleṅ in K. 691, translated derniers étaient plus des lieux de culte que des by Cœdès (IC, IV: 151, fn. 4) as ‘l’āśrama où réside lieux de conservation de manuscrits. La préle Feu sacré’, as referring to a building installed for sence d’un système d’aération pour faciliter la a sacred fire in a temple i.e. the southern piédroit combustion d’un feu rituel renforce cette idée. of the annex building of the temple of Trapeang (Cunin 2004: 59–60)19 Ropou (Lunet De Lajonquière 1911: 132). On the basis of the combined inscriptional evidence of 18. And yet, inscription K. 661 (IC, I: 197), engraved on the K. 691 and K. 937, as well as Dagens’ observation East Gopura, suggests that fire rituals dedicated to Śiva were that the architectural features responded to the among the daily activities going on at the temple: see side needs of aeration rather than illumination (Dagens B, st. LXXIV: śivārcanāgnihotrāditapasyāsādhanāni yaḥ / mantratantrāṇi saṃśodhya vidhaye ’rañjayad dhiyā //, ‘Par 2003: 193), Soutif cautiously supports the view that l’effet de son activité intellectuelle, il [King Śrī Sūryavar- the libraries may be identified as fire shrines, yet man I] fit briller, en les éditant pour la célébration des rites, without unduly generalizing, as other such annex les mantra et les tantra, guides des pratiques religieuses des buildings could have been libraries indeed, like ascètes, à commencer par le culte de Śiva et les oblations which is a Fire temple. Just as was the custom in India, Fire had to be ritually re-kindled each morning before the ceremonial worship of the shrine’s main deity, but in Cambodia from the ninth century onwards the sacred fire had its own dedicated sanctuary. There must also have been kitchens close by to cook the god’s food, but as they would certainly have been temporarily shacks made of perishable materials, no trace of them has survived.

au Feu’ (trans. Bhattacharya 1961: 47); side D, st. CVIIIab: śivāgnigurupūjāsu nipunaḥ … ‘[Śrī Kavīndrapaṇḍita] habile à célébrer les cérémonies en l’honneur de Śiva, d’Agni, et de son maître…’ (trans. IC, I: 218); compare side A, st. LXIV. 19. English translation: ‘Other indications, however, seem to contradict this hypothesis and leave us doubting that the whole of this type of building was indeed the place where the manuscripts of the monuments were preserved.

07_CreSouthV1_8P_11Feb22.indd 234

Indeed, it is not uncommon to find traces of a pedestal of an idol accompanied by the sections of a platform. It seems, then, that these were more places of worship than places of preservation of manuscripts. The presence of an aeration system to facilitate the burning of a ritual fire strengthens this idea’. 20. Cf. also agniveśman, in K. 432 st. V (Soutif 2017: 10).

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples for example the one at Prasat Khna (Soutif 2009: 330–331). The inscriptions mentioned above might shed some light on the ‘Type A’ buildings. K. 937, found at Prasat Srangé, near Preah Ko, Roulos, is the earliest inscription (883 CE) associated to an annex building, and is inscribed on the left door-jamb of a ‘Type A’ building. The inscription introduces the fire priest of King Indravarman I, ācārya Nandikācārya, a man of great asceticism (IC, IV: 46): rājā rājadhirājo yaś Śrīndravarmmeti viśrutaḥ / ācāryyo nandikācāryyaḥ taddhotā sumahātapāḥ // I śarakhāṣṭāṅkite śāke devāgnir ṇnandikeśvaraḥ / sthāpito vidhinā tena hutvāgniñ ca sadārccayet // II [pa]ñcatanmātram asyedaṃ pañcakāman dadāti yaiḥ / patreṇa kampitaṃ tebhyaḥ suvarṇnavyajanāni ca // III The king of kings, Śri Indravarman had a priest called Nandikācārya who practiced great austerities. In the śaka year 805 (883 CE), he (Nandika) established, according to precept, a particular sacred fire, devāgni [named] Nandikeśvara [as his personal fire]. May he always worship [the god] after sacrificing to the fire! Rich rewards such as fans made of gold will be granted to those who will fan the sacred fire, [to keep it alive] even with a simple palm leaf. (trans. Maxwell 2009: 162–163)

The interesting element in the inscription is that the priest is establishing a particular sacred fire, remaining personally associated with the fire and keeping it alive, possibly with the help of his disciples. Besides Maxwell, Soutif (2009: 323) too notes the personal character of the cult reserved to the fire, namely the relationship between the founder and the sacred fire, and also hypothesizes that Nandikeśvara is to be regarded as a divinity venerated through an ancillary cult of the fire, probably related to Śiva (ibid.). Some 300 years later an association between fire, an ācārya, and a building is found in an inscription of 1216 CE (N1/17, written by Madhurendra) en-

07_CreSouthV1_8P_11Feb22.indd 235

235

graved on the right-hand doorjamb of the structure BC45 at Banteay Chhmar, illustrated in Cunin’s archaeological reconstruction (Fig. 7.11). The structure is in the southeast and displays all the features of an annex building. Architecturally this structure is a ‘Type A’ building’ with ventilation holes. A row of ascetics appears on the superstructure next to the Rāvaṇānugrāhamūrti and above the lintel. In a similar manner to K. 937, the inscription states the name of the donor, names the sacred fire after his personal vision, and ends with a list of anticipated rewards: vraḥ vleṅ svargga iman divaś śrīmadhurendrasūrir agniṃ patantaṃ samaālam atra / śrīvīraśaktyāgamanena so ’ham adrākṣam aṣṭātriśaśāṅkarūpaiḥ // 1 dṛṣṭo mayā dīpitadiṅmukhyo ’gnir divaḥ patan śrījayavarmmarājye / atyadbhutaṃ yatsaphalāni manye jātañ ca netre ca kulañ ca me syuḥ // 2 The sacred Fire from heaven. I myself, Śrī-Madhurendrasūri, saw this Fire falling from heaven, here, at the time of the arrival of Śrī-Vīraśakti in 1138 [Śaka]. I have seen the Fire lighting up the directions of space [and] falling from heaven into the kingdom of Śrī-Jayavarman—a great wonder, which I believe will entail good results, [namely] high birth, two [good] eyes, and eminent family. May these be mine! (trans. Maxwell 2009: 155)

Noting the strong similarities between inscriptions K. 937 and N1/17, Maxwell (2009: 159) suggests that, while the name Agni translates vraḥ vleṅ, divaḥ patan is a Khmer rendering of the Sanskrit word svarg(g)a, and both of these were clearly intended to convey an alternative conception of fire-god, one that had special significance for him as the expression of what he had witnessed. The first among the unusual aspects of this inscription is, therefore, that the individual fire named in it was a particular manifestation of sacred fire that is not mentioned in any other Jayavarman VII inscription.

The nature of the fire is not easy to define here, but Maxwell thinks the essential reason for engraving this text on the building is to explain and define the

11/2/22 3:44 PM

236

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.11: Structure 45 situated in the southeastern part of the east complex of the temple. It has a ventilation system similar to ‘Type A’ annex buildings. (3D archaeological reconstruction: Olivier Cunin)

sacred fire associated with it. Another short inscription, N2, names the sacred fire (ibid.: 145–146). It is inscribed on the left side door pillar of the entrance to a now collapsed building BC15, placed in the southeastern quadrant of the first enclosure of the temple, which houses a ventilation system in its architecture (Fig. 7.12). This building is a ‘Type B’ building with some variations. These two inscriptions leave no doubt that structures 45 and 15 were fire shrines, where the sacred fire, to which offerings must have been regularly made to the accompaniment of mantras, was kept burning (ibid.: 162). Both the Prasat Srangé and Banteay Chhmar inscriptions are important as they are inscribed on the doorjamb of the annex building and identify that particular building upon which it is inscribed as the one where the sacred fire was established.21 Maxwell (ibid.: 33) has dis21. In his work on the Banteay Chhmar inscriptions, Maxwell (2009: 171) argues that ‘fire shrines differed from temples in their ritual function, in their architecture, in

07_CreSouthV1_8P_11Feb22.indd 236

cussed the combination of fire cult and image cult in one building in the light of his readings of Angkorian inscriptions.22 Like the gods that were represented by statues, each devāgni was also regarded as an individual deity, a particular manifestation of fire which had been instituted by, and which could even be named after, a person of rank in the religious establishment (ibid.: 162; cf. Soutif 2009: 323). Inscription K. 691 (1002 CE), engraved on the piers of south annex building at Prasat Trapeang Ropou (Siem Reap province), mentions a certain Loñ Dān installing a fire in this very building. He their orientation, and in being regarded as independent foundations. As such their establishment and that of the fire which they housed was recorded and dated in their own foundation texts written by the fire-priest himself.’ 22. K. 950 (IC, VI: 115–118), where Caitanyaśiva and his brother are said to have founded a Viṣṇu temple in the 10th century; K. 254 (ibid.: 183, 187–188), mentioning provisions for maintaining the fire cult for three statues representing Śiva, Viṣṇu, and Devī (cf. Bhattacharya 1961: 147).

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

237

Fig. 7.12: The structures BC15 and BC45 are variants of ‘Type B’ and ‘Type A’ buildings and are located in the southeastern corner of the first enclosure and the Eastern complex of the temple. (3D archaeological reconstruction: Olivier Cunin)

presented eight servants to tend the rice fields belonging to this fire shrine, which delivered a fixed quantity of husked rice daily as sacrifice to the sacred fire. Today, we see the mound with some brickwork in the southeastern corner in front of the central sanctuary, but the old EFEO photographs (CAM 10148, 10153) clearly show the construction of an independent building much like an annex building or fire shrine of BC45 at Banteay Chhmar.23 K. 258, the stele of Samrong, mentions the establishment of a sacred fire in the tapovana (grove for the ascetic practices) in 1079 at the āśrama of the Śiva temple of Bhadreśvara by a fire priest called Yogīśvarapaṇdita.24 Here, again, the 23. K. 691 (IC, IV: 151). For architecture and archaeology of the temple site, see Pottier 2000: 95; for the old photos of the site, see Marchal 1931: 614 (pl. CXVB). 24. K. 258 (IC, IV: 175–192). Note that, as we have noted above, the word āśrama in this context could be interpreted

07_CreSouthV1_8P_11Feb22.indd 237

relationship between a sacred fire, an (ascetic?) priest and a building or religious foundation seems clear, even though its exact location is not known.

libraries or fire shrines? So far, we have one inscription engraved on an annex building, calling it a pustakāśrama (Prasat Khna, K. 958), and three inscriptions (one from Prasat Srangé, two from Banteay Chhmar)—or four if we include Trapeang Ropou—mentioning the same building as a place where the sacred fire was established. Does Prasat Khna then represent the exception rather than the rule, as suggested by Cunin (2004: 59)? Thus, two main opinions have been advanced by scholars to understand these buildings: one either as ‘religious foundation’ or as ‘building’.

11/2/22 3:44 PM

238

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

proposes they were used for holding/storing manuscripts, and the other argues that they were fire shrines or homa structures. On the one hand, inscription K. 355 from Prasat Khna would support their identification as libraries. But the architecture of libraries in the late mediaeval Indian subcontinent as well as the textual descriptions of earlier libraries do not show any similarities with the annex buildings of the Khmer temples, apart from, perhaps, the relatively small dimensions of such buildings.25 On the other hand, inscriptions on the annex buildings mentioning the establishment of a sacred fire open up the possibility of the interpretation of this kind of building as a fire shrine. Furthermore, the architectural features of the buildings—i.e. the venting system and thick walls—are perfectly adapted to the use for fire-rituals, and the archaeological finds seem to support this (see below). These so-called ‘library’ buildings are the only buildings in Khmer temple complexes suitable for housing fire.26 On the basis of the inscription and the architecture of the building, Goodall (2017: 144; cf. above) attempts to combine the two interpretations and understands these buildings as libraries where manuscripts were copied, stored, and possibly hanged on the ceiling so as to be fumigated by fire. We find this interpretation, attributing a ‘dual-function’ to the buildings, to be convincing. However, differently from Goodall, we also 25. The most prestigious and important mediaeval library of India is the Saraswati Mahal Library at Tanjavur, established under the Nayaka kings (1535–1675), which we assume would have been modelled on earlier buildings. For a synthesis of secondary sources discussing epigraphic evidence (mainly from the 11th–13th century) on libraries in the Indian subcontinent, see De Simini 2016: 143–145; cf. Datta 1960: 42, 182; Sankaranarayanan 1993: 28ff. 26. While a fuller comparative discussion must be left for another occasion, we should like to point out here that a similar typology of shrines of relatively small size and with uncharacteristic ventilation holes has survived in Central Java in the period from ca. 8th to early 9th century, namely Candi Semar in Dieng and Candi Pawon in the vicinity of, and in a straight line with, the Borobudur. Interestingly, pawon in Javanese means ‘kitchen’ (i.e., place of the ashes, awu), and local people also name this temple as Bajranalan (from vajrānala, where ānala = fire); both names suggest the possibility that this temple might have functioned as a fire-shrine for homa rituals (see Lokesh Chandra 1980: 317).

07_CreSouthV1_8P_11Feb22.indd 238

contend that the interiors of these building do not have sufficient light for inscribing or reading manuscripts. Given the thickness of the walls and nature of the small diamond-shaped holes, these structures were not intended to be well-lit places and any visitor would easily note that almost no light filters through them, despite the strong tropical sunshine. Since the ancient Khmer manuscripts are incised, the copyist had to clearly see what he was writing and good light was very important; thus, it is likely that these activities were carried out in open or semi-open spaces, such as pavilions, or larger buildings.27 We agree with De Simini’s (2016: 142) observation, in connection with the small building mentioned in Śivadharmottara 2.117 as a vidyālaya, that, although the text is not explicit about its precise function, its dimensions would be sufficient to qualify it as a library ‘that could serve as a storage place for manuscripts rather than a place suitable for studying them’.28 It is not easy to integrate such widely differing purposes, but the combined epigraphic, textual, and architectural evidence, the doctrinal and ritual background, as well as practical aspects connected with manuscript conservation, may all have contributed to shape and define the function of these buildings over time. It is not impossible that they were originally intended for fire rituals and only at a subsequent stage they also became places for the storage of manuscripts—the fumigation of manuscripts being added as a welcome ‘side-effect’, which would have helped the preservation of the leaves by eliminating insects and preventing the formation of fungi.29 Going back to inscription 27. Olivier de Bernon, email dated 24 April 2018. 28. Cf. De Simini 2016: 142, fn. 348: ‘From the modern accounts on traditional libraries we learn, for instance, that Jain libraries were limited to small, dark—and sometimes even subterranean—rooms […]. The archaeological remains identified as temple libraries in Cambodia have limited dimensions, which suggest that these places did not allow the use of manuscripts in situ for personal study.’ 29. The South Indian practice of hanging palm-leaf manuscripts above fire places or in kitchens mentioned by Goodall (2017: 144; cf. Dr Perumal’s blog, https://drperumal.wordpress.com/2013/05/10/) was (and is still) known in Indonesia, too: see van der Meij 2017: 47: ‘in Lombok, lontar manuscripts may be stored in the roof of the kitchen so that they are exposed to kitchen smoke to prevent insect

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

239

of all the main activities connected to the conservation of manuscripts. Having discussed the observations by previous scholars with respect to the possible function of the annex buildings, in what follows we would like to draw attention to some relevant passages of Sanskrit sources from the Indian subcontinent to advance a hypothesis concerning the origin and functions of these buildings in a Khmer religious and ritual context dominated by Pāśupata Śaivism in the pre-Angkorian period, and its evolution in the early and late Angkorian periods.

K. 958 from Prasat Kôk Čak, we wonder whether the expression pustakāśrama (occurring after śraminām āśramaṃ, meaning ‘an āśrama for [the repose of] the weary’) could metaphorically allude to a ‘resting-place’ for manuscripts, not implying a place for the accession and use of the manuscripts, i.e. a library, but a building where manuscripts were temporarily laid down for periodic ‘servicing’, so to speak, in an environment that might have been fumigated most of the time thanks to the presence of a sacred fire. But it is also possible that these buildings were meant for the permanent storage of books in a fumigated environment—the premodern equivalent of a low-humidity, pest-free controlled microclimate. On the top of that, the possible ritual function of such buildings should be considered, too. For instance, the Northeastern Indian Buddhist monasteries seem to have known small edifices where old and worn-out manuscripts were permanently stored. De Simini (2016: 17), summarizing the scholarly discussion on this issue, remarks that ‘the British Library scrolls [of Gilgit] represented a ritual burial for old, “dead” manuscripts, which would have formed a sort of “Buddhist genizah”.’30 These books would also be worshiped as cult objects, as described in such texts as the Śivadharma, which characterizes the abode of knowledge (vidyāyatana) ‘as a place where the manuscript(s) of the Śaiva knowledge will, “like Śiva” (śivavat, 2.122), be worshipped daily’ (ibid.: 145).31 To De Simini, it therefore appears that Śaivism, too, admitted a ritualization

The Pāśupatas were one of the earliest organized and widely distributed Śaiva ascetic orders of the Indic world in the early mediaeval period. Several inscriptions document the spread of this order across the Indian subcontinent and beyond to Southeast Asia, clearly indicating the missionary character of the movement (Bhattacharya 1955; 1961: 43–76). By the 7th century, Pāśupata masters had been able to secure a place for themselves with the Khmer royalty.32 Although one notes the concomitant rise to popularity of mainly ritualistic Mantramārga Śaivism from the 9th century onwards,33 its influence on the ritual practices of the

damage, but in consequence they become black and dirty and seem old while in fact they are not.’ Andrea Acri is aware of similar practices in Java and Bali. Agrawala (1984: 34) describes the practice of hanging fresh palm leaves (that is, before their engraving) for several days in a place where smoke is emitted to make them insect- and fungus-proof in Thailand. 30. De Simini (2016: 17–18, 149–150) presents interpretations of this phenomenon by Solomon, Schopen, and Strauch, according to which this building represented a combination of genizah and scriptorium, located within the precinct of the Buddhist monastery, would have stored old or unusable texts while still allowing them to remain accessible to the monks. 31. It is perhaps significant that the text declares the worship to be performed thrice daily, at the main sandhyā, just like for the observance of bhasmasnāna (see below).

32. One of the earliest inscriptions in Cambodia, K. 604 (IC, IV: 17–19) of 627 CE from the reign of Īśānavarman I (616–635), founder of the Īśānapura kingdom with Sambor Prei Kuk as its capital, commemorates the erection of a liṅga by a Brahmin named Vidyāviśesa. According to Cœdès, Vidyāviśesa was the first Pāśupata master recorded in the Khmer domains. 33. Bhattacharya 1961; Sanderson 2003–2004: 435. Goodall (2015: 28) suggests that the earliest known evidence of the introduction of the Mantramārga in Kambujadeśa belongs to the middle of the 8th century; however, he (ibid.: 25) also notes that, although Mantramārga initiation-names ending in -śiva only appear from the late 9th century and become common in the 10th, ‘this does not necessarily mean that we can conclude with certainty that the Mantramārga only reached the Khmers in the Angkorian period’, and that ‘the relatively sparse documentation for the late

07_CreSouthV1_8P_11Feb22.indd 239

matching the textual, epigraphic, and architectural datas: pāśupatas and rituals relating to fire and ash in south and southeast asia

11/2/22 3:44 PM

240

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Khmer court appears to have continued well into the 13th century.34 Our sources for the knowledge of Pāśupata doctrine (vidyā, jñāna) and observances (vrata, vidhi) are limited to a handful of Sanskrit texts, containing both first- and second-hand accounts.35 From these accounts we know that, in order to obtain union with the supreme deity Rudra-Śiva, the Pāśupata adept was supposed to perform his ascetic observances in five stages while living in the following five places: in a temple, among people, in an empty house, in the cremation ground, and in Rudra (Hara 1966: 3–5). In the first stage, the disciple retreats to, and dwells in, a Śaiva temple and carry out worship activities including bathing in ashes three times a day (bhasmanā triṣavaṇaṃ snāyīta, PS 1.2), sleeping in a bed of ashes (bhasmani śayīta, PS 1.3), wearing a garland taken from the image of the deity (nirmālya), etc. Some among pre-Angkorian period makes it dangerous to make any sort of statement about religion, politics, literary taste or whatever at that time’ (ibid.: 26); cf. Sanderson 2003–2004: 444, who points out that certain iconographical elements found in Mantramārgic sources from the subcontinent are already attested among the Khmers in the 7th century, ‘well before the Mantramārga reached their shores’. 34. See Finot 1928: 44 (our translation from the original French): ‘It should be noted that the servant of the temple must be a Pāśupata Brahmin; thus, one can note the importance, from the very beginning of the Khmer monarchy, of a sect that, more than six centuries later, Tcheou Ta-kouan still found flourishing in the capital of Cambodia.’ See also Chapter 6 in this volume, and Goodall 2015: 26, recording the presence of names of ascetics ending in -rāśi—a marker of Atimārga affiliation—in inscriptions throughout the 11th century, as well as traces of the doctrines of Pāśupata subgroups in 10th-century non-royal caves. 35. These, however, are mainly philosophy-oriented, like the Pāśupatasūtra (PS) with the commentary Pañcārthabhāṣỵa (PABh) by Kauṇḍinya (prob. 4th–6th century), and Haradaṭṭa’s Gaṇakārikā (GK) along with the commentary called Ratnaṭīkā (RṬ) attributed to Bhāsarvajña (prob. 8th and 10th centuries, respectively). Also relevant are the Pāśupata ritual manuals recently discovered by Diwakar Acharya, such as the Saṃskārakārikā (Acharya 2007), and other, non-Pāśupata sources that may have incorporated Pāśupata accounts, namely the early Siddhāntatantras Niśvāsatattvasaṃhitā and Mataṅgapārameśvara (Sanderson 2006: 156; Kafle 2020; Acri 2014), as well as the Atharvavedapariśiṣṭas (Bisschop and Griffiths 2003) and the (Ur-)Skandapurāṇa (Bakker 2014; Bisschop 2006).

07_CreSouthV1_8P_11Feb22.indd 240

these various practices of a Pāśupata ascetic’s life, enumerated in the PABh and other sources, can be seen depicted on reliefs of Khmer temples (Chemburkar and Kapoor 2018). According to the PABh, the pāśupatavrata commences with the ritual of the ash-bath (bhasmasnāna). The text (1.1, p. 4) describes how Rudra himself, having incarnated as Lakulīśa in the body of a dead brahmin at Kārohaṇa in present-day Gujarat, stated the injunctions and marks of the Pāśupata order, and began living on an altar of ashes, in a pure spot defined as āyatana, in order to establish a connection with his first disciple, Kuśika. In 1.2 (p. 8), both the sūtra and its commentary stress that the foremost vidhi for ascetic practice is that of bhasmasnāna, which must be performed three times a day, and that the ashes should be bright, pure, and should be obtained in a large quantity (see Hara 1966: 171–172). Supplementary bathing is required after any act like eating, spitting, etc. The injunction is to sleep on ash and nothing else (PABh ad 1.3, p. 9). The PABh (ad 1.2, p. 8) also mentions that the initiate is not allowed to make his own ashes or light a fire. The RṬ states that the ascetic must rub himself vigorously from the soles of his feet to the top of his head with the ash.36 The ash should be of the purest quality37 and the Pāśupata should take a large quantity, because they are necessary for the accomplishment of his duties. In case a large quantity cannot be obtained, then a practitioner may take less. The container for carrying ashes is a well-known implement of the sect (Hara 1966: 172). Mentions of ashes and ash-related practices are found in Khmer and Cam epigraphy. Inscription K. 733 (IC, I: 3–5), dated to 639 CE, of Phnom Preah Vihear talks about Vidyāpuṣpa, a Pāśupata ācārya who practised Pāśupata austerities. With a donation from a king, he established an āśrama for his coreligionists. This inscription speaks of Vidyāpuṣpa visiting a sacred place in a dream, where he saw the footprints of Śiva, a cow, and 36. RṬ ad GK 7, p. 17; cf. Śivadharmottara 11.46ab. 37. RṬ ad GK 7, line 11 remarks that the ashes must be white (śukla). The 12th-century Tewar Inscription (line 9) of Gayākarṇa praise an ascetic named Bhāva Brahma because he sleeps on pure ashes (Indian Antiquary vol. XVIII: 211; cf. Pathak 1980: 16).

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples ashes (bhasman). The inscription of Glai Lomov (Bergaigne 1893, no. 393A, st. V) of King Indravarman I (801 CE) from the neighbouring kingdom of Campā mentions white ashes (sitabhasman), yoga, muttering, huṃkāra,38 and the ultimate goal of union with Rudra, in concurrence with the Pāśupata regimen described in the PS (Bhattacharya 1955: 481). Another inscription, K. 853 at Phnom Bayan (IC, I: 258, st. VII) from Yaśovarman I’s reign (889–c.915), praises an ascetic named Amarbhāva,39 who is appointed as the head of the Indrāśrama, and is gifted a pot of gold filled with ashes.40 This is significant, as it casts light on one of the central practices of the sect. K. 380 (IC, VI: 264) from Phnom Preah Vihear in the reign of Sūryavarman I (r. 1002–1050) mentions bhasman. Though the inscription is too degraded to translate, Kamaleswar Bhattacharya (1961: 46, fn. 5) regarded this inscription too as being associated with the Pāśupatas. The epigraphic references listed above clearly indicate the importance of ash-related practices in the Śaiva/Pāśupata religious landscape of the ancient Khmer domains over a period of nearly 400 years. The inscriptional record also gives us glimpses of real-life practices by Pāśupata ascetics and hermitages relating to ash that are seldom if ever described in the Sanskrit transmitted literature from South Asia. For instance, some inscriptions differentiate between normal ash and pungent ash, possibly indicating their usage in different rituals— the latter being used for cleaning the jaṭā or the

38. Bergaigne 1893: 43–46; compare Bhattacharya 1955: 481. This may be a variant or ‘development’ of the original sound huḍukkāra or hudduṅkāram (also attested as duṃḍuṃ, huṃḍum, huḍuṃ in the sources), the imitation of the bull’s cry made by touching the palate with the tip of the tongue (Kafle 2020: 255–256, fn. 443). 39. On the basis of his study of Śaiva cult in northern India, Pathak (1980: 19, fn. 2) argues that ascetics of Pramāṇa and Ananta gotras of the Pāśupata sect bore names beginning with Bhāva-, such as Bhāvavālmīki, Bhāvatejas, Bhāvodyota. Estève (2009: 476–480) and Goodall (2015: 26–27) also advance this view. According to Bhandarkar (1913: 120), names ending in -rāśi are Pāśupata and/or Kālamukha. 40. K. 853, st. VII (IC, I: 257): ‘In his gratitude, this king duly honoured the ascetic, with precious objects: ashes in a pot of gold, a rosary etc.’

07_CreSouthV1_8P_11Feb22.indd 241

241

coil of hair worn by Śaiva ascetics as a chignon.41 While comparing the āśramas founded by Yaśovarman I, viz. a Buddhist Saugatāśramas at Prasat Ong Mong, a Vaiṣṇavāśrama at Prasat Komnap, and a Brāhmaṇāśrama and a Māheśvarāśrama at Prasat Prei, Cœdès (1932: 103) points to the list of objects required in each āśrama. While the Śaivas required ashes, the Vaiṣṇavas required razors and scissors. Inscriptions at Yaśovarman I’s āśramas for Śaivas and Pāśupatas specifically direct that each initiate should be given, among other things, One āḍhaka of ash, one āḍhaka of pungent ash to clean the hair bun. All these objects will be given every four months individually to the brahmans (dvija), the senior teachers and ascetics (tapasvin). (ibid.: 102)42

The term āḍhaka is a measure by weight of food grains; in the Khmer context, according to Claude Jacques, it would be equal to 5 kilograms.43 Therefore each ascetic, brahmin, and senior teacher in the āśrama required 10 kilograms of ash every four months. Given the number of ascetics/brahmins/ senior teachers who must have resided at each āśrama and performed the ash rituals three times a day, along with the requirement for ash for supplementary cleansing and sleeping, the amount of ash which each āśrama needed to generate and store would have been considerable. The requirement of ash mentioned in the Khmer inscriptions listed above in connection with populous Śaiva āśramas44 indicates the importance of, and adherence to, observances and rituals re41. K. 774 (IC, IV: 64); stele of Thnal Baray (K. 279D, st. IV–V, 9th century; Bergaigne 1893: 424, 430). 42. Cœdès notes that as Yaśovarman I founded several āśramas, these inscriptions may have been meant to introduce standardized regulations to earlier practices. 43. Āḍhaka conversion differs from time to time and region to region; Claude Jacques calculated it as ca. 5 Kgs in the ancient Khmer context (email dated 19 June 2017). 44. An indication of the number of people inhabiting these āśramas is suggested by the number of slaves donated to them: see C, XCVIII in Cœdès 1932: 104–105. He comments that in both the Vaiṣṇava and Buddhist āśramas there could be no less than fifty slaves; while the lines referring to the Brāhmaṇāśrama have been obliterated, a similar amount of slaves must have been attached to that āśrama.

11/2/22 3:44 PM

242

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

quiring this resource, and raises the question of where, in a monsoonal country, such a ritual could take place, and where the agnyāgāra or ‘house of fire’ for the production of ashes (and, perhaps, the performance of homa rituals) was situated within āśramas and, possibly, temple complexes as well.45 In the light of the above, we wonder whether the annex buildings discussed here could have provided suitable candidates for this role. In the lack of detailed and explicit data drawn from Khmer epigraphy and transmitted texts, we must turn to Sanskrit sources to find clues about the doctrinal and ritual background as well as the architectural context from which the annex buildings could have originated. The Atharvavedapariśiṣṭa (henceforth: AVP) 40 describes the duration, place, and methods involved for observing the Pāśupata observance, including bhasmasnāna at the end of the ‘offerings’ (i.e. dancing, singing, etc. while residing in a temple or āyatana: cf. PS 1.8, 3.9–4.5). Verse 1.4 prescribes dancing, singing, etc. while residing in a temple for this observance: athāsyāyatanāni mahādevāyatane ’pāṃ samīpe giriguhāyāṃ gavāṃ goṣṭhe ’gnyāgāre vā Now the places of this [observance]. At a sanctuary of Mahādeva, in the vicinity of water, In a mountain cave, in a cow-pen or in a firehouse. (trans. Bisschop and Griffiths 2003: 326)

The agnyāgāra for the observances of the pāśupatavrata has been loosely defined as a ‘fire-house’.46 Bisschop and Griffiths (ibid., fn. 53), while pointing at the occurrence of the word in the Vedic corpus with various meanings, note that in the context of the passage in question ‘the precise meaning of the word is not clear’. Payne (2011: 253, fn. 62) understands the word agnyāgāra and its many synonyms

45. For, indeed, the Pāśupata adept in the first stage of his career was prescribed to live in a temple (āyatanavāsī). 46. Dresden (1941: 151) describes agnyāgāra as a place for keeping the sacred fire. In Bhāradvājaśrautasūtra 1.6.14, agnyāgāra is juxtaposed with āyatana (parisamūhanty agnyagārāṇy upalimpanty āyatanāni; cf. Kashikar 1964: 10).

07_CreSouthV1_8P_11Feb22.indd 242

as a shrine devoted to homa rituals.47 Gonda (1969: 16) noted that in Vedic literature āyatana is ‘the place for keeping the three sacred fires’, as well as the ‘(receptacle of the) fire in which a dead body is to be cremated’, and to ‘the fire-place when it is removable or transportable’. Could the passage quoted above be referring to a fire-chamber associated with a specific ritual function? In 2.1, the text describes the preparation of the āyatana and the ritual: gocarmamātraṃ sthaṇḍilam upalipya gomayenollikhyābhyukṣyāgne prehīty agniṃ praṇīyopasamādhāya paristīrya brahmāṇaṃ kalpayitvā nānyadevatādiśi rudrasya dakṣiṇodapātraṃ sthāpayitvā mahāvyāhṛtibhir agnyāyatane nidhāya rudram āvāhayati // Having smeared a piece of ground of the size of a cow’s hide with cow-dung, having drawn [an auspicious sign] on it, having besprinkled it, having brought forth fire [with the mantra] ‘agne prehi’, having added [fuel to it], having spread [grass] around [it], having prepared a brahmin (?), having placed a water-vessel to the right side of Rudra—not in any other deity’s direction—he invites Rudra, installing him in the fire-place with the Great Utterances. (trans. Bisschop and Griffiths 2003: 328–329)

Subsequently the adept, having invited Rudra in the fire, identifies him with that fire (2.7).48 He then kindles a sacred fire (for the preparation of the ashes) (3.3) and bathes in ashes (3.9). Verses 3.4ff. contain a description and praise of bhasmasnāna. These passages show that the agnyāgāra, besides hosting the fire-ritual, would have served for the performance of ash-related observances. The mention of the ‘right side of Rudra’ (rudrasya 47. ‘Agni-niketana, -sadana, -āgāra: the place at which one performs the ritual of fire. For grand occasions, temples have a hall especially reserved for the cult; but on ordinary days it is performed in the ardhamaṇḍapa, in a kuṇḍa permanently dug into the floor.’ 48. yo agnau rudra ity anumantrayed āvāhane devadevasyāvāhayāmy aham iti // ‘He should speak the mantra “yo agnau rudraḥ” [over the fire (cf. 40.2.1) / image (?)], at the Invitation of the God of gods, [adding the words] “I perform the invitation”’ (trans. Bisschop and Griffiths 2003: 331).

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

243

dakṣiṇa°) in the context of the installation of the of Śaiva manuscripts, prescribes the building of a sacred fire and the performance of bhasmasnāna is ‘kitchen’ (mahānasa) in the southeast: notable as it mirrors the occurrence of the equivasuvīthyāḥ purataḥ kāryaṃ lent word dakṣiṇāmūrti, the ‘right side of the Lord’, diśy āgneyyāṃ mahānasam / in similar contexts, which we will discuss below. dhūmanirgamanopetaṃ A passage in Chapter 11 of the Śivadharmottara pūrvataḥ satramaṇḍapam // 2.131 (also found in the related text Śivadharma)—an early text that we have mentioned above with In front of a nicely built terrace, in the Agni diregard to the vidyāyatana, stemming from a milieu rection (Southeast), a kitchen endowed with a of lay Śaiva worship that has preserved Pāśupachimney has to be built; in the East, a pillared pavilion for the sattra. (trans. De Simini 2016: ta-derived material—describes practices of adepts 386) abiding in the Śaiva āśrama compound that are in harmony with those laid out in first-hand Pāśupata accounts. As such, it constitutes a mix between lay Thus, while the Śivadharmottara does not explicmodes of worship and ascetic lifestyles. It mentions itly identify the ‘library’ or vidyālaya mentioned of the ancillary buildan agnyāgāra as one of the prescribed places to earlier in Chapter 2 with any 52 it mentions a kitchen ings listed in this account, be built by the adept to carry out his observances: or fireplace-building endowed with ‘issues/exits/ outlets for the smoke’ (dhūmranirgamaṇa) in the svāśramād uttare kuryāt puṣpārāmaṃ suśobhanam / southeast. agnyāgārakasaṃyuktam49 A list of (annex) buildings (śālā) with similar aiśānyām īśvarālayam // 11.3 functions is mentioned in later Śaiva Saiddhāntiprādeśasaṃbhave50 kuryāt ka treatises on temple architecture, each linked to yatra syāt saṃbhavo bhuvi / cardinal and intermediary points of the compass śivād dakṣiṇataḥ kuryāt (see, e.g., Ajitāgama, Kriyāpāda 38.43ff., Dagens tadbhaktābhyāgatālayam51 // 11.4 1977: 112–113). In these sources, too, the kitchen (pacanasthāna) is usually located in the southeast, [An adept of the śivāśrama] should make, to or east of the enclosure. On the other hand, some the north of his own residence, a beautiful flowtexts, like the Rauravāgama (Kriyāpāda 26.32, er-garden, furnished with a fire-shed [in the Dagens and Barazer-Billoret 2000: 130), mention southeast] and a temple of Śiva in the northa chapel for oblations (homakuṭi, homasthānā, yaeastern direction. jñālaya, yāgaśālā, yāgasthāna, cf. Colas 1986: 274); If there is no space for it, he should make it while the ca. 8th-century Marīcisaṃhitā (11.3.2, wherever land is available. To the south (or: Colas 1986: 143) locates it in the southeast, the right side) of Śiva (śivād dakṣiṇataḥ) [i.e., the Āgamas place it in the northeast (Dagens 1977: 88). main temple, liṅga, or image], he should make Furthermore, a vidyāsthāna (‘place for knowledge’) a guesthouse for his devotees. is mentioned in Ajitāgama, Kriyāpāda 38.47, but That the agnyāgāraka or fire-shed was built in the in the direction marudvaruṇa (between northwest southeast can be inferred from the context of the and west?). De Simini (2016: 177) regards these passage, mentioning the other directions, as well vidyāsthānas as ‘places specifically meant for teachas other passages in the text. Chapter 2, which is devoted to the ritual of vidyādāna or ‘gift of knowledge’ associated with the copying and worshiping

49. Em.; agnyaṅgārakasaṃyuktam ŚDU; anyāgārasamāyuktam ŚD. 50. ŚD; prādeśasaṃbhave ŚDU. 51. ŚD; tadbhaktānāṃ tathā balam ŚDU.

07_CreSouthV1_8P_11Feb22.indd 243

52. The buildings prescribed for the other directions are a ‘greenhouse of fragrant flowers, provided with cloths’ (gandhapuṣpagṛhaṃ… paṭṭasaṃyutam, northeast), a treasury-house (bhāṇḍāgāraṃ, north), a storeroom (koṣṭhāgāraṃ, northwest), a pond (west), a deposit for fuel and kuśa grass as well as weapons in the southwest: cf. Śivadharmottara 2.132–133.

11/2/22 3:44 PM

244

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

ing activities’.53 Uttarakāmika 67.2 gives details on the vidyāsthāna, to be placed either in the main or intermediary directions, but without specifying where (ibid.: 341). Thus, the direction allotted in relatively late Saiddhāntika sources for this type of ‘library’ building does not seem to match well with the Khmer archaeological evidence of the annex buildings;54 on the contrary, buildings associated with fire, such as kitchens, are usually placed in the southeastern quadrants of temples or āśramas in early sources like the Śivadharmottara, and also in some later Saiddhāntika scriptures. Let us now go back to Chapter 11 of the Śivadharmottara, prescribing the acts of worship in the śivāśrama by the adept. The text explains that one who has taken up residence in the Śiva-temple (śivagṛhāśramin) should worship the Lord at the three junctures of the day, constantly [perform] the kindling of the [sacred] fire (trisandhyam arcayed īśam agnikāryañ ca śaktitaḥ, v. 5ab), and also be devoted to the worship of the image of Śiva and [his] fire (śivārcāgniparo bhavet, v. 16). He should perform worship at the three ‘baths’ (trisnānapūjā, v. 21)—the ash-bath to be taken at the three sandhyās, mirroring the prescriptions found in such Pāśupata sources as the PS/PBh and GK/RṬ. An eulogy of thrice-daily bhasmasnāna is presented in verses 44–54. It lists the manifold fruits bestowed by this practice, which is superior to other baths in water, and calls the ash-bath an 53. A 13th-century inscription from the Naṭarāja temple at Chidambaram describes the activity of a library (sarasvatībhaṇḍāra) situated in the northwestern part of the second enclosure of the temple, in the western maṇḍapa located to the north of the Subrahmaṇya shrine (Sankaranarayanan 1993: 28ff.). This building included a vikramaśolaṉtirukkaiyoṭṭi or ‘storeroom of sacred manuscripts’ entitled to King Vikrama Cōḻa (1118–1135 CE), and was animated by staff members conducting activities of reading and copying of texts. This is not only in harmony with the descriptions of the prerogatives of vidyāsthānas mentioned in Śaiva texts, but also with the prescription given by the Ajitāgama with regard to its location, i.e. to the northwest of the main temple. 54. With the exception of the prescription in the architectural section of the early Saiddhāntika text Sarvajñānottara 19: 30c–32 that ‘A shrine to the East and South of one’s settlement should face west’ (Goodall 2017: 74); this matches with the direction of the doors of the Khmer ‘annex buildings’ with respect to the main sanctuary.

07_CreSouthV1_8P_11Feb22.indd 244

āgneya-bath, a bath ‘belonging or relating or consecrated to fire or its deity Agni’ (āgneya being also the southeastern direction).55 As per the practice of lay persons, he should make a tripuṇḍramark with ashes on the forehead with ‘white ashes’, while the yogin should take a full bath (vv. 45–46). Verse 48 explains that ashes obtained from a Śivāgni (sacred fire of Śiva) are pure (pavitra).56 Verses 49–54 extol the qualities and superiorities of the ash-bath, which leads to unity with Rudra. Along similar lines, a passage of the same text (ch. 5.195–197, shared also by the Śivadharma) lists the benefit of performing the fire-oblation (agnikārya) and erecting a building in bricks (iṣṭakāgāra) for Rudrāgni, which leads to the obtainment of unity with Rudra (rudrasāyujyam) through said ritual, while the ash bath leads to purification from the ten sins.57 The above passages highlight the importance of the ash-bath and the connections existing between that observance and the sacred fire established in the place of worship, possibly in the kitchen or fireshed located in the southeast. Another—more ambiguous, yet early—reference to a Pāśupata ritual involving fire may be found in the Mattavilāsa, a farce composed in Kāñcī in the 7th century. The character of the ‘madman’ (unmattaka), which Ferstl (2015) has proposed to identify with some kind of Pāśupata ascetic (unmattavrata being performed by the Lākula subgroup)—probably a sādhaka, as opposed to his ācārya, Babhrukalpa: The unmattaka walks towards his ācārya and offers the skull bowl he has found. Babhrukalpa, 55. Cf. AVP 40, 4.5, distinguishing it from the vāruṇa and saumya types, on which see Bisschop and Griffiths 2003: 336, fn. 109; cf. Padoux et al. 2000: 173–174, s.v. āgneyasnāna. 56. Svataś śuddhakulāgāvaḥ saṃbhūtantā sugomayam / śivāgninā punaḥ paktaṃ pavitraṃ bhasma tat smṛtam // 11.48. 57. śivāhutyā [śivaṃ stutvā ŚD] jagatsarvaṃ sṛṣṭidvāreṇa [vṛṣṭidvāreṇa ŚD] dhāryate / yat saṃdhāya jagatpuṇyaṃ agnikāryeṇa tad bhavet // 5.195 yaḥ kuryād iṣṭakāgāraṃ rudrāgner atiśobhinam [atiśobhanam ŚD] / sakṛt tatrāgnikāryeṇa rudrasāyujyam āpnuyāt // 5.196 śivāgnikāryaṃ yaḥ kṛtvā [kuryāt ŚD] triyāyuṣam ihātmanaḥ [kuryād āyuṣyam ātmanaḥ ŚD] / mucyate daśabhiḥ [pṛthubhiḥ ŚD] pāpaiḥ śivena [sitena ŚD] sitabhasmanā [śivabhasmanā ŚD] // 5.197.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples however, refuses to accept it and sets off to perform an obscure ritual called the ‘hour of smoke’ (dhūmavelā). This might refer to a certain ritual for the preparation of ashes, a most important requirement for Pāśupata rituals. I am, however, not aware of an exact record for the ‘hour of smoke of the Venerable residing in the eastern region’, as it is called in the MVP. (ibid.: 14)

To explain this idea, Ferstl refers to the prescription concerning the preparation of ashes in AVP 2.1 (Bisschop and Griffiths 2003: 328ff; cf. supra). It is possible that the ‘Venerable residing in the eastern region’ (bhagavataḥ pūrvasthalīnivāsinaḥ) stands for the Sun (Lockwood and Bhat 1995: 79, fn. 49), indicating that this ritual was enacted at the sandhyās, towards Śiva-Rudra conceived or manifested as the Sun. A passage adumbrating just such ritual for the preparation of ashes for the morning sandhyā worship may be RṬ on GK 1.7 (p. 28): tatra pūrvasaṃdhyayoḥ śaucaṃ kṛtvā bhasma saṃskartavyam // kartṛkārakādidoṣarahitaṃ śuklādiguṇayuktaṃ ca bhasmārjitaṃ śivadakṣiṇamūrtau mantraiḥ saṃskṛtya pradakṣiṇaṃ ca dattvā sūryarūpiṇaṃ bhagavantaṃ locanatrayeṇa prasannadṛṣṭyā bhasma paśyantaṃ dhyāyet // In the first two [sections] of this period [that is, the two hours before sunrise] he should cleanse himself and prepare the ashes. The ashes which he obtains should be free from any impurity of agent or other factor in their making and should possess such qualities as whiteness, etc. He should sanctify these ashes to Śiva of the southern image by prayers and after circumambulating the image should meditate on the Blessed One in the form of the sun, who looks at the ashes with kindly glance from his three eyes. (trans. Hara 1966: 548)

Ferstl wonders whether the word dhūmavelā in the clause bhagavata pūrvasthalīnivāsino dhūmavelā, pratipālayāmi ‘I’ll observe the “hour of smoke” for the Venerable residing in the eastern region’ (Mattavilāsaprahasana, p. 60, 1f.), may be ‘a corruption of dhūm(r)avedyām (cf. bhasmavedyām uṣitaḥ in PABh, p. 4,4)’, which would imply some kind of ritual activity of the homa type. But he points out

07_CreSouthV1_8P_11Feb22.indd 245

245

that ‘Any activity with fire, however, is forbidden in order to observe the general commandment (yama) of non-injury (ahiṃsā) (cf. PABh, p. 17, lines 7–8)’. However, textual evidence from South and Southeast Asia suggests that some Pāśupatas (or, rather, the post-Pāñcārthika developments/ subgroups that arose within the Atimārga movement in the early mediaeval period) might have actually performed fire rituals. Compare, for instance, Śivadharmottara Ch. 11 vv. 16–18, describing a Śaiva hermit—in strikingly Pāśupata terms—as entirely devoted to the fire of the Śaiva worship (sivārcāgniparo), constantly engaging in meditation of Śiva (śivadhyānarataḥ sadā), depending on ashes (bhasmaniṣṭho), having as abode the liṅga/the hermitage of Śiva (text edition as in De Simini 2016: 51, fn. 148); and Niśvāsaguhya 9.8cd, referring to a Pāśupata or Lākula practitioner as follows: japahomarato nityam pratigrahavivarjjitaḥ / triṣkālam arcayed devaṃ triṣkālasnānam ācaret // Constantly devoted to muttering and fire-ritual, not accepting any gifts, one should worship God three times [a day]; one should carry out [ash-] bathing during the three times [of the day].

The above material would seem to reflect an evolution of Pāśupatism from a fireless ascetic movement to one admitting homa rituals. An analogous stance is reflected in the characterization of Atimārga from the perspective of (a localized form of) Mantramārgic Saiddhāntika Śaivism in the Sanskrit-Old Javanese Śaiva text Bhuvanakośa, preserved on Bali. Dyads 25–30 of Chapter 8 refers to sakala (‘material, visible’) practices, which are inferior to the gnosis-oriented niṣkala (‘immaterial, invisible’) ones of the (Śaiva)siddhāntajñāna, as a mixture of homa ritualism with typically Pāśupata modes of worship, like ashes (homabhasma 8.25; homabhasmavidhāna 8.28c), mantras, salutations (namaskāra), hūṃdhrūtkāra,58 etc. As suggested by the use of such technical terminology, the passage 58. This is probably a variant form or local development (attested multiple times as such in the text) of huḍḍukkāra; see supra, fn. 38.

11/2/22 3:44 PM

246

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

may reflect a localized evolution of Pāśupata Śaivism along ritualistic lines, which is consistent with a trend documented across the wider Indic world. The textual passages discussed above confirm the importance of ashes and fire-rituals in early Śaiva contexts, and suggest the connection of ashes with a place (āyatana, fireplace-house, or kitchen) in the southeast or south of the image/main temple to carry out observances and rituals. These descriptions are compatible with the interpretation of Khmer annex buildings as fire shrines, and we surmise that Pāśupata Śaivism must have played a role in setting the doctrinal and ritual background for the establishment and functioning of these edifices across the ages in the Khmer domains. Indeed, the conservative nature of Khmer Śaivism has been noted by scholars (see, e.g., Sanderson 2001: 22–23, fn. 28). It is therefore possible that the Atimārga developed independently in the Khmer domains and remained influential, and representative, in the Śaiva paradigm throughout the Angkorian period, that is well after the movement lost its prominence over large areas of the Indian subcontinent.

pāśupata initiation rituals, the ‘giFT of knowledge’, and śiva dakṣiṇāmūrti As we have mentioned above, ash plays a crucial role in the Pāśupata ascetic regimen. It is not only essential for purification, but it also constitutes a prime ingredient in the Pāśupata initiation ritual (dīkṣā), when the teacher transforms his disciple using ashes empowered with Sadyojāta and other mantras. The ācārya is an integral part of the initiation ritual and Pāśupata texts accord the greatest importance to his role. Śiva imparts knowledge through the ācārya and by the act of offering service, the disciple receives the vidyā or knowledge (or: ‘book’) from him.59 Some of these interactions—such as rubbing oil, cleaning the guru’s feet, etc.—between a Pāśupata ācārya and his disciples are carved on the Mỹ Sơn E1 pedestal from Campā, Vietnam (Fig. 7.13; see Chemburkar and Kapoor 2018). Here we see an image of a 59. We will elaborate on vidyā as ‘book’ below.

07_CreSouthV1_8P_11Feb22.indd 246

bearded ācārya wearing a yogapaṭṭa (cloth meditation band) across his knees and another instructing a practitioner who touches his guru’s feet. The transmission of the religious knowledge depends on a strong personal link between the ācārya and practitioner and is crucial for the propagation of the tradition. The ācārya provides spiritual support and sustenance along with doctrinal and yogic instruction. It is the ācārya who creates a spiritual environment or āyatana for the disciple to practise: in PABh 1.1 (p. 4), the āyatana was the place established by Lakulīśa to ‘initiate’ his first disciple, Kuśika, into the system: Having assumed the marks of the transcendent stage of life he delivered his teaching. From the words ‘bathing and lying in ashes, supplementary bath, wearing flowers taken from an image and wearing a single cloth’ [PS. 1.2–5 and 10], We may infer that he dwelt on an altar of ashes in a place pure for the reception of pupils, at a temple as in his scripture is set forth in order to make the location clear. And then, the blessed Kuśika came to this teacher and seeing in him the marks of excellence such as perfect contentment and in himself just the opposite, he clasped the teacher’s feet60 and declared in accordance with rule his caste, his family-name, his education and his freedom from debt.61 (trans. Hara 1966: 157–158)

Thus, the āyatana in origin was thought of as a pure place to reside (PS 1.7: āyatanavāsī; PABh, p. 12), receive pupils,62 and conduct rituals. This āyatana, initially understood as a sacred space where a disciple receives the transmission of the

60. This passage may provide the context for identifying the relief on the Cam pedestal depicted in Fig. 7.13 as sādhaka and his prospective ācārya ‘reenacting’ an all-important moment in the history of the tradition, namely Kuśika coming to Lakulīśa, representing the former clasping the feet of the latter. 61. atyāśramaprasiddhaṃ liṅgam āsthāya pravacanam uktavān bhasmasnānaśayanānusnānanirmālyaikavāsograhaṇād adhikaraṇaprasiddhyarthaṃ ca svaśāstrokte āyatane śiṣyasambandhārthaṃ śucau deśe bhasmavedyāmuṣitaḥ //. 62. Compare with the placement ‘at the right side of Śiva’ of a guesthouse for the reception of devotees (tadbhaktābhyāgatālayam) in Śivadharmottara 11.4, quoted above.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

247

Fig. 7.13: Mỹ Sơn E1 Pedestal relief depicting a disciple attending to his guru. (Photo: Trần Kỳ Phương )

tradition from Śiva through the lineage of a guru,63 may also have had a function in the ritual of initiation into the Pāśupata order. While keeping in mind the textual passages discussed above, we will explore more below. In stanza 5 of the GK we find a listing of the items constituting the last of the five categories of the Pāśupata system: ‘The material adjunct, the proper time, the proper ritual, the image of god and verily the master’. The commentary on the last three items runs as follows: By the proper ritual (kriyā) is meant the ritual of initiation pertaining to the image of God (kāraṇa) and to the pupil. The Order of this ritual is to be understood from the Saṃskārakārikā. By image (mūrti) is meant a spot of ground to the south of, and not separated by the ascetics’ hut etc., from the seat of worship of God 63. It has been argued that the term gurvāyatana attested in Sanskrit inscriptions from the subcontinent has been variously interpreted as a sacred space where a disciple receives the transmission of the tradition, or as a temple dedicated to the liṅgas of deceased Pāśupata masters. The Mathurā Pillar Inscription of Candragupta II mentions the installation of two liṅgas by Uditācārya for his teachers in the gurvāyatana (Bhandarkar 1931: 4). However, Sircar (1942: 271) argues that the teacher’s shrine was not a memorial but a place for the living teachers (for a recent discussion, see Lefèvre 2011: 48–49). The remnants of such ash-related practices may still be seen in the Śaiva landscape of the holy town of Rishikesh in northern India. At the Someśvara temple, in the southeastern section, is the space for the priest and ascetics to honour their guru, with a log of white wood kept burning in front of the guru’s photograph, which also provides the ashes for the Śaiva ascetics.

07_CreSouthV1_8P_11Feb22.indd 247

(mahādeva), which should be [an image] characterized by erect phallus, etc. as described in the sūtras on presentation (upahārasūtra). By master is meant teacher […].64 (trans. Hara 1966: 501–502)

The passage identifies the mūrti in the kārikā as referring to a spot on the ground in which the initiation should be carried out, at the south of the image of Lakulīśa. The initiation involves the substances (dravya) of ashes, produced from the sacred fire. This passage recalls the preparation of an āyatana by Lakulīśa, Śiva’s incarnation, initiating Kuśika in PABh 1.1, as if to ‘repeat’ that important moment in the mythical history of the movement. The text mentions the Saṃskārakārikā, a ritual manual of the Pāśupatas. This source elaborates on the practice mentioned in the RṬ in a more detailed manner: [Answer:] The [appropriate] time [of the transformatory rite] is the morning, and the [appropriate] place is [a place] where the southern face [of the image (i.e. liṅga) located]. Materials required are the kuśa-grass and so on, and the mantras [commended for the rite] are Sadyo [-jāta] and the other brahmamantras. (4)

64. kālaḥ pūrvāhṇaḥ // kāraṇamūrtiśiṣyayoḥ saṃskārakarma kriyety ucyate // tatkramaś ca saṃskārakārikāyāṃ draṣṭavyaḥ // mūrtiśabdena yad upahārasūtre mahādevejyāsthānam ūrdhvaliṅgādilakṣaṇaṃ vyākhyātaṃ tatsamīpadakṣiṇabhūpradeśaḥ kuṭyādyavyavahito ’trābhipretaḥ // gurur ācāryaḥ […].

11/2/22 3:44 PM

248

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

   The teacher is the agent of the transformatory rite; the cause is God. […] (5)65 (trans. Acharya 2007: 37).

indicates the Pāśupata’s position with respect to Mahādeva [Supreme] and his embodiment.

According to Bakker, this would be applicable The disciple is instructed to bring the sacred ashes ‘to every situation when the Pāśupata practitionand sanctified water-jars (41), which should be er enters into contact with his object of worship.’ sanctified by a teacher who has good qualities and Therefore, the ācārya would constantly personify is bathed in ash (43–45). Then the teacher, med- Śiva for the initiate during the initiation rite, in itating on Lakulīśa, and cognizant of the way of which the place of the physical image, i.e., the union with Śiva, should perform ritual acts and mūrti, may be taken by the guru who represents offerings to the bodies of God, then bathe the God Śiva (ibid.) Bakker (ibid.: 128) further points to a possifirst in ashes and then in pure water (58–61ab). ble confusion regarding the word, which in the This brings about the commencement of union course of the centuries became reinterpreted as the with Śiva (62cd–63ab). The second phase of the southern face or form of Śiva, especially in South ritual is described in vv. 81–84 (quoted in full infra, p. 251). The adept, whose members are besmeared India: ‘Temple southern walls were suitable for with ashes, and who is meditating on Śiva’s da- showing his role as teacher and tradition become kṣiṇāmūrti, undergoes the imposition of the hands reoriented and his right side became its southern face’ (ibid.: 129). He (ibid.: 133) then discusses a on his head by the ācārya (śivahasta). There are several interesting elements in this lintel that originally may have been placed at the passage, which merit a detailed discussion. We entrance of a Pāśupata temple found in Nāgārī would like to start by from the concept of da- (Rajasthan) depicting a scene from Dakṣa’s sacrikṣiṇāmūrti, already mentioned in many of the fice, where Mahādeva instructs Dakṣa on the Pāśupatayoga. Mahādeva (facing east) leans towards the sources describing the places for the Pāśupata right, casting a sidelong glance 45 degrees south observance and/or initiation quoted above. As towards Dakṣa, who is facing north. It seems to us Pāśupata texts prescribe that the initiate should that this arrangement could reflect the position of consecrate the ashes with mantras in the vicinthe adept who receives initiation from Mahādeva ity of the dakṣiṇāmūrti,66 literally ‘the southern/ (dakṣiṇāyāṃ mūrtau), ‘and by analogy the position rightward image/body/figure’, Hans Bakker (2001) of the student and the teacher, which replicates argues that the initiation site of a practitioner is the one prescribed for teacher and pupil in some at the mūrti, a consecrated place on the right side branches of Vedic literature’ (De Simini 2016: 56, (dakṣiṇa, or south side) of the manifestation of the fn. 163), namely the southeast, or āgneya, direcsupreme God or Mahādeva. On the basis of his tion, where the Cambodian annex-buildings are readings of Kauṇḍinya’s PABh and other sources, usually located with respect to the main temple. Bakker (2004: 127) argues that the term mūrti in Furthermore, in Sanskrit texts, whether the dathe Pāśupata context refers to the bodily manifes- kṣiṇāmūrti represents a liṅga, a temple, the guru, tation as well as to the place of the image: or an actual image, the disciple should approach it facing north while standing at God’s right side, and The concept of daksiṇāmūrti thus comprises the thereby seeing the southern face of the ācārya, liṅga, manifested form (rūpa) of God, the (physical) or image (Bakker 2004). It is perhaps significant image or body (mūrti) in which it may be envisthat the fire-ritual section of the Saiddhāntika text aged and the right side (daksiṇe parśve), which Somaśambhupaddhati prescribes that the adept should seat himself facing north with the centre 65. [kālaḥ] (prātas tathā deśo yatra) mūrtis tu dakṣiṇā / of the fire-pit (kuṇḍa) in front of him to perform upakaraṇaṃ darbhādyaṃ mantrāḥ sadyādayaḥ smṛtāḥ // the homa in the fire-house (agnyāgāra).67 Does this 4 saṃskārakartā ācāryo nimittaṃ kāraṇaṃ smṛtam […] // 5 (ed. Acharya 2007: 29). 66. PABh 1.1 (p. 8), 1.9 (pp. 14–15); RṬ 7 (pp. 17–18); cf. Bakker 2004: 128.

07_CreSouthV1_8P_11Feb22.indd 248

67. Somaśambhupaddhati, section 4: ‘1. And now, after having obtained the permission of the master, he makes

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

249

injunction reflect the continuation of an originally the Pañcārthavidyā, a term possibly denoting the Pāśupata practice? Pāśupatasūtra (see Acharya 2007: 42–43, fn. 95). It must be noted that some scholars have pro- Śiva imparts knowledge through the ācārya and posed to interpret some of the images at Ellora the disciple receives the vidyā or knowledge (or: and Elephanta, previously identified as Yogīśvara ‘book’) from him.69 We have already mentioned or Lakulīśa (Collins 1982: 612), as Dakṣiṇāmūrti above depictions of Pāśupata ascetics carrying a (Mankodi 1988); and that architectural sections manuscript on Khmer and Cam temple reliefs. found in some Southern Indian Saiddhāntika The Śivadharmottara (Chapter 2) associates the treatises locate icons of either Dakṣiṇāmūrti or ritualized transcription and ‘offering’ of the manLakulīśa in the southern quadrant of Śaiva temples uscript with the Śaiva knowledge with worship (e.g. Rauravāgama, Kriyāpāda, 39.33–35; Ajitāga- for Śivāgni and the teacher,70 and also prescribes ma, Kriyāpāda 14.75; see Dagens 1977: 106). the installation of that manuscript ‘in a cleansed, Clearly, an iconographical and symbolical connec- pleasant place in the presence of Śiva’ (or, ‘in front tion between these three manifestations of Śiva can of Śiva’, śivasya purataḥ), to be worshiped ‘with be established. Lakulīśa represents the prototypi- the [same] devotion addressed to a teacher’.71 Is cal initiator, or ascetic Śiva Yogīśvara. The South it possible that vidyādāna evolved from an origiIndian Dakṣiṇāmūrti carries the book, symboliz- nally a Pāśupata ritual connected with initiation, ing the (Śaiva) knowledge and its transmission. In involving a ritual fire? fact, devotion towards Lakulīśvara is relevant in Having considered the practice of the donation the Śivadharmottara, ‘which gives prescriptions of a manuscript by the ācārya to the disciple, De for the installation of the icon of the god precisely Simini (2016: 319, fn. 792) discusses the opposite in the context of the gift of knowledge’ (De Simini scenario, and draws attention to the fact that the 2016: 52).68 A tentative yet fascinating link between Anteṣṭividhi early Śaiva initiation and the bestowal of knowlmentions a manuscript among the possessions edge as a book from a master to a disciple could of a sādhaka initiate that, at his death, must thus be proposed. De Simini (ibid.: 319, fn. 792) be donated to the ācārya who performed the elaborates on the hypothesis by Diwakar Acharya funerary rite: ‘Pavitra, manuscript, disciple, that pre-tantric (i.e. Pāśupata/Atimārga) initiation may have involved the donation of a manuscript. The Saṃskāravidhi in st. 54 seems to make a brief 69. The word pañcārthavidyā, and his synonym pañcarreference to the veneration of the manuscript of thajñāna, refer to both the ‘knowledge of the five categories’ his return to the residence of Agni, and circumambulates in silence, with all the utensils of the cult, and carrying in his hands a vase of arghya. 2. Then, looking upon all of the materials of the sacrifice with a divine gaze, he seats himself facing north, the center of the kuṇḍa in front of himself’. (trans. Payne [2011: 239] of the French translation by Brunner [Brunner-Lachaux 1963]). 68. Śivadharmottara 2.146–147: ‘There, according to rule, one should install Śiva, made of clay, wood or stone, who is the author of all treatises, omniscient, Lord who bears a club, (146) / Surrounded by pupils and pupils of pupils, with his hands raised in the act of teaching, seated in the lotus position, lord of the gods, a master whose speech is vivid (147)’; tatra mṛddāruśailam vā sthāpayed vidhivac chivam / sarvavidyāvidhātāraṃ sarvajñaṃ lakulīśvaram // 146 vṛtaṃ śiṣyapraśiṣyaiś ca vyākhyānodyatapāṇikam / padmāsanasthaṃ deveśaṃ prasannavadanaṃ gurum // 147 (text and translation De Simini 2016: 52).

07_CreSouthV1_8P_11Feb22.indd 249

(i.e, the Pāñcārthika Pāśupata system) and the ‘treatise about the five categories’, namely the PS. 70. ‘Once [the transcription] is completed, he should again perform worship, according to the former procedure, for Śivāgni, the teacher, and knowledge, having fasted with effort’ (42), sampūrṇe pūrvavidhinā punaḥ pūjāṃ prakalpayet / śivāgniguruvidyānāṃ sopavāsaḥ prayatnavān // (trans. and text as in De Simini 2016: 378, 395). Compare v. 14: ‘Having worshipped Śiva according to rule, one should then worship his knowledge, and [worship] with devotion the teacher as if he were Śiva, because this triad is similar’, śivaṃ sampūjya vidhivat tadvidyāṃ pūjayet tataḥ / guruṃ ca śivavad bhaktyā tulyam etat trayaṃ yataḥ // (ibid.: 375, 393). 71. Verse 60, ‘Having gently placed it in a cleansed, pleasant place in the presence of Śiva, having bowed to this [manuscript] with the [same] devotion addressed to a teacher, he should make offerings’, sthāne susaṃskṛte ramye śivasya purataḥ śanaiḥ / sthāpayitvā guror bhaktyā taṃ praṇamya nivedayet // (De Simini 2016: 380, 397).

11/2/22 3:44 PM

250

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.14: 11th-century Banteay Srei annex building in the southeastern corner of the first enclosure of the temple housing a Śiva dakṣiṇāmūrti pediment on its entrance that faces the main sanctuary. (Photo: Swati Chemburkar)

and whatever [possessions] may be there, such as a staff, should be handed over successively to the teacher who has helped [the deceased] to be [united] with the Lord’; pavitraṃ pustakaṃ śiṣyaṃ kiñcid daṇḍādikaṃ kramāt / tad ācāryāya dātavyaṃ sanāthaṃ yena kalpitam // 44.

If the Khmer Annex buildings were (or, rather, became over time) libraries of some sort, could they have served the function of gurvāyatanas, i.e. temples associated with teachers and sacred fires, which also functioned as ‘resting places’ for the books (pustakāśrama) donated in this manner? De Simini (ibid.: 146) notes that Chapter 11 of the Śivadharmottara and the inscriptional records reflect a situation in which a manuscript collection attached to an āśrama is formed by private donations, and that such ‘places intended for the storage of manuscripts might also have hosted the practice of ritual activities focused on the manuscripts.’ She refers to Buddhist examples from Northwestern India and Tibet discussed by Fussman, i.e. an isolated chapel or hermitage building detached from the main monastery, where ‘only a hermitic teacher (ācārya), part of a lineage of ācāryas officiating rituals for the laypeople, could have lived

07_CreSouthV1_8P_11Feb22.indd 250

before it was abandoned for good’, and which could have stored the manuscripts (ibid.: 147). The Śaiva tradition could have shared similar concerns for properties (including manuscripts) of the deceased teachers and adepts. The evidence analysed thus far would seem to establish a connection between a place at the right of the image (i.e. south, possibly southeast), bestowal of instruction (book(s), Śaiva knowledge) and fire (for homa and initiation, and perhaps also the production of ashes) in an early Śaiva context. This leads us to ask whether the Śiva dakṣiṇāmūrti pediment on the annex building (Type B) placed in the southeastern corner of the first enclosure of Banteay Srei hints at the above Pāśupata tradition (Fig. 7.14).72 Here Śiva as a guru is on the right side of the supreme god of the main sanctuary. Did the annex buildings represent an evolution, in terms of architecture, doctrine, and praxis, of an original (gurv)āyātana space where initiation rituals took place? 72. All the pediments on the southeastern annex building depict Śaiva themes, clearly indicating its Śaiva association, whereas the northwestern annex building bears reliefs of Kṛṣṇa, indicating a Vaiṣṇava association.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

sacred fire and significance of the ācārya The association between initiation- and/or homa-rituals, a sacred fire, and an ācārya (conceived of as an embodiment of Śiva) that existed in early Śaivism in the Indian subcontinent may have been current also among the Khmers, and could have developed into a ‘regional’ tradition that led to the creation of annex buildings in several Khmer Śaiva temples. The majority of early Khmer images of Śiva either depict him as a teacher, or as an ascetic. Furthermore, one notes an association between Śiva and fire (agni), and between the ācārya and agni in the context of initiation, which stems from the Indian tradition. Paśupati, one of the eight names of Śiva, is regarded as one of the names representing fire by the Śatapatha Brāhmaṇa (1.7.3.8), and the God of fire Agni and the master are seen as homologous in that text (Bakker 2004: 118–119). Similarly, in the Vāyu Purāṇa (27.53) the body of Paśupati is said to be Fire (paśupater […] tanur agnir), while the Brahmāṇḍa Purāṇa (1,10.45cd–48ab) states: The fifth name of yours that had been mentioned by me is Paśupati. Let this agni be the fifth body of your fifth name. On this being mentioned, the fiery brilliance that had been stationed in his body and is termed uṣṇa (hot) entered the fire. Hence it (fire) is Paśupati. Since agni is paśu and since he protects animals, the physical body of Paśupati is designated as agni.73 (trans. Tagare 1958: 106)

The ācārya’s sacred fire symbolizes his union with Śiva. It also provides the ashes necessary for the disciple to undertake ritual cleansing and attain the same state, and it is central in the ritual of initiation, in which the ācārya imposes the hand on the head of the sādhaka and then cuts his topknot. This is described in a passage of the Saṃskārakārikā. After the preparation involving the besmearing of 73. nāma yadvai paśupatir ity uktaṃ pañcamaṃ mayā // 45 pañcamī pañcamasyaiṣā tanur nāmnāgnir astu te / ity ukte yac charīrasthaṃ tejas tasyoṣṇasaṃjñitam // 46 viveśa tat tadā hy agniṃ tasmāt paśupatis tu saḥ / yasmād agniḥ paśuś cāsīd yasmāt pāti paśūṃś ca saḥ // 47 tasmāt paśupates tasya tanur agnir nirucyate /.

07_CreSouthV1_8P_11Feb22.indd 251

251

the disciple’s limbs with ashes and the purification of the body and soul (67–80), the teacher should approach the Southern face (dakṣiṇāṃ mūrtim āśritaḥ, 79), and perform śivahasta (82): Having meditated on Śiva in the [southern] face with devotion, he should meditate on the teacher who holds a staff, and think that his hand is placed on the head of the sādhaka. (83) The experienced teacher should perform the rite of renunciation with regard to material [possessions] (dravya), mental states (bhāva) and the [sacred] fire, and then cut off the topknot by the śivahasta reciting the mantras. (84)74 (trans. Acharya 2007: 46)

The prescription that the ācārya should approach, and meditate on, Śiva Dakṣiṇāmūrti and the teacher who holds a staff (śrīlakuladhāriṇam, i.e. Lakulīśa) while initiating the disciple near the ritual fire is relevant, and supports the associations we have suggested in the previous section. Indeed, beyond being crucial for the production of pure ashes, the sacred fire played a role in Pāśupata initiation and, possibly, soteriology too. While first-hand Pāśupata sources give us scant detail on this matter, some interesting information can be gleaned from the Paramokṣanirāsakārikā on a few (lost) verses of the Raurava by early Śaiva Saiddhāntika exegete Sadyojyotis (early 8th century) and the Vṛtti by Rāmakaṇṭha (early 11th century), where certain (pre-)Saiddhāntika Śaiva views, including various traditions of Pāśupata Śaivism, are refuted. While the discussion of these references would take us too far from the present subject, suffice here to say that the commentary on verse 3d associates the Pāśupatas with the soteriological position of ‘having agency that belongs to fire’ (āgneyakartṛtva), which leads to ‘becoming the same as “The Treasury of Light”’ (samatvaṃ tejasaṃ nidheḥ), i.e. Fire (or the Sun), referring to liberation as the powers of the Lord being transferred into the liberated soul, 74. dhyātvā mūrtau śivaṃ bhaktyā śrīlakuladhāriṇaṃ gurum / tasya hastaṃ punar dhyāyed dattam sādhakamūrdhani // 83 dravyabhāvāgnisaṃnyāsaṃ krtvā samyag vicakṣaṇaḥ / śikhācchedaṃ tataḥ kuryāc chivahastena mantrataḥ // 84 śivamūrdhnā trisūtrīṃ tu sādhakasya śikhāṃ nayet / pañcabrahmamayo hy eṣa śivahasto hy udāhṛtaḥ // 85.

11/2/22 3:44 PM

252

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

i.e. śikhāsaṅkrānti (Watson et al. 2013: 222).75 It is, therefore, possible that Pāśupata initiation, while not having the liberating power of Mantramārga initiation, was conceived of as a preliminary metaphorical act implying the transference of powers from the Lord (via the teacher) to the initiated disciple, visualized as the identity between the sacred fire and its scintillae. This process would culminate after the death of the sādhaka with his final and complete union with Rudra. The transference would be symbolically enacted in this world and life during the initiation ritual, in which the ācārya (who embodies Śiva, after having literally become him through the ritual procedure and meditation) ‘cuts’ the bond of the initiate, transfers the divine powers to him through the fire, and bestows upon him the gnosis (vidyā), intended as the Pāśupata scripture. This constellation of ideas and practices, especially the ritual of śivahasta discussed above, seems to be represented in the relief of the Bayon’s inner gallery, where an ācārya (doubling as Śiva), seated in the classic dakṣiṇāmūrti pose touches the head of the ascetic adept while both are surrounded by flames in the background (Fig. 7.15). Could the flames contouring both figures represent the ‘transfer’ of divine powers from the ācārya-Śiva to the sādhaka? Many Śaiva ascetics are depicted on the inner gallery of the Bayon, some of which contoured with flames, which Provost-Roche (2010: 53) interprets as a representation of their ascetic power (tapas). Apart from Pāśupata dīkṣā, analogous ideas might have been at play in the tantric fire-ritual of homa, which, according to Gonda (1970: 85), represents ‘by means of techniques partly Vedic, partly purāṇic and partly Tantric, the enactment, in a gradual process of development, of the realization of the unity of the worshipper’s soul, the place and material for worship and God.’ As we have suggested above, homa rituals may have been practised by some post-Pāśupata movements within the Atimārga, and such texts as the Saṃskārakārikā 75. According to another interpretation, taking the meaning of śikhā (in the compound viśikhā occurring next to āgneyakartṛtva) to be ‘topknot’, this would be a reference to the ritual of the removal of the disciple’s topknot during the śivahasta part of initiation. These interpretations may not be mutually exclusive.

07_CreSouthV1_8P_11Feb22.indd 252

Fig. 7.15: Ācārya (doubling as Śiva) is blessing the ascetic amid flames. Inner gallery relief on the north wall between BY32 and BY33, 13th-century Bayon temple. (Photo: Shivani Kapoor)

stress the salvific role of fire-rituals, leading to union with Rudra (cf. above, p. 248); hence, it seems likely that the properly tantric varieties of this ritual that rose to prominence during the medieval period could have stemmed from the pretantric tradition.

archaeological finds from the annex buildings The archaeological finds from the annex buildings have not been systematically studied, but they may be worth looking at.76 These include two types of navagraha or nine planetary divinity friezes77 76. For a study on the archaeological contexts of the ‘resthouse temples’, see Hendrickson 2008a. 77. The first type is usually found at the bottom of the wall that faces the eastern entrance of the annex build-

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

253

Fig. 7.16: 3D scanning of the navagraha panel seen in situ at the skirting level of the Preah Ko temple wall. (Photo: Olivier Cunin)

Fig. 7.17: Traces of a navagraha panel at Prasat Khna annex building. (Photo: Shivani Kapoor)

(Figs. 7.16, 7.17 and 7.18), which according to Dagens (2003: 193) could have been the object of ings. It can be still seen in situ at 9th-century Preah Ko (Fig. 7.16), while a trace of it can be seen at Prasat Khna (Fig. 7.17). The second type was seen at East Mebon according to Louis Malleret’s notes, with a reverse side carved with seven ascetics commonly understood as saptarṣi (Malleret 1960). The frieze is now kept at the Angkor Conservation Depot (Fig. 7.18). It should be mentioned that only four of the nine divinities have been certainly identified as grahas, while the other five could correspond to directional deities (Bhattacharya 1961: 142–143).

07_CreSouthV1_8P_11Feb22.indd 253

a specific cult until at least the 12th century. The worship of planets was associated in both India and Kambujadeśa with rites of oblation to fire (homa; see Bhattacharya 1961: 148; 1964). The cult of Fire held a special status in Khmer courts (Soutif 2009: 319), and might have been reserved an independent cult in temples (Bhattacharya 1961: 147). Numerous inscriptions referring to lakṣahoma, and koṭihoma are documented in Khmer epigraphy.78 78. K. 95 (st. XXVIII, Bergaigne 1893: 355); K. 300 (st. XX,

11/2/22 3:44 PM

254

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Fig. 7.18: Second type of navagraha frieze from East Mebon, kept at the conservation depot depicting navagrahas on one side and ascetics on the other (accession no. DCA.3506.A). (Photo: Swati Chemburkar)

The AVP (31, 10.2), in the context of the koṭihoma, specifies the importance of the ‘observance of Paśupati’ (paśupater vratam) also described in AVP 40, the knowledge of which is a requisite for the performance of this ceremony and the obtainment of its fruits (etaj jñātvā tu yaḥ samyak koṭihomaṃ prayojayet / sarvān kāmān avāpnoti brahmalokaṃ sa gacchati). Sections of the AVPs IC, IV: 254); K. 418B (Finot 1904: 672–679; Cœdès 1929: 305, 308); K. 692 (st. LIV, IC, I: 237); K. 933 (IC, IV: 47); K. 194 and K. 383 (Cœdès and Dupont 1943: 56–154); K. 806 (st. CCXXXVI, IC, I: 73); K. 872 (st. XIII, IC, V: 97).

07_CreSouthV1_8P_11Feb22.indd 254

were composed against a backdrop of Pāśupata Śaivism, and the performance of these rituals along with the presence of navagraha panels in annex buildings further corroborate their association with fire rituals and purohitas. Furthermore, there is a link between homa rituals and abhiṣeka or royal initiation, for the former were performed to overcome any obstacles that might have arisen during the latter. Lokesh Chandra (1980: 317) refers to a Japanese Buddhist maṇḍala ritual including among its steps the invocation of Agni as well as the navagrahas and signs of the Zodiac (rāśi).

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples

255

Fig. 7.19: 3D scanning of the archeological finds from the annex buildings by Olivier Cunin. The square shaped stone was found in the north annex building of the Banteay Srei-style monument to the east of the North Khleang at Angkor Thom. The circular shaped stone was found in the south annex building of Banteay Srei.

As pointed out by Soutif (2017: 14), no fixed kuṇḍa has ever been recovered from a Khmer temple. At the Bayon, four fixed kuṇḍas have been identified as being depicted in the bas-reliefs in the southeastern sections of the second and third galleries (ibid.), which is consistent with the direction of Agni and the location of the annex buildings. A quadrangular concave stone preserved at the National Museum of Cambodia (KA 442), discovered at Sambor Prei Kuk, bears the formula oṃ jaiminaye svāhā (addressed to Agni’s wife) engraved on his four sides, and as such constitutes one of the earliest examples of (movable) kuṇḍas (ibid.: 18). At Banteay Srei, a circular stone base was found in the south annex building, which bears certain markings. A similar square shaped stone was found in a north annex building of a Banteay Srei-style monument to the east of the North Khleang sanctuary in Angkor Thom. The function of these stones has not been ascertained yet, but they might be homa-kuṇḍas (Fig. 7.19).

07_CreSouthV1_8P_11Feb22.indd 255

conclusion A particular type of annex building was developed during the pre-Angkorian period within Khmer temples that continued to be erected for four centuries. Commonly called ‘library buildings’, these structures remain a puzzle. This chapter has made a new attempt to understand the possible function(s) of the annex buildings, based on the Khmer inscriptions and evidence of Khmer Śaiva religious practices of the period as well as archaeological finds. It has also situated the majority of annex buildings in the context of Pāśupata practices in an āyatana space for initiation and other rituals that formed a backbone of the movement. Having rejected the idea that these buildings constituted ‘libraries’ in the sense of spaces for copying and studying text, our hypothesis is that the architecture, symbology, and actual functions of these buildings reflect a constellation of ideas and ritual practices involving a sacred fire, preparation of ashes, initiation by an ācārya, donation of the Śaiva knowl-

11/2/22 3:44 PM

256

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

edge and its material medias, i.e. manuscripts, and the sacred fire. The kind of location prescribed for storage of the same. It is not impossible that the the ritual is usually at the southern/right side of the concept of the building stemmed from early ideas Lord, represented by an image, liṅga, or, in the case found in Pāśupata texts about the āyatana as a of a temple, the main sanctuary itself. The emplaceplace provided with a fire for the production of ment and architecture of the annex buildings do ashes, the performance of bhasmasnāna, and in- fit these descriptions, being located approximately itiation rituals, and then developed into a space 45 degrees southeast from the main shrine—the conceived for a ritualized Śaiva ‘cult of the book’, direction of Agni, associated with kitchens or fire along the lines suggested by De Simini, during the shrines in prescriptive Śaiva literature. While the development of the annex building period in which Mantramārga Śaivism was dominant. This evolution might explain the possible appears to be coterminous with the arrival and development of the Pāśupata sect in the Khmer dual purpose of the annex buildings. Khmer temples are normally viewed through domains, epigraphy from the Angkorian period the prism of Indian temple architecture, but the reflects the presence of Śaiva Mantramārga tradiKhmers’ adaptation of Indic texts and rituals ac- tions, similar to what may be seen in the Indian 79 tually resulted in a unique vocabulary of art and Subcontinent. But despite the arrival of new texts and the doctrinal and ritual evolution of Śaivism, architecture, which demands finding a different and independent perspective. To take the most evidence of the Pāśupata rituals established in striking examples, it is most rare to find an Indian the royal temples continues to be visible in the 80 temple that is a literal depiction of Mount Meru, Khmer domains; and, as remarked by Sanderson such as one sees at Bakong, Bakheng, Preah Rup (2003–2004: 435–436), Śaiva temples and associand Angkor Wat in Cambodia. Similarly, to seek an ated hermitages for the support of Śaiva ascetics equivalent Indian structure for these Khmer temple ‘were institutions that appear to have remained annex buildings would be futile. The Khmer annex unchanged in their fundamentals whatever the buildings are one of the many Indic art and archi- initiatory affiliation of the religious attached to tectural innovations created in Maritime Asia, and them.’ Furthermore, Pāśupata beliefs and rituals may have been absorbed and reenacted by the their function must be derived both from local evigenerations of local practitioners as inherently dence and South Asian prescriptive literature. The Khmer: after the mid-10th century reign of Rāannex building thus could represent either a purely jendravarman II, the Pāśupata movement may Khmer innovation, or reflect early ideas found in have lost its influence on the royal elites over time Sanskrit prescriptive literature that have not been as patronage specifically accorded to Pāśupata developed in the mediaeval temple architecture of masters dropped from the inscriptional record, the Indian subcontinent. As Minoru Hara (1966: and Pāśupata practices appear to have been sub70) observes: The Cambodian transplantation of the Pāśupata seems to be more faithful to the original than that of southern India. This Cambodian faithfulness to a mainland original is parallel to the phenomena of Cambodian Sanskrit.

Such texts as the AVP and the Śivadharma/ Śivadharmottara, which mention some of the practices expounded in the PS, describe a fireplace-building (agnyāgāra, etc.) for the observances of the pāśupatavrata (especially bhasmasnāna), while original Pāśupata sources like the GK and the Saṃskārakārikā give us more detailed descriptions of rituals of initiation carried out in the presence of

07_CreSouthV1_8P_11Feb22.indd 256

79. Sanderson (2003–2004: 435) argues that there were two waves of Śaivism in Kambujadeśa—the first, Atimārga, in pre-Angkorian inscriptions of the 7th and 8th centuries, the second, Mantramārga, from 9th to 14th centuries. 80. K. 95 (v. XXVIII, Barth 1885: 355); K. 300, v. xx (IC, IV: 254); K. 418B (Finot 1904: 672–679, Cœdès 1929: 289–330); K. 692, v. liv (IC, I: 237); K. 933 (IC, IV: 47); K. 194 and K. 383 (Cœdès and Dupont 1943: 56–154); K. 806, v. ccxxxvi (IC, I: 73); K. 872, v. xiii (IC, V: 97). See also Melville Bolling and von Negelein 1909: xxx–xxxi, 179–93. Cf. also the testimony about Pāśupata Brahmins of the Chinese envoy Zhou Daguan, who spent a year at Angkor in 1296–1297 (see Chemburkar, this volume), and the ‘postBayon-phase’ reliefs at the Bayon (Provost-Roche 2010), which appear to depict Śaiva ascetics who may very well have been Pāśupatas.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples sumed as standard Khmer Śaiva religious traditions. The architecture of annex buildings did see many modifications during this period, again reflecting a changing Śaiva religious programme, including a movement away from ascetic practices towards a temple-based, king-sponsored worship and ritual involving homa. It is therefore possible that the usage of those buildings changed over time, and the architectural development of these structures during the successive centuries is a silent testimonial to the evolution of religious practices, which caused the adaptation of those buildings by the Khmer royalty.

Primary Sources Ajitāgama: Ajitāgama, Vol. I, ed. by N.R. Bhatt. Pondichéry: Institut français d’Indologie, 1964. Atharvavedapariśiṣṭas: The Pariśiṣṭas of the Atharvaveda, Vol. 1, Part 1 and 2, ed. by George Melville Bolling and Julius von Negelein. Leipzig: Otto Harrassowitz, 1909–1910. [cf. Bisschop and Griffiths 2003] Bhuvanakośa: palm-leaf manuscript (lontar) Leiden Cod. Or. 5022 (ca. 1878 CE), Balinese script. Brahmāṇḍa Purāṇa: Brahmāṇḍa-mahā-purāṇa. [Mumbai:] Śrī-Venkateśvara, 1857. Gaṇakārikā of Ācārya Bhāsarvajña with the Ratnaṭīkā and four appendices including the Kāravaṇa-Māhātmya, ed. by C.D. Dalal. Gaekwad’s Oriental Series no. 15. Baroda: Oriental Institute Baroda, 1966 [1920]. Mattavilāsa: Mattavilāsa Prahasana, ed. and tr. by Narayanan Parameswaran Unni. Delhi: Nag Publishers, 1998. Niśvāsaguhya: e-text of in-progress edition of the Guhyasūtra of the Niśvāsatattvasaṃhitā, prepared by Dominic Goodall, with the contribution of Diwakar Acharya, Peter Bisschop and Nirajan Kafle, from MS NAK 1-227, supplemented with readings from its two Devanāgarī apographs, MS 5-2401, NGMPP Reel No. A 159/18 and Sanskrit MS i.33 of the Wellcome Institute for the History of Medicine, London. Pāśupatasūtra: Pasupata Sutras with Pancarthabhashya of Kaundinya, ed. by R. Anantakrishna Sastri. Trivandrum: The Oriental Manuscript Library of the University of Travancore, 1940.

07_CreSouthV1_8P_11Feb22.indd 257

257

Pañcārthabhāṣya, see Pāśupatasūtra. Ratnaṭīkā, see Gaṇakārikā. Saṃskārakārikā, see Acharya 2007. Śatapathabrāhmaṇa: Madhyamdina recension, ed. by H.S. Ananthanarayana and W.P. Lehman. GRETIL – Göttingen Register of Electronic Texts in Indian Languages and related Indological materials from Central and Southeast Asia (http://fiindolo.sub.uni-goettingen.de/gret_utf. htmSatBr). Śivadharma: transcript IFP/EFEO 860, copied from R2442 G.O.M.L., Madras. Śivadharmottara: transcript IFP/EFEO 72, from page 25 copied from Adyar Library Madras ms. no. 75425. Somaśambhupaddhati, see Brunner-Lachaux 1963. Vāyu Purāṇa: Vāyupurāṇam. Poona: Anandasrama Sanskrit Series, 1860.

References

Acharya, Diwakar. 2007. ‘The Saṃskāravidhi: A Manual on the Transformatory Rite of the Lakulīśa Pāśupatas’, in Dominic Goodall and André Padoux (eds.), Mélanges tantriques à la mémoire d’Hélène Brunner, pp. 27–48. Pondicherry: Institut français de Pondichéry/École française d’Extrême-Orient. Acri, Andrea. 2014. ‘The Śaiva Atimārga in the Light of Niśvāsaguhya 12.1–22ab’, Cracow Indological Studies 16: 7–49. Agrawala, Om Prakash. 1984. Conservation of Manuscripts and Paintings of Southeast Asia. London: Butterworths. Arahi, Hsiao. 2000. ‘Étude de la chronologie relative de la “bibliothèque” sud du temple de Banteay Kdei’, in Barom Neth, Chiev Khus, and Henri Locard (eds.), The 2nd International Conference on Khmer Studies Proceedings. Phnom Penh: Ministry of Education, Youth and Sport and Royal University of Phnom Penh. Bakker, Hans. 2001. ‘Dakṣiṇāmūrti’, in Klaus Karttunen and Petteri Koskikallio (eds.), Vidyārṇavavandanam: Essays in Honour of Asko Parpola (Studia Orientalia 94), pp. 41–53. Helsinki: Finnish Oriental Society. . 2004. ‘At the Right Side of the Teacher’, in Phyllis Granoff and Koichi Shinohara (eds.), Imagination, Imagery, and Image in Vedic and Śaiva Initiation, Images in Asian Religions: Texts and Contexts, pp. 127–129. Toronto: UBC Press.

11/2/22 3:44 PM

258

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

. 2014. The World of the Skandapurāṇa: Northern India in the Sixth and Seventh Centuries. Leiden: Brill. Barth, Auguste. 1885. Inscriptions sanscrites de Cambodge (extraits des notices et extraits des manuscrits de la Bibliothèque nationale 27, 1ère partie, 1er fascicule), pp. 1–181. Paris: Imprimerie Nationale. Bergaigne, Abel. 1893. Inscriptions sanscrites du Campā et du Cambodge (extraits des notices et extraits des manuscrits de la Bibliothèque nationale 27, 2e partie, 2e fascicule), pp. 181–632. Paris: Imprimerie Nationale. Bhandarkar, D.R. 1931. ‘Mathurā Pillar Inscription of Chandragupta II’, Epigraphia Indica 21. Bhandarkar, R.G. 1913. Vaiṣṇavism, Śaivism and Minor Religious Systems. Grundriss der indo-arischen Philologie und Altertumskunde Bd. 3, Hft. 6. Strassburg: Trübner. Bhattacharya, Kamaleswar. 1955. ‘La secte des Pāçupata dans l’ancien Cambodge’, Journal Asiatique 243/2: 479–490. . 1961. Les Religions brahmaniques de l’ancien Cambodge, d’après l’épigraphie et l’iconographie. Paris: Publications de l’École française d’Extrême-Orient. . 1964. ‘Notes d’iconographie khmère’, Arts Asiatiques 10/1: 91–94. Bisschop, Peter. 2006. Early Śaivism and the Skandapurāṇa: Sects and Centres. Groningen Oriental Studies XXI. Groningen: Egbert Forsten. Bisschop, Peter and Arlo Griffiths. 2003. ‘The Pāśupata Observance (Atharvavedapariśiṣṭa 40)’, Indo-Iranian Journal 46: 315–348. Brunner-Lachaux, Hélène (trans.). 1963. Somaśambhupaddhati: Le rituel quotidien dans la tradition śivaïte de l’Inde du Sud selon Somaśambhu. Vol. 1 of 4 vols. Pondicherry: Institut français d’Indologie. Chemburkar, Swati and Shivani Kapoor. 2018. ‘Pāśupata Sect in Ancient Cambodia and Champa: Art Historical Evidence’, in Trần Kỳ Phương, Võ Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 45–57. Bangkok: River Books. Cœdès, George. 1911. ‘Études cambodgiennes. VI. Des édicules appelés « bibliothèques »’, Bul-

07_CreSouthV1_8P_11Feb22.indd 258

letin de l’École française d’Extrême-Orient 11: 405–406. . 1929. ‘Études cambodgiennes. XXIII. La date du temple de Bantāy Srëi’, Bulletin de l’École française d’Extrême-Orient 29: 289–330. . 1932. ‘Études cambodgiennes. XXX. À la recherché du Yaçodharāçrama’, Bulletin de l’École française d’Extrême-Orient 32/1: 84–112. . [IC] 1937–1966. Inscriptions du Cambodge, éditées et traduites, vols. I–VIII. Hanoi and Paris: École française d’Extrême-Orient. . 1940. ‘Études cambodgiennes. XXXV. Les gîtes d’étape à la fin du XIIe sìecle’, Bulletin de l’École française d’Extrême-Orient 40: 347–349. Cœdès, George and Pierre Dupont. 1943. ‘Les stèles de Sdok Kăk Thoṃ, Phnoṃ Sandak et Práḥ Vihãr’, Bulletin de l’École française d’ExtrêmeOrient 43: 56–134. Colas, Gérard. 1986. Le temple selon Marīci. Extraits de la Marīci-saṃhitā étudiés, édités et traduits. Pondichéry: Institut français de Pondichéry. Collins, Charles D. 1982. ‘Elephanta and the Ritual of the Lakulīśa-Pāśupatas’, Journal of the American Oriental Society 102/4 (October–December): 605–617. Cunin Olivier. 2004. De Ta Prohm au Bayon: Analyse comparative de I’histoire architecturale des principaux monuments Khmers du style du Bayon, Volume I. PhD dissertation, Institut National Polytechnique de Lorraine. Dagens, Bruno. 1977. Les enseignements architecturaux de l’Ajitāgama et du Rauravāgama. Pondichéry: Institut français d’Indologie. . 2003. Les Khmers. Paris: Société d’édition des belles lettres. Dagens, Bruno and Marie-Luce Barazer-Billoret. 2000. Le Rauravāgama: un traité de rituel et de doctrine śivaïte. Pondichéry: Institut français de Pondichéry. Datta, Bimal Kumar. 1960. Libraries and Librarianship of Ancient and Medieval India. New Delhi: Atma Ram & Sons. De Simini, Florinda. 2016. Of Gods and Books: Ritual Knowledge and Transmission in the Manuscript Cultures of Premodern India. Berlin/ Boston: De Gruyter. Dresden, Mark Jan. 1941. Mānavagṛhyasūtra: A Vedic Manual of Domestic Rites. Translation, Commentary and Preface. Groningen/Batavia: J.B. Wolter.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples Estève, Julia. 2009. Étude critique des phénomènes de syncrétisme religieux dans le Cambodge angkorien. PhD dissertation, École Pratique des Hautes Études. Ferstl, Christian. 2015. ‘Overlooked Material for the Study of Pāśupata Śaivism: The satirical play Mattavilāsaprahasana’. Pre-print draft dated 17 July 2015, to appear in Ernst Prets and Hiroshi Marui (eds.), Transmission and Tradition. Vienna: VÖAW. [Last accessed September 2021 at https://www.academia.edu/12332050/ Overlooked_Material_for_the_Study_of_Pāśupata_Śaivism_The_satirical_play_Mattavilāsaprahasana] Finot, Louis. 1904. ‘Notes d’épigraphie. VII . L’inscription de Práḥ Khan’, Bulletin de l’École française d’Extrême-Orient 4/3: 672–679. . 1912. ‘Notes d’épigraphie. XIII. L’inscription de Ban That’, Bulletin de l’École française d’Extrême-Orient 12/2: 1–28. . 1925. ‘Dharmaçâlâs au Cambodge’, Bulletin de l’École française d’Extrême-Orient 25/3–4: 417–422. . 1928. ‘Nouvelle Inscriptions de Sambor’, Bulletin de l’École française d’Extrême-Orient 28/1: 43–46. Foucher, Alfred C.A. 1903. ‘Nouvelles et Mélanges. Critique de E. Lunet de Lajonquière, 1902, Inventaire descriptif des monuments du Cambodge’, Journal Asiatique 10: 174–180. Gonda, Jan. 1969. ‘Āyatana’, Adyar Library Bulletin 23: 1–79. . 1970. Vaiṣṇavism and Śaivism: A Comparison. London: Athlone Press. Goodall, Dominic. 2015. ‘On K. 1049, A Tenth-Century Cave Inscription from Battambang and on the Sectarian Obedience of the Śaiva Ascetics of Non-Royal Cave Inscriptions of Cambodia’, UDAYA: Journal of Khmer Studies 13: 3–34. . 2017. ‘What Information can be Gleaned from Cambodian Inscriptions about Practices Relating to the Transmission of Sanskrit Literature?’, in Vincenzo Vergiani, Daniele Cuneo and Camillo Alessio Formigatti (eds.), Indic Manuscript Cultures through the Ages: Material Historical Investigation, pp. 131–160. Berlin/ Boston: De Gruyter.

07_CreSouthV1_8P_11Feb22.indd 259

259

Hara, Minoru. 1966. Materials for the Study of Pāśupata Śaivism. PhD dissertation, Harvard University. Hendrickson, Mitch. 2008a. ‘People Around the Houses with Fire: Archaeological Investigation of Settlement around the Jayavarman VII “Resthouse” Temples’, UDAYA: Journal of Khmer Studies 9: 63–79. . 2008b. Arteries of Empire: An Operational Study of Transport and Communication in Angkorian Southeast Asia (9th to 15th Centuries). PhD dissertation, University of Sydney. IC, see Cœdès, George. Jacques, Claude. 1968. ‘La stèle de Phnom Srès’, Bulletin de l’École française d’Extrême-Orient 54: 605–622. Jacques, Claude and Philippe Lafond. 2004. L’Empire Khmer: Cités et sanctuaires, Vème-XIIIème siècle. Paris: Fayard. . 2007. The Khmer Empire, Cities and Sanctuaries from the 5th to the 13th Century. Bangkok: River Books. Kafle, Nirajan. 2020. Niśvāsamukhatattvasaṃhitā: A Preface to the Earliest Surviving Śaiva Tantra (on Non-Tantric Śaivism at the Dawn of the Mantramārga); Critical Edition, with Introduction & Annotated Translation and an Appendix Containing Śivadharmasaṅgraha 5–9. Pondicherry: Institut français de Pondichéry/ École française d’Extrême-Orient/Asien-Afrika Institut, Universität Hamburg. Kapoor, Shivani. 2019. ‘Iconography of Pashupata Ascetics in Cambodia’, Orientations 50/1: 122–131. Kashikar, C.G. 1964. The Śrauta, Paitṛmedhika and Pariśeṣa sūtras of Bharadvāja. Critically edited and translated, part 2. Poona: Vaidika Saṃśodhana Maṇḍala. Lefèvre, Vincent. 2011. Portraiture in Early India: Between Transience and Eternity. Leiden: Brill. Lockwood, Michael, A. Vishnu Bhat, and Mahendra Vikrama Varma. 1995. Metatheater and Sanskrit Drama. Madras/New Delhi: Tambaram Research Associates/Munshiram Manoharlal Publishers. Lokesh Chandra. 1980. ‘Chandi Mendut and Pawon: A New Interpretation’, Bijdragen tot de Taal-, Land- en Volkenkunde 136/2–3: 313–320.

11/2/22 3:44 PM

260

Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

Lunet de Lajonquière, Étienne. 1902–1911. Inventaire descriptif des monuments du Cambodge (3 vols.). Paris: Ernest Leroux. Majumdar, R.C. 1953. Inscriptions of Kambuja. Calcutta: The Asiatic Society. Malleret, Louis. 1960. ‘Contribution à l’étude du thème des neuf divinités dans la sculpture du Cambodge et du Champa’, Arts Asiatiques 7/3: 205–230. Mankodi, Kirit. 1988. ‘Śaiva Panels of Ellora and Elephanta: Yogíśvara or Dakṣiṇāmūrti’, in Ratan Parimoo, Deepak Kannal, and Shivaji Panikkar (eds.), Ellora Caves, Sculptures and Architecture: Collected Papers of the University Grants Commission’s National Seminar, pp. 278–284. New Delhi: Books and Books. Marchal, Hénri. 1931. ‘Chronique’, Bulletin de l’École française d’Extrême-Orient 31: 563–650. Maxwell, Thomas S. 2007. ‘The Stele Inscription of Preah Khan, Angkor: Text with Translation and Commentary’, UDAYA: Journal of Khmer Studies 8: 1–85. . 2009. ‘A New Sanskrit and Khmer Inscription at Banteay Chhmar’, UDAYA: Journal of Khmer Studies 10: 162–163. Maxwell, Thomas S. and Jaroslav Poncar. 2006. Of Gods, Kings, and Men: The Reliefs of Angkor Wat. Chiang Mai: Silkworm. van der Meij, Th.C. 2017. Indonesian Manuscripts from the Islands of Java, Madura, Bali and Lombok. Leiden/Boston: Brill. Monier-Williams, Monier. 1899. Sanskrit-English Dictionary. Oxford: Clarendon Press. Multzer o’Naghten. 2014. ‘The Organisation of Space in Pre-modern Thailand under Jayavarman VII’, in Nicolas Revire and Stephen A. Murphy (eds.), Before Siam: Essays in Art and Archaeology, pp. 396–419. Bangkok: River Books/The Siam Society. Padoux, André et al. 2000. Tāntrikābhidhānakośa: dictionnaire des termes techniques de la littérature hindoue tantrique (A Dictionary of Technical Terms from Hindu Tantric Literature). Wien: Verlag der Österreichischen Akademie der Wissenschaften. Pathak, V.S. 1980. History of Śaiva Cults in Northern India from Inscriptions (700 A.D. to 1200 A.D.). Varanasi: Ram Naresh Sharma.

07_CreSouthV1_8P_11Feb22.indd 260

Payne, Richard K. 2011. ‘Ritual Studies in the Longue Durée: Comparing Shingon and Śaiva Siddhānta Homas’, Pacific World, Journal of the Institute of Buddhist Studies, Third Series 13 (Fall): 223–262. Pottier, Christophe. 2000. ‘À la recherché de Goloupura’, Bulletin de l’École française d’ExtrêmeOrient 87/1: 79–107. Provost-Roche, Ludivine. 2010. ‘Nouvelles perspectives pour la connaissance de l’histoire de la fin de la période angkorienne au Cambodge: notes sur quelques bas-reliefs de la galerie intérieure du Bayon’, Arts Asiatiques 65: 79–94. Sanderson, Alexis. 2001. ‘History through Textual Criticism in the Study of Śaivism, the Pañcarātra and the Buddhist Yoginītantras’, in François Grimal (ed.), Les Sources et le temps. Sources and Time: A Colloquium, Pondicherry, 11–13 January 1997, pp. 1–47. Collection Indologie 91. Pondicherry: Institut français de Pondichéry/ École française d’Extrême-Orient. . 2003–2004. ‘The Śaiva Religion Among the Khmers, Part I’, Bulletin de l’Ecole française d’Extrême-Orient 90–91: 349–463. . 2006. ‘The Lākulas: New Evidence of a System Intermediate between Pāñcārthika Pāśupatism and Āgamic Śaivism’, Ramalinga Reddy Memorial Lectures, 1997, The Indian Philosophical Annual 24: 143–217. Sankaranarayanan, S. 1993. ‘Sanskrit Education in Ancient Tamil Country’, The Adyar Library Bulletin 57: 7–33. Shimoda, Ichita and Sae Shimamoto. 2012. ‘Spatial and Chronological Sketch of the Ancient City of Sambor Prei Kuk’, Aséanie 30: 11–74. Sircar, D.C. 1942. ‘A Note on the Mathurā Inscription of Chandra Gupta II’, Indian Historical Quarterly 18/2: 271. Soutif, Dominique. 2009. Organisation religieuse et profane du temple khmer du VIIème au XIIIème siècle. Volume I: Les biens du dieu. PhD dissertation, Université Sorbonne Nouvelle–Paris III. . 2017. ‘Le feu sacré: étude archéologique et épigraphique d’un rituel du Cambodge ancien’, in Azedine Beschaouch, Franciscus Verellen, and Michel Zink (eds.), Deux décennies de coopération archéologique franco-cambodgienne à Angkor, pp. 9–25. Paris: Académie des Inscriptions et Belles-Lettres.

11/2/22 3:44 PM

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples Tagare, C.V. 1958. The Brahmāṇḍa Purāna. Part I. Translated and Annotated by G.V. Tagare. New Delhi: Motilal Banarsidass. Törzsök, Judith. 2020. ‘Why are the Skull-Bearers (Kāpālikas) called Soma?’, in Dominic Goodall, Shaman Hatley, Harunaga Isaacson, and Srilata Raman (eds.), Śaivism and the Tantric Tradition: Essays in Honour of Alexis G.J.S. Sanderson, pp. 33–46. Leiden/Boston: Brill.

07_CreSouthV1_8P_11Feb22.indd 261

261

Watson, Alex, Dominic Goodall and Anjaneya Sarma. 2013. An Enquiry into the Nature of Liberation; Bhaṭṭa Rāmakaṇṭha’s Paramokṣanirāsakārikāvṛtti, a commentary on Sadyojyotiḥ’s refutation of twenty conceptions of the liberated state (mokṣa), for the first time critically edited, translated into English and annotated. Pondicherry: Institut français de Pondichéry/ École française d’Extrême-Orient.

11/2/22 3:44 PM

07_CreSouthV1_8P_11Feb22.indd 262

Ascetic figures on the south wall of a southeast annex building. (Photo: Swati Chemburkar)

West view of the southeast annex buildings. (Photo: Swati Chemburkar)

Southeast view of the annex buildings. (Photo: Swati Chemburkar)

Annex Buildings at Bakong—Type A

Appendix 7.1

Photographic Database of Annex Buildings in Khmer Temples

262 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 263

West view of the southeast annex building. (Photo: Swati Chemburkar)

Navagraha panel at skirting level of the east wall of the southeast annex building. (Photo: Swati Chemburkar)

Chimney-like interior of the southeast annex building. (Photo: Swati Chemburkar)

Annex Building at Preah Ko—Type A

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 263

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 264

Southwest view of the southeast annex building. (Photo: Swati Chemburkar)

Possible Navagraha panel at a skirting level of the east wall of annex building. (Photo: Shivani Kapoor)

Annex Buildings at Prasat Khna (r)—Type A

264 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

Phnom Krom: four annex buildings are seen from the eastern entrance (two situated in the northeast and two in the southeast quarter of the temple). (Photo: Swati Chemburkar)

07_CreSouthV1_8P_11Feb22.indd 265

East view of the southeast annex buildings: one in stone and one in brick. (Photo: Swati Chemburkar)

Annex Buildings at Phnom Krom—Type A

West view of the southeast annex buildings in brick. (Photo: Swati Chemburkar)

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 265

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 266

North view of the four annex buildings as seen on the left side: two in stone and two in brick (the brick buildings are in ruins). (Photo: Swati Chemburkar)

West view of the northeast and southeast annex buildings. (Photo: Swati Chemburkar)

Annex Buildings at Phnom Bok—Type A

266 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 267

Two annex building are seen from the east side. The brick one is on the left (southeast) and the laterite one without ventilation system is on the right side (northeast). (Photo: Swati Chemburkar)

Annex Buildings at Prasat Pram—Type A

East view of the southeast annex building in brick. (Photo: Swati Chemburkar)

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 267

11/2/22 3:45 PM

Northeast annex building. (Photo: Swati Chemburkar)

07_CreSouthV1_8P_11Feb22.indd 268

South view of the northeast annex building. (Photo: Chimney-like interior of the northeast annex Swati Chemburkar) building. (Photo: Swati Chemburkar)

Annex Buildings at Preah Rup—Type A

268 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

West view of the southeast annex building. (Photo: Swati Chemburkar)

07_CreSouthV1_8P_11Feb22.indd 269

South view of the southeast annex building. (Photo: Swati Chemburkar)

Annex Buildings at Ta Keo—Type B

Interior view of the southeast annex building. (Photo: Swati Chemburkar)

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 269

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 270

West view of the northeast annex building. (Photo: Swati Chemburkar)

Southeast annex building with false door on the western side. (Photo: Swati Chemburkar)

West view of the southeast annex building. (Photo: Swati Chemburkar)

South view of the southeast annex building. (Photo: Swati Chemburkar)

Annex Buildings at Banteay Srei—Type B

270 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 271

Southeast view of the northeast annex building. (Photo: Swati Chemburkar)

Northwest view of the southeast annex building. (Photo: Shivani Kapoor)

Annex Buildings at the Banteay Srei style temple at the east of North Khleang in Angkor Thom—Type B

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 271

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 272

Southwest view of the northeast annex building. (Photo: Olivier Cunin)

Northwest view of the southeast annex building. (Photo: Olivier Cunin)

Annex Buildings at Chau Srei Vibol—Type B

272 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 273

Phimai: northeast view of the west annex building, variant of Type B. (Photo: Olivier Cunin)

Phnom Rung: west view of the southeast annex building, Type B. (Photo: Olivier Cunin)

Annex Buildings at Phimai and Phnom Rung—Variant of Type B and Type B

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 273

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 274

Thommanon: northwest view of the southeast annex building. (Photo: Olivier Cunin)

Banteay Samre: west view of the southeast annex building. (Photo: Olivier Cunin)

Annex Buildings at Thommanon and Banteay Samre—Type B

274 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:45 PM

07_CreSouthV1_8P_11Feb22.indd 275

South view of the southwest annex building in the inner enclosure. (Photo: Swati Chemburkar)

Southwest annex building in the third enclosure. (Photo: Swati Chemburkar)

North view of the southwest annex building in the second enclosure. (Photo: Swati Chemburkar)

Northwest annex building in the third enclosure. (Photo: Swati Chemburkar)

Annex Buildings at Angkor Wat—Type C

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 275

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 276

(TP 3) Northwest view of the southeast annex building, Type B. (Photo: Olivier Cunin)

(TP 43) Northeast view of the southeast annex building, variant of Type B. (Photo: Swati Chemburkar)

Annex Buildings at Ta Prohm of Angkor—Type B and Variant Type B

(TP 43) Southeast annex building interior. (Photo: Swati Chemburkar)

276 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 277

(BK 7) West view of the southeast annex building of Banteay Kdei, variant of Type B. (Photo: Olivier Cunin)

(PK 5) Northwest view of the southeast annex building of Preah Khan of Angkor, Type B. (Photo: Olivier Cunin)

(BK 13) West view of the northeast annex building of Banteay Kdei, variant of Type B. (Photo: Olivier Cunin)

Annex buildings at Preah Khan of Angkor and Banteay Kdei—Type B and variant of Type B

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 277

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 278

(BC 15) Banteay Chhmar: southwest view of variant of Type B annex building. (Photo: Olivier Cunin)

(BC 64) Banteay Chhmar: northwest view of variant of Type B annex building. (Photo: Olivier Cunin)

(BC 45) Banteay Chhmar: northeast view of Type A annex building. (Photo: Olivier Cunin)

(BC 88) Banteay Chhmar: south view of Type C annex building. (Photo: Olivier Cunin)

Annex Buildings at Banteay Chhmar—Type A, variant of Type B and Type C

278 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

West view of the southeast annex building. (Photo: Olivier Cunin)

07_CreSouthV1_8P_11Feb22.indd 279

West view of the northeast annex building. (Photo: Olivier Cunin)

Annex Buildings at Wat Nokor—Type B

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 279

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 280

(BY54) South view of the northeast annex building in the outer enclosure. (Photo: Swati Chemburkar)

(BY53) West view of the southeast annex building in the outer enclosure: Type C. (Photo: Swati Chemburkar)

(BY50) and (BY51): Variant of Type B annex buildings. (Photo: Olivier Cunin)

(BY16-left) and (BY21-right): Type B annex buildings in front of the central sanctuary. (3D reconstruction: Olivier Cunin)

Annex Buildings at the Bayon—Type B and Type C

280 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 281

‘Resthouse buildings’: top left: Ta Prohm; top right: Preah Khan of Angkor; bottom left: Preah Khan of Kompong Svay; bottom right: Banteay Chhmar. (Photos: Olivier Cunin)

Gîte d’étapes, Dharmaśālā or Resthouse

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 281

11/2/22 3:46 PM

Jayavarman II

Jayavarman III

Indravarman I

790–835

835–877

877–c. 886

Īśānavarman

Reining King

800

Date

616–628

Year

700

  Province

07_CreSouthV1_8P_11Feb22.indd 282

315,00

Preah Vihear

Prasat Khna(r)

584,00

 

Rolous Group, Siem 585,00 Reap

Rolous group, Siem Reap

 

164,00 (following Bruguier)

IK No.

Preah Ko and Bakong (stage II)

Bakong (stage I)

 

Sambor Prei Kompong Thom Kuk, Temple N10

Temple

One (Type A)

One (Type A)

Two (Type A) with ventilation system in the southeast, two other annex buildings with same design in the northeast, one in the southwest, and one in the northwest but without any ventilation system.

 

An initial open pillared structure underwent four stages of construction showing strong similarities to later annex buildings (Shimoda and Shimamoto 2012).

No. of Annex Buildings/Type

Appendix 7.2 Brief Survey of the Annex Buildings in Khmer Temples1

Navagraha frieze, probably placed at skirting level of the wall facing the entrance of the southeastern building.

Inscription

K. 355 on the doorjamb of the south annex building, calling this building a pustakāśrama.

 

 

Navagraha frieze (see EFEO CAM19417-3DCA no. 4363) and Auṃ metal symbol (EFEO CAM250814- DCA no. 7313).

In situ frieze of nine planets at the bottom of the wall facing the main entrance of the library building in the southeast.

 

 

Navagraha frieze   from the Sambor Prei Kuk temple site (exact findspot unknown; published by Malleret 1960: 207, Fig. 1).

Archaeological Finds

282 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 283

Harṣavarman I

Jayavarman IV

c. 915–923

c. 928–941

900

 

Yaśovarman I

Reining King

 

889–c. 915

Date

 

Year Siem Reap

Province

Siem Reap

Prasat Phnom Bok

Prasat Pram

Work on Bayang

Koh Ker Group, Preah Vihear

 

 

Siem Reap

Phnom Krom

Work on Bayang

Siem Reap

Bakheng

Additions to Ta Keo pre-Angkorian temple of Phnom Bayang

Prasat Trapeang Srange or Dei Dom

Temple

265,00

 

 

547,00

501,00

496,00

3,00

588,00

IK No.  

A very ruined Navagraha frieze on the back of the pavilion, at the left of the esplanade (Malleret 1960: 206).

Archaeological Finds

 

 

Two (Type A):   southern in brick with ventilation system, northern in laterite without ventilation system and central pedestal.

 

 

Four: two in stone and   two in brick (Type A)

Four: two in stone and Navagraha two in brick (Type A) frieze from the northwestern library (CA: 4715 and EFEO CAM 00878).

Two (Type A)

At least one (probable   addition in the 10th century to the preAngkorian temple: see image EFEO CAM 00811)

Two (in ruins)

No. of Annex Buildings/Type

 

 

 

 

 

 

 

K. 937 on the left door pillar of the south annex building, calling it a fire shrine.

Inscription

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 283

11/2/22 3:46 PM

Year

Reining King

Rājendravarman

Date

944–968

Siem Reap

Siem Reap

Preah Rup

Province

East Mebon or Mebon Oriental

Temple

07_CreSouthV1_8P_11Feb22.indd 284

538,00

531,00

IK No.

Two (Type A)

Two (Type A) in the southeastern corner with possibly some ventilation system and three others with similar design but without ventilation system.

No. of Annex Buildings/Type

A Navagraha frieze   from the northwestern building (EFEO CAM 13083) and a seven ascetic-frieze from the southeastern building (EFEO CAM 13082).

A Navagraha frieze   placed at the skirting level of the wall facing the eastern entrance (EFEO CAM 19416.3). A 15 cm wide pedestal from the far end of the southeastern building, which possibly housed a Navagraha slab. A slab with nine holes from the second southeastern building, which would have possibly housed nine individual planets (in situ). A row of seven ascetic-frieze from the northwestern building (EFEO CAM 19417.1), currently kept at the Depot DCA no. 3506-a-b (Malleret 1960: 209).

Archaeological Finds Inscription

284 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

968–1000

 

 

 

 

Date

 

Year

07_CreSouthV1_8P_11Feb22.indd 285

 

 

Jayavarman V

Reining King

Banteay Srei style temple at the east of North Khleang

Banteay Srei

Ta keo

Temple

Angkor Thom, Siem Reap

Siem Reap

Siem Reap

Province

480,06

546,02

533,00

IK No.

Two (Type B: see image EFEO CAM 02138)

Two (Type B)

Two (Type B)

No. of Annex Buildings/Type  

Navagraha frieze from the northwestern building (EFEO CAM 13083); and seven ascetic-frieze from the southeastern building (EFEO CAM 13082).

 

  Three Gaṇeśas, one small liṅga, one kneeling male figure and one male figure on Garuda from the southern building (for identification, see Multzer o’Naghten 2014; CA: 3157, MNC: CA_02816 and EFEO CAM 0736)1. One circular object with central pedestal (CA: 3216; see reconstruction images by Cunin).

Pedestal with nine mortises (DCA.6788/N.179) from the southeastern library at the second level, according to the report of Angkor Conservation.

Archaeological Finds Inscription

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 285

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 286

Jayavīravarman

Sūryavarman I

Udayādityavarman II

Harṣavarman III

1002–1010

1002–1049

1050–1066

1066–1080

Ta Keo

Phnom Chisor

Phnom Rung Angkor Wat

 

 

1113–c.1149 Sūryavarman II

Work on Bayang

Dharaṇīndravarman I

1107–1112

1100

Banteay Samre

Phimai

 

 

 

Preah Vihear

Preah Vihear

Siem Reap

Buriram, Thailand

 

Siem Reap

Nakhon Ratchasima, Thailand

 

 

Siem Reap

Prasat Kuk Trapeang Ropou

 

Siem Reap

 

 

Province

Chau Srei Vibol

Continuation of Ta Keo

 

Temple

 

1080–c.1107 Jayavarman VI

Udayādityavarman I

1001–1002

1000

Reining King

Date

Year

497,00

596,00

 

541,00

447,00

 

 

23,00

398,00

518,00

564,00

 

 

IK No.

Six (Type C: two in the third enclosure, two in the second enclosure, and two in the first enclosure).

One (Type B) in the southeast

 

Two (Type B)

One (variant of Type B)

 

 

One (Type B: see images EFEO CAM 15908.07 and CAM 15909.19)

Two (Type B)

One (In ruin; see image EFEO CAM 1014)

Two (Type B)

 

No. of Annex Buildings/Type

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Inscription

K. 691 on the door pillar of the southern annex building, mentioning it as a fire shrine.

 

 

 

 

 

 

 

 

Archaeological Finds

286 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

Year

07_CreSouthV1_8P_11Feb22.indd 287

1181– c. 1220

c. 1149– 1180

Date

Jayavarman VII

Siem Reap

Siem Reap

Siem Reap

Ta prohm

Preah Khan of Angkor

Banteay Kdei

Siem Reap

Beng Melea

535,00

522,00

534,00

214,00

490,00

Siem Reap

IK No.

Thommanon

Province 489,00

Temple

Tribhuvanāditya- Chau Say Tevoda Siem Reap varman

Reining King

Two: BK7, BK13 enclosure I, variant of Type B

Eight (all Type B): PK3, PK35, PK5 (enclosure I), PK64, PK67 (enclosure III), PK75, PK98 and PK122 (in the south, west and north annex complexes).

Two: TP3 (enclosure I, Type B), TP43 (enclosure II, variant of Type B)

Four (two are Type B, placed in enclosure I, and two belong to Type C, placed in enclosure III; see image EFEO CAM 10964).

One (Type B)

Two (Type B)

No. of Annex Buildings/Type

 

  Navagraha frieze from enclosure III, northern cloister, southeastern courtyard (published by Bhattacharya 1957; kept at the Depot, no. 7501; old photo: EFEO CAM 13834).  

 

 

 

 

 

 

 

 

Archaeological Finds Inscription

Reinterpreting the Function of ‘Annex Buildings’ in Khmer Śaiva Temples 287

11/2/22 3:46 PM

07_CreSouthV1_8P_11Feb22.indd 288

Date

Reining King Siem Reap

Kompong Cham Siem Reap

Wat Nokor Bayon

Province

Banteay Chhmar

Temple

IK No.

474,00

85,00

816,00

1.  Timeline adopted after a lecture delivered by Claude Jacques at SOAS in May 2016.

Year

Six: BY50, BY51, BY16, BY21 enclosure I, variant of Type B); BY53, BY54, enclosure III, Type C.

2 (Type B)

Eleven: BC15, BC20, BC64 (enclosure I, variant of Type B), BC45 (enclosure II, Type A), BC81, BC88 (enclosure III, Type C), BC103, BC92SW (north and south annex complexes, Type B), BC125, BC126, BC127 (west annex complex, variant of Type B).

No. of Annex Buildings/Type Inscription

 

N1/17 from BC45 and N2/N18 from BC15, mentioning these two as fire shrines.

At least two   Navagraha friezes. One was found in 1937, while clearing the north side courtyard, and is presently kept at the Conservation depot, no. 11 (see EFEO CAM 02177). The second can still be seen in the Northwest Commaile heap in situ.

 

According to NMC records, a male figure on Garuḍa (see Multzer o’Naghten 2014 for a possible identification), a four-armed seated Avalokiteśvara, and a Yama on a buffalo from BC64 of the west group of the second enclosure. Presently at NMCKa. 1720 (B.362, B.73.2); Ka. 504 (B.162, B.115.1) and Ka. 86 (B.363, B.731.4).

Archaeological Finds

288 Shivani Kapoor, Swati Chemburkar, Andrea Acri, and Olivier Cunin

11/2/22 3:46 PM

Chapter 8

Śaiva Religious Iconography: Dancing Śiva in Multi-Polity Medieval Campā1 M ai Bùi Diệu Linh

A

s early as the 5th century CE, Śaivism had risen to eminence on the Indian subcontinent as an established religious institution. Prosperous maritime trade routes between India and China brought Śaivism to the new lands of Southeast Asia, including Campā, a chain of small coastal polities that developed during the first millennium CE in present-day central and southern Vietnam. The name ‘Campapura’ first appeared in a local Cam inscription, the Mỹ Sơn Stele of Prakāśadharma (C. 96), which is dated to 658 CE.2 A few years later, in the Khmer Kdei Ang inscription (K. 53), the name ‘Campeśvara’ is mentioned as well (Cœdès and Parmentier 1923: 46; Schweyer 2010). But Campā did not exist throughout its history under the same name, as it was referred to in several Chinese sources under different names such as Linyi 林邑 (Viet. Lâm Ấp) from 192 to 758, Huanwangguo 環王國 (Viet. Hoàn Vương quốc) from 758 to 886 and Zhancheng 占城 (Viet. Chiêm Thành) from 886–1471 (Vickery 2011: 369). In addition, if early French scholarship often described Campā as a unified kingdom, the last few decades have seen a wave of new studies concerning Campā not as a single entity, but as a multi-polity Cam ‘network state’ comprised of five regions or principalities located from North to South in present-day Vietnam: Indrapura, Amarāvatī, Vijaya, Kauthara, and Pāṇḍuraṅga.3 Keith Taylor (1999: 153) defined 1. I thank professors Leslie C. Orr and Trần Kỳ Phương for their valuable comments and suggestions. 2. Here and below I refer to numbers of the EFEO inventory of Campā inscriptions; for this particular stele, see Cœdès and Parmentier 1923: 22. 3. See, for instance, Po Dharma 2001. However, Hardy,

08_CreSouthV1_5P_28Jan22.indd 289

Campā as ‘a generic term for the polities organized by Austronesian-speaking peoples along what is now the central coast of Vietnam’, while Vickery (2011: 378) asserts that there was no single kingdom named ‘Campā’, and that the regions as distinguished in epigraphy corresponded to distinct geographical, and even rival, polities.4 This prompts the question as to what kind of political relationships and connections there existed between different parts/regions of Campā, and whether these regions can be conceptualized as polities, states, or principalities. There are of course differences between these terms but for the time being, as my chapter focuses more on artistic rather than political aspects, I will use these terms interchangeably. And for the sake of convenience, I will use the term Campā to include all Cam polities. Modern reconstructions of the early history and religion of Campā rely mostly upon stone inscriptions, archeological monuments, religious sculptures, and a few texts from neighbouring countries such as the account found in the Chinese Sui shu 隋書 (Book of the Sui [Dynasty]), Song Huiyao Jigao 宋會要積搞 (Draft of the Collected Statues Griffiths, and Wade (2019: 14) recently suggested that this model of Campā as a spatial organization of five territories is also obsolete and provided instead a new interpretation that emphasizes the centrality of river valleys in Campā’s political geography. 4. Interestingly, based on epigraphical texts and sculptural steles asociated with the name Śrī Viṣṇujāti Vīrabhadravarmadeva dating between 1401/2 and 1443/4, Griffiths (2019: 193) refutes revisionist claims that Campā was never a unified kingdom and instead suggests that in the 15th century, Campā was a land of significant territorial extent ruled by a powerful king.

28/1/22 7:44 PM

290

Mai Bùi Diệu Linh

of the Song [Dynasty]), and the Da Ming shilu

大明實錄 (Veritable Records of the Great Ming

[Dynasty]).5 A survey of extant Cam inscriptions points to the primacy of Śiva worship at the royal courts of mediaeval Campā.6 The earliest Cam indication of Śiva worship is the 5th century CE Mỹ Sơn Stele Inscription of Bhadravarman I (C. 72), which documented land endowments that were dedicated to the benevolent Śiva (Bhadreśvara) by the king.7 This inscription reveals that by this time Śaivism was a mature, well-established practice in Campā, and that it received royal and elite patronage. Examining the Pāśupata elements in ancient Cambodia and Campā, Chemburkar and Kapoor (2018: 47) argue that Bhadreśvara was also the state god of the Khmer at Sambor Prei Kuk and the Cam at Mỹ Sơn. As for religious iconography in Campā, depictions of Śiva appeared in both anthropomorphic form and aniconic symbolic form (śivaliṅga). This chapter focuses only on the artistic representations of one of the anthropomorphic forms of Śiva, the nṛttamūrti or dancer. The reasons why I am interested in this specific form of Śiva are first because of the lingering effect of A.K. Coomaraswamy’s popularization of South Indian cult of Naṭarāja (lit. Dancing Lord), and second because Naṭarāja was promoted and deployed by Cōḻa queen Sembiyan Mahādevī to become an international Cōḻa emblem from the 10th century onward. The role played by religious iconography in the establishment of political authority, such as in the case of Pallava8 or Cōḻa dynasties, inspired me to consider questions of how popular the dancing Śiva was in Campā, what were the political purposes behind adopting this representation, and what were the sources of transmission and diffusion of this embodiment of Śiva. It should be stressed that not

5. For a thorough study of these three Chinese accounts and their mention of Campā, see Wade 2003, 2005; Shiro 2011; and Aspell 2013. 6. According to Griffiths et al. (2012), the total number of inscriptions that had been found prior to 2010 was 233. They were written in Sanskrit and Old Cam, and dated mainly from the 7th to 15th centuries. 7. For a full translation of this inscription, see Majumdar (1985, Book III: 4) and Golzio (2004: 2). 8. See Gillet 2010.

08_CreSouthV1_5P_28Jan22.indd 290

all the images of dancing Śiva are Naṭarāja and that Naṭarāja was not the first sculptural form to illustrate Śiva dancing. Kaimal (1999: 392–396) argues that Naṭarāja is the representation of Śiva in the ānanda-tāṇḍava (‘dance of furious bliss’) in which Śiva bends his right leg and lifts his left leg to waist height and crosses the hips, while one of his four arms crosses the body with its relaxed ‘elephant hand’ gesture (gajahasta-mudrā) above the left ankle, and one of his hands makes the ‘open-palm gesture’ (abhaya-mudrā). He has long matted tresses spread out from his head and among the locks is the personification of the goddess of Ganges river with her hands in a respectful gesture (añjali-mudrā). He also displays a crescent moon, a flower and a topknot, and he stands on a trampled dwarf. Kaimal suggests that early (i.e., pre-10th century) dancing forms of Śiva usually had no lifted left leg crossing in front of the hips, and no drum or flames in their hands. She maintains that the Indian form of Naṭarāja is unique and emerges in both text and iconography from the beginning of the 10th century. In Campā we do not see this form of South Indian Naṭarāja discussed by Kaimal, and for this reason I will use the term ‘dancing Śiva’ instead of ‘Naṭarāja’ hereafter. The problem in undertaking this research is that there are no written records that mention how people worshipped the dancing Śiva or how exactly the Cam responded to this dancing figure.9 The treatment of these stone sculptures in their original temple locations remains unknown, as does the matter of how important this form of Śiva was in Campā. However, what we do know is that Cam dancing Śivas were commonly carved in stone and sculpted in the tympana placed above a temple’s main entrance door.10 The image of the deity that occupies this position usually indicates the main deity of the temple, or that the temple ‘belongs’ to

9. While the bhakti movement became popular among the mass population in South India from the 8th century, we have no evidence of devotional worship following a kind of personal love relationship to particular deities in Campā. 10. See, for example, the dancing Śiva at the main entrance of the temple Po Klaong Girai (dated 12th–13th centuries) in Pāṇḍuraṅga (present-day Ninh Thuận) in Figs. 8.11–8.12.

28/1/22 7:44 PM

Śaiva Religious Iconography that particular deity.11 The fact that Cam dancing Śiva were situated in this architectural position demonstrates the significance of this manifestation of Śiva. In India, dancing Śiva is carved in a circular niche set in the lower part of the tower, in the façade, and above the main entrances of temples, as for instance at Pattadakal (dated 7th–8th centuries) and at Bhubaneswar (dated 7th–8th centuries).12 Modern scholars working on Campā have yet to explore this dancing Śiva in much detail, either geographically or temporally. Prominent scholars of Cam art have studied sculptures individually rather than comparatively. In this chapter, I present a comparative approach by examining ten Cam dancing Śiva sculptures found in the four main polities of Campā, viz. Indrapura, Amarāvatī, Vijaya, and Pāṇḍuraṅga, and dated between the 8th and 13th centuries.13 I argue that by examining all the extant statues of dancing Śiva in chronological order we can get a broader picture of the transmission of this icon and its influence on the religious art of this region, specifically the stylistic changes over time and, more generally, the degree to which Cam art was influenced by Indian and Khmer art. Importantly, as the following discussion will reveal, Indian iconography influenced Cam depictions of dancing Śiva in northern Campā, whereas Khmer artistic style left its imprint on dancing Śiva from southern Campā. Visual depictions of dancing Śiva from northern Campā (Indrapura, Amarāvatī) often represent the deity with family members and attendants. By contrast, Khmer-influenced dancing Śiva from southern Campā (Vijaya and Pāṇḍuraṅga) depict the Hindu god dancing alone but with attributes. Interestingly, dancing Śiva from the early period (dated 8th to 10th centuries) 11. Another way to find out who is the main deity in a temple that enshrines multiple deities is through examining the dvārapālas, ‘the guardians of the door’. 12. See figs. 161–169 of Sivaramamurti 1974: 292 –295. 13. No sculptures of dancing Śiva were found in the Kauthara area, where the renowned temple Po Nagar devoted to the worship of the mother goddess Yā Pu Nagara is located. Schweyer (2004: 109–144) maintains that the cult of this female deity at Po Nagar in Nha Trang parallels the cult of the male deity Bhadreśvara at Mỹ Sơn; Trần and Nakamura (2008) share the same opinion. For a comprehensive and thorough study of this Mother goddess and Po Nagar temple, see Noseworthy 2015 and Nguyễn Thế Anh 1995.

08_CreSouthV1_5P_28Jan22.indd 291

291

were from northern Campā, whereas sculptures from the later period (dated 11th to 13th centuries) were predominantly from the South. For the dating of sculptures, I rely on the scholarship of French pioneers H. Parmentier (1909, 1918), P. Stern (1942), and J. Boisselier (1963), all affiliated with the École française d’Extrême-Orient (EFEO). It was these scholars who first established the practice of naming the style of the sculptures after the names of monuments or places where they were discovered.

dancing śiva from northern campā (indrapura, amarāvatī) Here I examine northern dancing Śivas from the early period (dated 8th to 10th centuries), which show strong connections with Indian iconographical elements, particularly through the representation of Śiva’s family (Pārvatī and Skanda), attendants (dwarf Apasmāra and Nandin), Śiva’s dancing poses (chatura or lalita with both knees bent), and multiple arms with hand gestures (abhaya-mudrā, gajahasta-mudrā). These elements are the most popular accompanying aspects in Indian depictions of dancing Śiva. The first three Cam dancing Śiva sculptures under analysis are from the 8th century sites of Mỹ Sơn A1 and Mỹ Sơn C1, Quảng Nam province, and the 9th century site of Bích La, Quảng Trị province (Figs. 8.1, 8.2, and 8.3).14 I group these three sculptures together because they share many similar characteristics. Henri Parmentier was one of the first scholars who documented these three sculptures of dancing Śiva. In his assessment, Śiva from Mỹ Sơn A1 (Fig. 8.1) has twelve arms. 14. Mỹ Sơn sanctuary is one of the earliest and most important religious centres of Campā. The second one in importance is the Po Nagar at Nha Trang as mentioned in the previous footnote. There were many temples built successively from the 5th to 14th centuries in the Mỹ Sơn sanctuary complex. For this reason, there are different name codes and dates for different temples within the sanctuary complex. Even though Campā was not a unitary kingdom, but rather a polity organized as a federation or confederation (Po Dharma 2001), Mỹ Sơn continued to hold its prominent position throughout the history of Campā.

28/1/22 7:44 PM

292

Mai Bùi Diệu Linh

Fig. 8.1: Dancing Śiva. Mỹ Sơn A1. (Drawing: Parmentier 1909: 361)

Fig. 8.2: Dancing Śiva. Mỹ Sơn C1. Mỹ Sơn Sanctuary. (Photo: author)

Considerable damage to the sculpture, though, makes it impossible to precisely determine Śiva’s attributes. I do not know where this sculpture is currently located and have to depend solely on Parmentier’s drawing.15 Parmentier did not identify the figure under Śiva’s feet but stated it was a man who appeared to be suffering. He also mentioned that to Śiva’s right was a kneeling woman praying with a standing naked child covered with jewellery and with arms crossed (Parmentier 1909: 366). I suggest that the figure under Śiva’s foot is the dwarf Apasmāra, the woman is Śiva’s consort Pārvatī, and the child is their son Skanda. What is of most interest in this tympanum is that the dwarf’s lower body and his left leg are supporting a pedestal carrying Pārvatī and Skanda. This gives a sense of the function of the dwarf as a vāhana, a vehicle or a chariot that carries Śiva’s entire family. Parmentier (1909: 360–361) identified the dancing figure on the left of Śiva as Bhṛṅgin; next to him is Gaṇeśa with his trunk raised toward Śiva; on each side and furthest from Śiva is a standing warrior with hands joined in devotional position.16

Sharing similar characteristics with the Mỹ Sơn dancing Śiva A1 is the Śiva Mỹ Sơn C1 (Fig. 8.2), which is currently preserved in Mỹ Sơn Sanctuary. Parmentier states that this Śiva C1 is a larger version of Śiva A1 and it could be the model that inspired Śiva A1. He also suggests that this dancing Śiva seems to have ten arms and the missing hands were probably holding some attributes. Similarly, as with Śiva A1, he did not identify the sitting woman and the naked child but suggested that the standing worshipper in a rich dress who has hands joined together might represent a founder king. On the right side of Śiva, Parmentier identified two musicians and a skinny character who, according to him, is Bhṛṅgin.17 Lastly, he mentions that on top and around the god, in the clouds, there are two or three apsarases holding lotus buds (Parmentier 1909: 390–392). It is unfortunate that since the upper part of this tympanum is badly mutilated it is almost impossible to identify the flying celestial beings. However, upon examining this Mỹ Sơn C1 dancing Śiva, Thierry Zéphir (2005: 188) contends that at the left of Śiva, on the upper part of the tympanum, is depicted the god Sūrya who is holding a lotus. The lotus is one of the attributes of Sūrya, and this probably led Zéphir to his suggestion. In

15. Sivaramamurti (1974: 360) also discussed this Cam dancing Śiva sculpture and provided a photograph of the tympanum, however he did not mention if he took the photo himself and where the sculpture was located when the picture was taken. 16. Discussing this sculpture, Sivaramamurti (ibid.) states that the figure to the extreme left of Śiva is probably Indra.

08_CreSouthV1_5P_28Jan22.indd 292

17. According to Sivaramamurti (ibid.: 361), the two musicians are Vaiśravaṇa playing the ūrdhva drum and Indra seated under a tree playing the flute.

28/1/22 7:44 PM

Śaiva Religious Iconography addition, examining this sculpture, Sivaramamurti (1974: 361) asserts that this image closely resembles Śiva dancing in chatura from Badami. The third sculpture in this group was found further north in Bích La of Quảng Trị province, where a dancing Śiva was also sculpted within a tympanum (Fig. 8.3). Śiva here has ten arms but the eight back arms were not very carefully carved, and are perhaps unfinished work. The hand gestures of the back arms are worn out and Parmentier (1909: 532–533) suggested that some possible attributes might have been badly damaged by dampness. The main front arms are in the crossed position that is very typical of Indian dancing Śiva iconography. The figure on the left of Śiva could be Skanda, one of the two sons of Śiva, while the figure on his right is a worshipper. Śiva is wearing a long garment, a loose skirt that is quite different from shorter Indian garments that usually come with decorative jewellery (see Fig. 8.6). This long garment shows that even though the sculpture is clearly influenced by Indian religious iconography, it is still a product of Southeast Asian sculptors. All of the three statues of dancing Śiva from Mỹ Sơn A1 and C1, as well as the Bích La from Quảng Trị appear to be in the chatura pose, with the right foot placed on the demon Apasmāra or on a pedestal and the left foot resting on the toes. Indian depictions of Śiva dancing in chatura pose appeared as early as the 6th century CE in the art of the Vākāṭakas (250–500 CE, contemporaries of the Guptas but in the South), which can be seen in various caves at Ellora.18 Moreover, either Skanda or Gaṇeśa appeared in all the sculptures. If the dancing Śiva from Mỹ Sơn A1 and Bích La have their arms crossed in exactly the same position, what Mỹ Sơn A1 and C1 have in common is the presence of Śiva’s consort Pārvatī, as well as Śiva’s attendant Bhṛṅgin.19 It is unclear if Bhṛṅgin in Mỹ Sơn A1 was holding a ḍamaru, but in Mỹ Sơn 18. See for instance figs. 11 and 70 of Śiva dancing in chatura Ellora cave in Sivaramamurti 1974: 174, 219. Early Cōḻa depictions of Śiva in chatura pose can be seen in the 10th-century sculptures originating from Tiruvarangulam currently on exhibit at the National Museum in New Delhi, India. 19. For the myths on Bhṛṅgin dancing around Śiva, see Rao (1916 [1993], vol. 2: 138, 227, 323).

08_CreSouthV1_5P_28Jan22.indd 293

293

Fig. 8.3: Dancing Śiva. Bích La, Quảng Trị. (Photo: Trần Kỳ Phương)

C1 he is holding one. The ḍamaru is one of Śiva’s attributes which appeared quite early in Indian iconography. For example, a Gupta sculpture of a multi-armed Śiva dancing with one of his hands holding a ḍamaru can be seen in Sirpur, Chhattisgarh.20 An emphasis on music is clearly shown in Mỹ Sơn C1 where musicians playing the flute and tambour are beautifully depicted. Examining Śaivism in the ancient Khmer and Cam domains, Chemburkar and Kapoor (2018: 52) suggest that the Mỹ Sơn C1 tympanum shows an offering of music to Śiva. Since the Mỹ Sơn inscriptions of Harivarman II mention an offering of dance and music to the god, this evidence of the musical instruments (flute, tambour, ḍamaru) led them to suggest that they represent the Kāpālika cult that emerged from the Pāśupata sect. It remains a challenging topic to trace the types of schools of Śaivism that were practised throughout the history of multi-polity Campā because of the lack of evidence. However, it seems reasonable to suggest that the dancing Śiva from northern Campā were strongly influenced by Indian Śaiva religious art because of the representations of Śiva’s family members, attendants, his dancing poses, and multiple arms and hand gestures as they appeared in the above-mentioned sculptures (see Fig. 8.6 for an example of an Indian dancing Śiva). 20. See fig. 9 of Sivaramamurti 1974: 172.

28/1/22 7:44 PM

294

Mai Bùi Diệu Linh

The next two statues of dancing Śiva to examine are from Lương Hậu village, Hương Thuỷ district of Huế city (Fig. 8.4), and from Phong Lệ, Đà Nẵng (Fig. 8.5); the main reason for grouping these two together is that both have sixteen arms and the arms are prominently depicted. In both of these sculptures, we do not see the members of Śiva’s family; instead, the emphasis is on his worshippers. The Śiva tympanum was discovered in Lương Hậu village and thus called Lương Hậu tympanum (Fig. 8.4). This is the most recently discovered of the statues discussed here, as it was found a little over a decade ago. As far as I know, there are no publications on this sculpture in Western languages yet. The archeologist Lê Đình Phụng (2007: 113–117) was the first scholar to publish information on this sculpture, which he dated to the 9th to 10th centuries. However, the art historian Ngô Văn Doanh (2012) suggested that this dancing Śiva shares common characteristics with the 8th century Mỹ Sơn Śiva tympana A1 and C1 mentioned above and thus dated it to the 8th century. It was discovered near the Bà Yàng shrine (the shrine of the ‘Heavenly Lady’), which is also known as Bà Chuẩn Đề (‘Lady of compassion, saviour, protector of Dharma’). After the discovery of this tympanum, local villagers brought it to the shrine and have worshipped it ever since as the main deity because they thought that she was the multi-armed Buddhist bodhisattva Phật Bà Quan Âm (Guan-yin/ Avalokiteśvara). Śiva here has sixteen arms but holds no attributes at all; instead, the focus is on his hand gestures. Aside from his front right hand being in abhaya-mudrā (‘non-fear’ gesture) while his front left hand is in gajahasta-mudrā (‘elephant hand’ gesture), all the right-hand gestures are symmetric with the left side. The two uppermost hands seem to be holding something like a banded snake over his head. This is reminiscent of a 5th century Gupta multi-armed Śiva dancing sculpture from Sirpur (see the image in Sivaramamurti 1974: 172) whose topmost pair of hands holds up a snake while the two main front hands are in the gestures of abhaya-mudrā and gajahasta-mudrā, the same as the Śiva from Lương Hậu. The only difference between the Indian’s Gupta dancing Śiva and the one from Lương Hậu lies in the attributes the Indian Śiva carries in other arms

08_CreSouthV1_5P_28Jan22.indd 294

Fig. 8.4: Dancing Śiva, Shrine Bà Yàng. Hương Thuỷ, Huế. (Photo: Trần Kỳ Phương)

such as the triśūla, ḍamaru, and kapāla. Indeed, it is from the Gupta period that representations of dancing Śiva in visual form became identifiable. In this Lương Hậu tympanum, Śiva is standing on a dwarf or the demon Apasmāra and, interestingly, the dwarf is looking straight at us. This is the only Cam representation of Śiva dancing on a dwarf that faces directly to the viewers. It is also very rare in India to have the dwarf placed in this position while supporting dancing Śiva. I was able to find only a few such cases, the first being the depiction of the 2nd century BC–1st century CE Gudimallam liṅga (in southeastern Andhra Pradesh), which shows a merging of aniconic and anthropomorphic manifestations of Śiva, and the second being the 9th-century CE Pallava bronze sculptures from Kuram, Nallur, and Kilakkadu in Tamil Nadu.21 An Apasmāra under dancing Śiva’s foot is common in South Indian art, though the dwarf more often turns toward the left, and sometimes to the right, but rarely directly toward the viewers. Furthermore, in the Lương Hậu tympanum, there are two devoted worshippers, a male figure on Śiva’s right who looks more like a sage than an ordinary, and another male figure on his left who is most probably a devotee. None of Śiva’s family members nor his favourite attendant Bhṛṅgin were depicted. 21. For some illustrations, see Doniger 2009: 22–23; Sivaramamurti 1974: 168–169, 204–206.

28/1/22 7:44 PM

Śaiva Religious Iconography The dancing Śiva from Phong Lệ, Đà Nẵng (Fig. 8.5) is one of the best-carved and preserved Cam sculptures among all the depictions of dancing Śiva. There is a proportional balance between left and right and between the upper and lower parts of the sculpture. All the figures fit perfectly in the space of the arched form of the tympanum. The artist managed to bring out the overall harmony among all the figures in this sculpture in terms of the facial expressions of Śiva and his worshippers, and the techniques for proportional measurement of the statue. Śiva here is depicted with a crescent moon inserted in his headdress, one of his first identifying characteristics. Śiva has sixteen arms with all of the hands in identical gestures. Pierre Baptiste (2005: 237; 2018: 210–211) asserts that the hand gestures of the main arms, even though looking like vitarka-mudrā (hand gesture of argumentation or discussion in which the tips of the thumb and the index finger touch while the other three fingers remain straight), are more likely the arālahasta (the hand gesture in which the index finger is bent and the thumb is held close to the palm, the tips do not necessary touch one another, and the three other fingers remain straight) described in the dance manual Nāṭyaśāstra.22 In the upper part of the right main arm we see the serpent Śeṣa that curls as his armband or bracelet ornament. There is a harpist and a drummer on each side of Śiva. There are also several floating 22. Sivaramamurti (1974: 362) suggests that the hands of Śiva are in patāka signifying abhaya, the mudrā for ācamana (partaking of a small quantity of water). See Iyer 1998: 51 for a more detailed explanation of the arāla hand gesture. It remains unclear whether the Cam had access to any Sanskrit manuals on temple architecture and religious iconography such as the Śilpaśāstra and Nāṭyaśāstra. There is an ongoing debate on the transmission of knowledge that made it possible to create Hindu temples and sculptures in Southeast Asia. There were attempts to connect temple architecture and iconography to textual evidence and to claim that Southeast Asian architects might have received some texts from the Indian subcontinent and then built temples on the basis of prescriptions contained in the texts. But it remains unknown whether the builders and the sculptors were able to read the texts or they constantly needed an ācārya or a Brahmin to supervise them. Also, another problem is that there were temples that had been built before those manuals were actually produced or codified.

08_CreSouthV1_5P_28Jan22.indd 295

295

Fig. 8.5: Dancing Śiva. Phong Lệ, Đà Nẵng. Đà Nẵng Museum of Cham Sculpture. (Photo: author)

worshippers looking as half-bird and half-human celestial figures on each side of Śiva; they are presented with their hands pressed together in the devotional position (añjali-mudrā) (Tingley 2009: 222). Baptiste suggests that this tympanum shares similar characteristics with the art monuments of the Cāḷukyas at Badami, Karnataka, and especially the tympanum of dancing Śiva at the entrance of the Taraka Brahma temple in Alampur.23 The dancing Śiva here is in the lalita pose; in India we can see this type of dancing Śiva pose in a 6th century depiction in Ellora cave (Fig. 8.6). In this Indian icon, dancing Śiva has eight arms and the main ones are in gajahasta-mudrā (‘elephant hand’) and āhūyavarada-mudrā (hand gesture of invitation and charity) carrying several of his typical attributes. On the right of Śiva are three musicians who play the ūrdhva drum, flute, and cymbal. Peeping out from behind Śiva’s right leg is a tall and skinny figure, most probably Bhṛṅgin. On the left side of Śiva is Devī holding the hand of baby Skanda watching her husband’s dance. These figures were adopted directly to the northern Cam dancing Śiva iconography while we do not see them in the southern Cam depictions of dancing Śiva (see below). The last two northern statues of Cam dancing Śiva are both from Khương Mỹ of Quảng Nam province (Figs. 8.7 and 8.8). They are dated to the 23. See Baptiste 2005: 237–331, and note 2.

28/1/22 7:44 PM

Mai Bùi Diệu Linh

296

Fig. 8.6: Dancing Śiva. Vākāṭaka, Ellora. Aurangabad, Maharashtra, India. (Photo: author)

Figs. 8.7–8.8: Dancing Śiva. Khương Mỹ, Quảng Nam. Đà Nẵng Museum of Cham Sculpture. (Photo: author)

08_CreSouthV1_5P_28Jan22.indd 296

28/1/22 7:44 PM

Śaiva Religious Iconography 10th century CE and are currently exhibited in the Đà Nẵng Museum of Cham Sculpture. The lower part of the dancing Śiva in Fig. 8.7 was totally broken; however, Fig. 8.8 gives us a sense of what the other sculpture would look like if the Nandin was still intact. In both tympana, the face of Śiva is unfortunately badly damaged and it is unclear what kind of attributes each deity was holding in his hands. In Fig. 8.7, Śiva has twenty-eight arms; some of the hands are holding his typical attributes such as a rosary, a ball, a club, a disk, a vase; some others hold a rope while the topmost pair of hands holds up the snake. In both sculptures, Śiva holds a vīṇā (a stringed instrument) or a ṭuila (bar-zither) and is twisting backwards showing both his torso and his lower back at once. Dancing Śiva on a bull with his two main hands holding a vīṇā was depicted in the Pāla period at Natghar in the Brahmanbaria subdivision of Tippera and at Ranihati of Dacca in the 10th century; however, in these images Śiva’s body is not twisted.24 The images of dancing Śiva twisting his back and hips can be seen in cave fourteen of Ellora and at the temple of Markand in Chanda district of Maharashtra.25 However, in the Indian examples, Śiva does not dance on a Nandin nor hold a vīṇā. Interestingly, Sivaramamurti (1974: 363) mentions that the Nandin from Khương Mỹ (Fig 8.8) has a large necklace of bells and this is usually seen in the case of Cāḷukyan bulls. Accordingly, it is reasonable to conjecture that the northern Cam could have adopted Indian iconographical elements from Ellora art in depicting Śiva dancing. However, while examining dancing Śiva from Bengal, Ślączka (2015: 133) suggests that Cam dancing Śiva from Khương Mỹ (Figs. 8.7 and 8.8) share some elements with the twelve-armed images from Bengal. In particular, the uppermost arms are raised with the hands clasped above the head, the second pair holds a snake, and the main pair of hands plays a vīṇā. She maintains that those elements occur exclusively in Bengal, thus signifying that the Cam dancing Śiva was developed under 24. See figs. 173, 174 in Sivaramamurti 1974: 299–300. 25. See the images in Sivaramamurti 1974: 218, 337. However, the dancing Śiva from the Markand temple is of the Paramāra tradition and dated to the 11th century, which is much later than the one from Ellora and also later than the ones from Khương Mỹ.

08_CreSouthV1_5P_28Jan22.indd 297

297

the influence of Bengal art. Consequently, I suggest that the northern Cam might have received Indian iconographical influences from various sources through different periods of time, and that transmission and adoption happened before the 10th century. It is important to note that dancing Śiva did not appear in Khmer art until the 10th century and this form of Śiva did not attract the Javanese at all. This leads to the idea that the northern Cam dancing Śiva received Indian iconography directly from India and not through the intermediaries of Khmer or Javanese art. Interestingly, depictions of dancing Śiva dated prior to the 10th century all come from the northern part of Campā and emphasize Śiva with family members, musicians, and attendants. This is in contrast to the dancing Śiva from the southern parts of Campā, as not only they are all dated to a later period, from the 12th to the 13th centuries, but they also have a different iconographical focus. For example, such Śivas were typically depicted alone or with only two worshippers on each side, the emphasis instead being on their hand gestures and attributes.

dancing śiva from southern campā (vijaya and pāṇḍuraṅga) From the 10th to the 11th centuries, we do not see any dancing Śiva sculpture in any part of Campā. However, between the 12th and 13th centuries, a new representation of dancing Śiva was constructed in the Mỹ Sơn sanctuary for the main temple of group H1 (Fig. 8.9). This sculpture technically belongs to the North, but since it was sculpted in the later period and with strong influences from the Tháp Mẫm decorative art style further south, I think it would be more proper to group it together with the southern ones. According to Zéphir (2005: 301), this tympanum used to be on the top of the main entrance at the East side. Looking very different from the other two tympana of Mỹ Sơn A1 and C1, this sculpture has neither Śiva family members nor Nandin or Bhṛṅgin. The tympanum was constructed with three different pieces of sandstone rather than one piece like other tympana. The shape of the tympanum is round and large in the lower part

28/1/22 7:44 PM

298

Mai Bùi Diệu Linh

Fig. 8.9: Dancing Śiva, Mỹ Sơn H1. Mỹ Sơn Sanctuary. (Photo: Trần Kỳ Phương)

of the base and slowly converges toward the top as a pointed arch. This is a new feature of tympana from this period onward. The main figure of Śiva is recognized by the typical rosary attribute and his sacred thread tied in the form of a snake. Here the act of dancing is shown by the bent thighs of the deity in chatura position, the same as the other two statues of dancing Śiva found in Mỹ Sơn. Zéphir (2005: 301) suggests that the two hands of Śiva pointing upwards is an innovation in Cam art introduced at that time, since we do not see anything like this before the 12th century; however, he finds this hand gesture to be obscure. Śiva here is standing on a lotus petal base with two makaras holding lotus buds on each side. The appearance of the two makaras is interesting as we do not see makara in any other Cam dancing Śiva statues. Zéphir (ibid.) believes that the style of the makaras in this Mỹ Sơn H1 belongs to the Tháp Mẫm decorative art style of the 12th and 13th centuries and thus may reflect

08_CreSouthV1_5P_28Jan22.indd 298

a link between Mỹ Sơn and southern Campā. In addition, he suggests that this icon shows a strong influence from Khmer art: in particular, the facial expression (the large mouth and the closed eyes) would be due to Khmer influence, as well as the garments, with the large earring and the large body features being very similar to the Khmer style of Bayon during the period from the 12th and the 13th centuries. This leads Zéphir to suggest that this sculpture may point to contacts between Cambodia and Campā during the reign of the Khmer King Jayavarman VII (c. 1181–1220?) (ibid.). This idea of contact or mutual influence between Cambodia and Campā becomes more apparent when examining the next set of sculptures, also dated between the 12th and the 13th centuries. Further South, in Vijaya (present-day Bình Định province), there appeared a distinct type of dancing Śiva represented by a sculpture which was found in Tháp Mẫm in 1934 and is currently exhibited in the Đà Nẵng Museum of Cham Sculpture (Fig. 8.10).26 Guillon (2001: 142) speculated that this sculpture belonged to a group of four identical images, although he gave no details concerning the other three. He also suggested that these sculptures were intended to be attached to a wall rather than be part of a tympanum. Zéphir (2005: 318–319; 2018: 250–251), however, asserts that this sculpture was separated from a set of three sculptures in the photo archive of the Guimet museum. He proposed that the sculpture could be a cornice of a principal shrine or an antefix. I saw an almost identical version of this Tháp Mẫm dancing Śiva sculpture in the National Museum of Vietnamese History in Hà Nội, but I do not know where the third one is and whether there were four of them, as Guillon claims. In this Tháp Mẫm sculpture, Śiva holds an axe or a lance in his right hand and a trident (triśūla) in his left hand. The trident is the most famous attribute of Śiva and in India the earliest depiction of the triśūla is from the period prior to the Gupta; it can be seen in coins of Maues and Gondophares and in the copper seal discovered at Sirkap (Sivaramamur26. Tháp Mẫm is the name of a Cam temple and it is also used to refer to the 12th to 13th centuries Cam art of this region. The Tháp Mẫm site is located to the north of the Cam citadel Chà Bàn of Vijaya polity, about 25 km to the north of modern-day Qui Nhơn.

28/1/22 7:44 PM

Śaiva Religious Iconography

Fig. 8.10: Dancing Śiva, Tháp Mẫm, Bình Định. Đà Nẵng Museum of Cham Sculpture. (Photo: author)

ti 1947: 168). Interestingly, even though the triśūla is the most typical attribute of Śiva, it was only from this period and from the southern part of Campā that we find it depicted with dancing Cam Śiva. In the Tháp Mẫm sculpture, other typical attributes of Śiva are also present, such as a crescent moon inserted into his headdress and a third eye decorating his forehead. It is intriguing that the three most important of Śiva’s attributes were all represented in this sculpture while they were not adopted and presented in the images of dancing Śiva before the 10th century in the northern parts of Campā. In the Tháp Mẫm dancing Śiva, the two rear arms of Śiva are held up and the hands linked together above the head. Zéphir (2005: 318–319) suggests that this could be añjali-mudrā (‘gesture of worshiping’) or naivedya-mudrā (‘gesture of offering nutrition’, either concrete or spiritual). He adds that if this is a depiction of Śiva in naivedya-mudrā then it may suggest that Śiva here is in an offering position to the principal deity. But one can wonder who this

08_CreSouthV1_5P_28Jan22.indd 299

299

deity may be that such a powerful god as Śiva could be venerating. In researching Cam-Khmer interactions, Peter Sharrock (2018: 111) stated that the period of the early 12th to early 13th centuries saw ‘a succession of friendships, and military, economic and cultural alliances.’ He explains that the Khmer King Sūryavarman II (1113–1149 CE) helped the Cam to fight against Đại Việt to recover their lands during the 1120s and that, after the Chinese Song empire opened up foreign trade by sea, the Cam and Khmer became closer allies by developing the Thị Nại port in Vijaya. This alliance can be seen in the growth in trade and the visible Khmer influence in architecture and statuary in southern Campā. Particularly, Sharrock (ibid.: 116) suggests that the evidence of Khmer influence on Cam temple art and statuary can be seen in the Tháp Mẫm style, especially the high reliefs of dancers and the Garuḍa decorative motifs. The dancing Śiva from Tháp Mẫm also attests to Khmer influence in religious art, for instance the detailed decorations of heavy jewellery. Actually, prior to the 10th century, there is only a single anthropomorphic image of Śiva in any form in all of Khmer art (James 2011: 199). However, from the 10th to 13th centuries, there were quite a number of tympana of Khmer dancing Śiva, such as those from Phnom Rung, Banteay Samre, Banteay Srei, Phnom Chisor, and Phimai.27 In Campā, no dancing Śiva has been dated to around the 11th century; they appeared only in the 12th to 13th centuries, when the southern Cam of Vijaya were in close contact with the Khmer. In the statue of dancing Śiva from Tháp Mẫm (Fig. 8.10), the forward fold of the sampot hanging freely between Śiva’s legs enhances the image of dance movement. This fold of cloth appears again in the tympanum of dancing Śiva at Po Klaong Girai temple in Pāṇḍuraṅga in present-day Ninh Thuận province (Figs. 8.11 and 8.12). This is the last dancing Śiva sculpture I examine here. This pointed tympanum from the 12th to 13th centuries is currently mounted facing East above the main door of the main shrine of Po Klaong Girai temple complex. One can recognize the figure of Śiva 27. See the images in James 2011 and Sivaramamurti 1974: 354–357.

28/1/22 7:44 PM

300

Mai Bùi Diệu Linh

Figs. 8.11–8.12: Dancing Śiva, Po Klaong Girai temple, Ninh Thuận. (Photo: author)

because his sacred thread and third eye are clearly depicted. He has six arms, four of which hold attributes such as a lotus bud, a bow, a triśūla, and a sword (khaḍga). The two rear arms are held up with the two hands linked together above the head, exactly as in the sculpture of the dancing Śiva from Tháp Mẫm.28 This hand gesture is not, however, known in any Khmer statue of Śiva. If Śiva is in naivedya-mudrā, as Zéphir suggests, it implies a deity superior to Śiva. However, I believe that the principal deity of this temple was Śiva himself, represented in his symbolic form—as a Śivaliṅga, which remains in situ in the main sanctuary of the temple.29 Śiva here is in lalita dancing pose and the 28. Sivaramamurti (1974: 364) describes the topmost pair of hands raised over his head for simply for clapping and not in añjali-mudrā (‘gesture of worshiping’) or naivedya-mudrā (‘gesture of offering nutrition’). 29. This liṅga is currently covered with a paper case (liṅgakośa or mukhakośa) created by contemporary Cam to represent the legendary King Po Klaong Girai. This king

08_CreSouthV1_5P_28Jan22.indd 300

dynamism of movement is stressed by the cloth between his legs and the two ends of a scarf floating out on each side of the figure. This hanging cloth between Śiva’s legs looks almost the same as that of the dancing Śiva from Tháp Mẫm, suggesting a connection between these two Cam polities. Other elements such as heavy necklace jewellery, high hair cover, and detailed skirt decorations portrayed in dancing Śiva from both Vijaya and Pāṇḍuraṅga also point to the idea of a mutual artistic influence between the two Cam polities. Indeed, the Khmer were present not only in Vijaya; Sūryavarman’s nephew Jaya Harivarman (also known as Prince Śivānanda) and his father stayed in Pāṇḍuraṅga, as did Sūryavarman’s reputed general (senāpati) Śankara (Sharrock 2018: 116). This could explain why the representation of dancing Śiva in Po is currently worshipped by the Cam as the main deity of the temple instead of Śiva. As for the topic of mediaeval Śaiva ritual and the practice of liṅgakośa and mukhakośa, see Bhattacharya 1966 and Guy 2018: 89–96.

28/1/22 7:44 PM

Śaiva Religious Iconography Klaong Girai temple shares common features with the dancing Śiva from Tháp Mẫm, as both these Cam polities might have received influences from Khmer art. And this could explain why southern dancing Śivas look quite different from the northern ones, especially insofar that in the southern images, the statues of Śiva are not accompanied by images of his family members or attendants but there is a greater emphasis on the deity’s attributes with two back hands linked above his head.

conclusion Dancing Śiva was visualized and iconographically represented in Campā as early as the 8th century and continued to hold an important place in Cam religious art up until the 13th century. Interestingly, the dancing Śiva dated between the 8th and 10th centuries were all found in the northern regions of Campā (Indrapura, Amarāvatī), while the sculptures from the later period, between the 12th to 13th centuries, came mainly from the southern part (Vijaya, and Pāṇḍuraṅga), except the dancing Śiva from Mỹ Sơn H1, which is a special case (it was found in Mỹ Sơn but bears strong similarities with the Tháp Mẫm decorative art style). Dancing Śivas found in the North were portrayed very often with his family members such as Pārvatī, Gaṇeśa and Skanda, as well as with his attendants Bhṛṅgin, Nandin, and the dwarf Apasmāra. These characters were adopted directly from Indian iconography. In addition, the crossed arms position and the dancing poses seem to be influenced by Indian Śaiva iconography of dancing Śiva. However, in the statues of dancing Śiva from the South, we do not see the portrayal of Śiva’s family members nor his attendants Bhṛṅgin, Nandin, and the dwarf Apasmāra. Instead, attributes of Śiva such as the trident (śūla or triśūla), a club (khaṭvāṅga), a rosary, and an axe were depicted. These attributes were not main features of the 8th to 10th centuries sculptures found in the northern regions of Campā even though they had Indian antecedents. It is clear that the Cam selectively appropriated a number of Śaiva religious art features from various sources. One can ask why the northern Cam were not interested in those attributes of Śiva that were of the utmost importance. Śaivism was clearly a

08_CreSouthV1_5P_28Jan22.indd 301

301

dominant religion of various Cam polities, but on the basis of this preliminary survey of the dancing Śiva iconographical representation it seems reasonable to suggest that perhaps the form of Śaivism that was practiced in the northern part of Campā could have undergone modifications in the Southern part. After all, Campā never existed as a single entity and thus different varieties of Śaivism could have been practised more or less independently in different parts of Campā. For the time being, it is important to note that dancing Śiva was not a common element within Khmer religious art prior to the 10th century and in fact there was only a single Khmer anthropomorphic image of Śiva in any form prior to the 9th century. In insular Southeast Asia, images of dancing Śiva were not adopted and the only example of a dancing deity tentatively identified as Śiva can be seen is the dancing figures on the walls of Candi Śiva at the Prambanan temple complex in Central Java, dating from the 9th century, but this is still a hypothesis and needs further research before it can be definitely confirmed (Iyer 1998). Thus, the 8th-century dancing Śiva tympana from Campā, of Lương Hậu, and Mỹ Sơn (A1 and C1) may be regarded as the earliest depictions of the dancing Śiva in Southeast Asia. In the light of this fact, I suggest that Campā might have received a direct early transmission of religious art from India to its northern regions. Furthermore, on account of the close political and trading connections between southern Campā and the Khmer kings in the 12th to 13th centuries, and of the elements of Khmer decorative motifs seen in dancing Śiva from Tháp Mẫm and Po Klaong Girai, I would suggest that Campā might have also received Khmer art influence during this period. In short, at this stage it is possible to speculate that in the early period, from the 8th to 10th centuries, Northern Campā seems to be influenced by Indian iconography through a direct transmission while from the 12th century onwards the presence of Khmer religious art influences and connections can be seen in the Cam art of Vijaya and Pāṇḍuraṅga. What is important to stress here is that the Cam did not simply copy and reproduce the religious art elements they received from Indian and Khmer sources, but rather adapted them and ex-

28/1/22 7:44 PM

302

Mai Bùi Diệu Linh

pressed new forms based on their own concepts. For instance, the uppermost hands interlocked and joined above Śiva’s head are a feature unique to southern Campā, which is especially strongly present in Cam art from the 12th century in Vijaya and Pāṇḍuraṅga, even though the exact meaning of this hand gesture as intended by the Cam remains unclear. The extant sculptures of dancing Śiva from Indrapura and Amarāvatī in the north of Campā did not employ this type of hand gesture, nor can this element be found in the religious art of Southeast Asia outside of Campā, thereby suggesting a local interest. In addition, most of the Cam statues of dancing Śiva were depicted in the form of tympana that were intended to be placed at the main entrances of temples. This is the second most prominent position in a temple, second only to the central sanctuary icon. To have a dancing Śiva placed above the main entrance signifies that the temple belongs to Śiva as the central deity. Śiva has many manifestations and the fact that this manifestation was selected to be represented on the tympanum suggests that the Cam were especially attracted to this form of Śiva.

References

Aspell, William. 2013. ‘Southeast Asia in the Suishu: A Translation of Memoire 47 with Notes and Commentary’, Asia Research Institute Working Paper Series No. 208: 3–33. Singapore: ARI, National University of Singapore. Baptiste, Pierre. 2005. ‘Les piédestaux et les soubassements dans l’art cham hindou: une spécificité?’, in Pierre Baptiste and Thierry Zéphir (eds.), Trésors d’art du Vietnam. La sculpture du Champa Ve-XVe siècles, pp. 109–121. Paris: Réunion des musées nationaux and Musée des arts asiatiques-Guimet. . 2018. ‘Pediment of Dancing Siva’, in Trần Kỳ Phương, V. Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 210–211. Bangkok: River Books. Baptiste, Pierre and Thierry Zéphir (eds.). 2005. Trésors d’art du Vietnam. La sculpture du Champa Ve-XVe siècles. Paris: Réunion des musées nationaux and Musée des arts asiatiques-Guimet.

08_CreSouthV1_5P_28Jan22.indd 302

Bhattacharya, Kamaleswar. 1966. ‘Liṅga-Kośa’, Artibus Asiae, Supplementum 23: 6–13. Boisselier, Jean. 1963. La statuaire du Champa: Recherches sur les cultes et l’iconographie. Paris: École française d’Extrême-Orient. Chemburkar, Swati and Shivani Kapoor. 2018. ‘The Pāśupata sect in Ancient Cambodia and Champa’, in Trần Kỳ Phương, Võ Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 45–56. Bangkok: River Books. Cœdès, George and Henri Parmentier. 1923. Listes Générales des inscriptions et des monuments du Champa et du Cambodge. Hanoi: Imprimerie d’Extrême-Orient. Coral Remusat, Gilberte de. 1934. ‘Art Cam, le problème de la chronologie’, in P. Dupont et al. (eds.), Musée Guimet: Catalogue des collections indochinoises, pp. 35–44. Paris: Musée Nationaux. Davis, Richard H. 2017. ‘Śiva’s Multiplicity of Presence in Aniconic and Iconic Form’, Religion 47/3: 459–482. Doniger, Wendy. 2009. The Hindus: An Alternative History. Oxford: Oxford University Press. Gillet, Valérie. 2010. La création d'une iconographie śivaïte narrative: incarnations du dieu dans les temples pallava construits. Paris: École française d’Extrême-Orient. Golzio, Karl-Heinz. 2004. Inscription of Campā. Berlin: Shaker Verlag. Griffiths, Arlo. 2019. ‘Epigraphical Texts and Sculptural Steles Produced under the Vīrabhadravarmadevas of 15th-Century Campā’, in Arlo Griffiths, Andrew Hardy, and Geoff Wade (eds.), Champa: Territories and Networks of a Southeast Asian Kingdom, pp. 193–220. Paris: École française d’Extrême-Orient. Griffiths, Arlo, Amandine Lepoutre, William A. Southworth, and Thành Phần. 2012. The Inscriptions of Campa at the Museum of Cham Sculpture in Đà Nẵng. HCM: VNU-HCM Publishing House. Griffiths, Arlo, Andrew Hardy, and Geoff Wade (eds.). 2019. ‘Preface: Frameworks of Champa History’, in Arlo Griffiths, Andrew Hardy, and Geoff Wade (eds.), Champa: Territories and Networks of a Southeast Asian Kingdom,

28/1/22 7:44 PM

Śaiva Religious Iconography pp. 13–20. Paris: École française d’ExtrêmeOrient. Guillon, Emmanuel. 2001. Cham Art: Treasures from the Da Nang Museum, Vietnam. Bangkok: River Books. Guy, John. 2014. Lost Kingdoms: Hindu-Buddhist Sculpture of Early Southeast Asia. New York: Metropolitan Museum of Art. Guy, John. 2018. ‘Śaiva Ritual: liṅgakośa and mukhakośa in Champa’, in Trần Kỳ Phương, Võ Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 89–96. Bangkok: River Books. Hall, Kenneth R. 2018. ‘Champa Ports-of-Trade Networking on the Coastline, ca. 300–1500 CE’, in Trần Kỳ Phương, Võ Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 19–30. Bangkok: River Books. Iyer, Alessandra. 1998. Prambanan: Sculpture and Dance in Ancient Java. Thailand: White Lotus Press. . 2000. ‘Śiva’s Dance: Iconography and Dance Practice in South and Southeast Asia’, Music in Art 25/1–2: 25–32. James, Soumya. 2011. The Symbiosis of Image, Monument and Landscape: A Study of Select Goddess Images at Prasat Kravan, Kbal Spean and Banteay Srei in Cambodia. PhD dissertation, Cornell University. . 2014. ‘Durga and Śiva at Banteay Srei: Blurring Boundaries between Monument, Images and Practice’, in Upinder Singh and Parul (eds.), Asian Encounters: Exploring Connected Histories, pp. 137–158. Oxford: Oxford University Press. Kaimal, Padma. 1999. ‘Shiva Nataraja: Shifting Meanings of an Icon’, The Art Bulletin 81/3: 390–419. . 2000. ‘The Problem of Portraiture in South India, circa 970–1000 A.D.’, Artibus Asiae 60/1: 139–179. . 2011. ‘Shiva Nataraja: Multiple Meanings of an Icon’, in Rebecca M. Brown and Deborah S. Hutton (eds.), A Companion to Asian Art and Architecture, pp. 390–419. Chichester, West Sussex; Maiden, MA: Blackwell Publishing Ltd.

08_CreSouthV1_5P_28Jan22.indd 303

303

Lê Đình Phụng. 2007. Văn Hoá Champa ở Thừa Thiên Huế. Hà Nội: Nxb. Văn Hoá-Thông Tin và Viện Văn hoá. Majumdar, R.C. 1985. Champa History & Culture of an Indian Colonial Kingdom in the Far East 2nd–16th Century A.D. Delhi: Gian Publishing House. [Reprint; first edition 1927] Ngô Văn Doanh. 2012. ‘Lương Hậu sculpture and Dancing Shiva in Cham art’. A paper delivered at the International Conference on Cham Civilization, Đà Nẵng, 26–27 June 2012. Nguyễn Thế Anh. 1995. ‘The Vietnamization of the Cham deity Po Nagar’, Asia Journal 2/1: 55–67. Noseworthy, William. 2015. ‘The Mother Goddess of Champa: Po Ina Nagar’, Suvannabhumi: Multi-Disciplinary Journal of Southeast Asian Studies 7/1: 107–137. Parmentier, Henri. 1909. Inventaire descriptif des monuments cams de l’Annam. Publications de l’École française d’Extrême-Orient. Paris: Leroux. . 1918. Inventaire Descriptif des Monuments Cams de l’Annam. Tome II. Paris: Leroux. Po Dharma. 2001. ‘The History of Champa’, in Emmanuel Guillon (ed.), Cham Art: Treasures from the Da Nang Museum, Vietnam, pp. 14–27. Bangkok: River Books. Rao, T.A. Gopinatha. 1916. Elements of Hindu Iconography, vol. 2, pt. 1. Madras: The Law Printing House. [Reprints of 1971, 1993, New Delhi] Schweyer, Anne-Valérie. 2004. ‘Po Nagar de Nha Trang (Part 1)’, Aséanie 14: 109–140. . 2005. ‘Po Nagar de Nha Trang (Part 2)’, Aséanie 15: 87–119. . 2010. ‘The Birth of Champa: Crossing Borders in Southeast Asian Archaeology’. [A conference paper available online at https:// halshs.archives-ouvertes.fr/halshs-00828812/ document] Sharrock, Peter D. 2018. ‘Cham-Khmer interactions 1113–1220 CE’, in Trần Kỳ Phương, V. Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Đà Nẵng Museum of Cham Sculpture, pp. 111–119. Bangkok: River Books. Shiro, Momoki. 2011. ‘“Mandala Champa” seen from Chinese Sources’, in Trần Kỳ Phương

28/1/22 7:44 PM

304

Mai Bùi Diệu Linh

and Bruce M. Lockhart (eds.), The Cham of Vietnam: History, Society and Art, pp. 120–137. Singapore: NUS Press. Sivaramamurti, C. 1974. Naṭarāja in Art, Thought, and Literature. New Delhi: National Museum. Ślączka, Anna. 2015. ‘Dancing Śiva Images from Bengal’, in Mokammal H. Bhuiyan (ed.), Studies in South Asian Heritage: Essays in Memory of M Harunur Rashid, pp. 125–155. Dhaka: Bangala Academy. Stern, Philippe. 1942. L’art du Champa (ancien Annam) et son évolution. Toulouse and Paris: Les Frères Douladoure and Adrien-Maisonneuve. Taylor, Keith. 1999. ‘The Early Kingdoms’, in The Cambridge History of Southeast Asia: Vol. 1, Pt. 1 (From Early Times to c. 1500), pp. 137–182. Cambridge: Cambridge University Press. Tingley, Nancy. 2009. Arts of Ancient Vietnam: From River Plain to Open Sea. Houston: The Museum of Fine Arts. Trần Kỳ Phương and Bruce M. Lockhart (eds.). 2011. The Cham of Vietnam: History, Society and Art. Singapore: NUS Press Trần Kỳ Phương and Rie Nakamura. 2008. ‘The Mỹ Sơn and Pô Nagar Nha Trang Sanctuaries: On the Cosmological Dualist Cult of the Champa

08_CreSouthV1_5P_28Jan22.indd 304

Kingdom in Central Vietnam as seen from Art and Anthropology’, Asia Research Institute Working Paper Series No. 100. Singapore: ARI, National University of Singapore. Vickery, Michael. 2011. ‘Champa Revised’, in Trần Kỳ Phương and Bruce M. Lockhart (eds.), The Cham of Vietnam: History, Society and Art, pp. 363–420. Singapore: NUS Press. Wade, Geoffrey. 2003. ‘The Ming shi Account of Champa’, Asia Research Institute Working Paper Series No. 3. Singapore: ARI, National University of Singapore. . 2005. ‘Champa in the Song hui-yao: A draft translation’, Asia Research Institute Working Paper Series No. 53. Singapore: ARI, National University of Singapore. Zéphir, Thierry. 2005. ‘Introduction à la Sculpture du Champa Ancien’, in Pierre Baptiste and Thierry Zéphir (eds.), Trésors d’art du Vietnam: la sculpture du Champa, Ve–XVe siècles. pp. 123–131. Paris: Musée national des arts asiatiques-Guimet. . 2018. ‘The Sculpture of Champa: Specificity & Evolution’, in Trần Kỳ Phương, Võ Văn Thắng, and Peter D. Sharrock (eds.), Vibrancy in Stone: Masterpieces of the Da Nang Museum of Cham Sculpture, pp. 65–70. Bangkok: River Books.

28/1/22 7:44 PM

Chapter 9

The Colossal Trà Kiệu Pedestal in Campā and its Relationship to Courtly Culture in Cambodia, East Java, and China1 m ya c h au 1

T

introduction

he polities of Campā along the coast of central Vietnam were not politically united, but the Cams appear to have participated in temple and courtly culture, as can be seen through their inscriptions and image making. A 16th-century Spanish account of the customs in Campā mentions at least five courtly festivals held annually by the Cam king, including feasts, plays, public races, and celebrations of elephant and tiger hunts (Souza and Turley 2016: 677–680). In addition, Chinese histories describe the Cam kings performing public processions with elephants, parasols, drums, and conches. These observations indicate a public display suggesting that the Cam polities shared political and cultural expression of power and religion.2 This chapter explores the extent to which the Cams incorporated themselves into the international political and economic sphere in the 9th–10th centuries by projecting onto their artistic 1. I am indebted to research carried out by the late Julie Romain in her 2015 dissertation, ‘Courtly Culture and Visual Art in India: Rāmāyaṇa Reliefs on Hindu Temples of the Sixth to Eighth Century’. She also offered advice and encouragement from the first time I met her in 2011 and throughout my graduate studies at UCLA in 2012–2017. In her dissertation, she argues that the Rāmāyaṇa reliefs in Hindu temples reflected ‘secular, courtly culture’, and that the images of Rāma emphasize an honorific function, and highlight his role as a courtly hero. I thank Robert L. Brown, Peter Sharrock, and Andrea Acri for their comments on earlier versions of this chapter. In addition, two anonymous peer reviewers also provided helpful suggestions for revision. 2. Cefu yuangui, 959.11288; Taiping yulan (Encyclopedia Assembled for Imperial Inspection during the Taiping Era) (983; Taibei: Shangwu yinshuguan, 1980), 786.3611.

09_CreSouthV1_8P_11Feb22.indd 305

productions a complex and creative military and courtly culture. One of the most well-known sacred epics in South and Southeast Asia is the Rāmāyaṇa, the life story of Rāma, an avatar of Viṣṇu. Using textual sources, visual evidence, and recent scholarship, this chapter analyses the colossal Trà Kiệu pedestal within the context of the Rāmāyaṇa in Khmer and Javanese cultures during the 9th–10th centuries. Visual evidence includes the Rāmāyaṇa sandstone reliefs preserved at the Đà Nẵng Museum of Cam Sculpture and imagery from three central Vietnam archaeological sites: Khương Mỹ, Chiên Đàn, and Chánh Lộ. The visual evidence shows three shared iconographic elements: archers with bows, architectural features, and dancing apsarases found in Campā, East Java, and Cambodia. I suggest that the Rāmāyaṇa reliefs reveal that the Trà Kiệu pedestal found in a temple context also promoted courtly culture with close relationships to the art and cultures of Cambodia, East Java, and China in the 9th–10th centuries. I define courtly culture as comprising ceremonial, celebratory, and martial activities with an emphasis on symbols of power, hierarchy, and etiquette for political motivations. Within the sacred sphere, the stone Trà Kiệu pedestal also commemorates royal-divine interactions through the narrative of the Rāmāyaṇa.3

3. Peter Sharrock importantly notes that stone was normally a material reserved only for the gods (personal communication, 2019). In addition to religious motives, I suggest that the 10th-century Trà Kiệu pedestal’s imagery has political implications as well. The royal court was also seen through the lens of the sacred.

11/2/22 3:48 PM

306

Mya Chau

Fig. 9.1: Plan of Trà Kiệu, Campā. © JY Claeys, Bulletin de l’École française d’Extrême-Orient 27, 1927, Pl. XXXVIII

history of the trà kiệu pedestal The colossal Trà Kiệu pedestal and the liṅga-yoni were discovered in the 19th century in Trà Kiệu village of Quảng Nam province in central Vietnam. The objects probably originated from point B in Trà Kiệu beside the main sanctuary, but not having been scientifically excavated, the original archaeological context of the pieces is unknown (Fig. 9.1).4 Early photographs of the Trà Kiệu pedestal suggest that it was once topped by a liṅga and thus focused on the worship of Śiva, but it is 4. The earliest record which documents the Trà Kiệu pedestal’s transfer to Tourane is by E.L. Lajonquière (1901).

09_CreSouthV1_8P_11Feb22.indd 306

possible that other deities were instead the objects of devotion. The first images of the Trà Kiệu pedestal were published in 1894 by French Resident Charles Lemire in an excursion report to Tourane, now modern-day Đà Nẵng, Vietnam. We do not know what sculptures were originally placed on the pedestal. During the colonial period, several different icons were displayed on the pedestal due to archaeological and curatorial considerations.5 For example, Lemire’s report shows a statue of a 5. Regardless of the reasons behind the arrangement, local and foreign visitors still continue to pray to and worship Cam objects in the Đà Nẵng Museum of Cham Sculpture.

11/2/22 3:48 PM

The Colossal Trà Kiệu Pedestal in Campā

307

Fig. 9.2: Trà Kiệu Pedestal, Campā, 10th century. © Doktor Max, CC0 1.0

female goddess seated on the Trà Kiệu pedestal (Lemire 1894: 402).6 French archaeologists erected a monumental liṅga-yoni on top of the pedestal in the Tourane outdoor garden in a modern arrangement in the 1930s. The decorated yoni with doubled lotus petals was placed on a separate base. In a third instance, the doubled lotus petals yoni was not paired with the liṅga, but with a sandstone sculpture of an elephant. The Trà Kiệu pedestal with the liṅga-yoni remains displayed as a unit in the Đà Nẵng Museum of Cham Sculpture and documented by photographs in various publications (Fig. 9.2). Photographs link the Trà Kiệu pedestal and the liṅga-yoni as one entire structure, often referred to as an altar-pedestal, or simply a ‘pedestal’. The 6. This statue may be identified as the goddess Lakṣmī who is represented seated with her legs crossed and hands resting on both of her thighs. However, Lakṣmī is often flanked by two elephants which are not represented here, suggesting that she might also be a local Cam goddess.

09_CreSouthV1_8P_11Feb22.indd 307

term pedestal is used for the four largest extant Cam stone artefacts,7 including the Trà Kiệu structure. Scholars assume that these constructions were intended to support some sacred object or multiple objects placed on top of the structure, as a seat of the divine, or as an offering pedestal (pīṭha or balipīṭha, Tingley 2009: 106). A second type of object, of which Cam art has many examples, is a base with spouts for funneling off sanctified liquids used to anoint the object they support. I use the term yoni for these objects, although they do not always support only liṅgas or spouts. The Trà Kiệu pedestal must be studied independently from the liṅga-yoni because we have no evidence that the two objects were created to be displayed as a unit. We now consider whether the narrative scenes on the pedestal indicate its original purpose. Scholars have proposed a variety of interpretation re7. The four large Cam structures referred to as pedestals were found in the archaeological sites of Mỹ Sơn, Trà Kiệu, Đồng Duơng, and Vân Trạch Hòa.

11/2/22 3:48 PM

308

Mya Chau

Fig. 9.3: Trà Kiệu Pedestal, Campā, 10th century. © George Cœdès, Bulletin de l’École française d’Extrême-Orient 31, 1931, Pl. XVII

garding the narrative scenes and the identity of the characters carved on the four sides of the Trà Kiệu pedestal (Fig. 9.3). All of the theories try to connect the scenes to a narrative of a myth or a text. For example, Jean Przyluski (1929) suggested that the pedestal depicts the legend of Kauṇḍinya, the mythical king of Funan. George Cœdès (1931) challenged Przyluski’s interpretation, arguing that the imagery reflects three life scenes of Kṛṣṇa from the Hindu text Bhāgavatapurāṇa. Przyluski (1933) countered Cœdès’ reading of the Trà Kiệu reliefs as the miracle of the hunchback, the merchants’ offering to Kṛṣṇa, and the breaking of Kaṃsa’s arch. He also suggested that the Trà Kiệu pedestal should be viewed in the direction of the right (pradakṣiṇa) to honour the divinity supported on

09_CreSouthV1_8P_11Feb22.indd 308

top. Phạm Hữu Mý (1995) has suggested that the images reflected the royal family of Bhārata from the Mahābhārata while Trần Kỳ Phương (2000) has argued that the pedestal depicts the wedding of Sītā and Rāma from the Rāmāyaṇa epic. Some of the figures can be confidently identified as related to Indian texts, but many cannot. Therefore, any monothetic hypothetical interpretations would unduly impose an identification on the figures. Without dismissing the possibility that the visual narrative may be connected to a particular episode of the Bhāgavatapurāṇa, the Mahābhārata or the Rāmāyaṇa epics, I focus on the Trà Kiệu pedestal and its reliefs as a visual document reinforcing royal or court identifications. The circulation of the narrative may be combined with an

11/2/22 3:48 PM

The Colossal Trà Kiệu Pedestal in Campā

309

oral tradition that used various Indic heroes to tell dating of the liṅga-yoni and the local stories. The reliefs display what to modern trà kiệu pedestal audiences would seem a disjointed series of excerpts from epic or sacred narratives, but were It is possible that the large liṅga-yoni still paired likely well understood by the Cams through their with the Trà Kiệu pedestal in the Đà Nẵng Museum of Cam Sculpture is dated earlier than the pedestal own cultural lens. itself. Trần Kỳ Phương argues that a 7th-century James Scott argues that ‘in the case of oral histories and narratives, the concept of “the orig- yoni from Hà Trung is similar to the form of the liṅinal” simply does not make sense. Oral culture ga-yoni on top of the Trà Kiệu pedestal (Trần 2018: exists and is sustained only through each unique 128). The Hà Trung yoni is elaborately decorated performance at a particular time and place for an with two rows of petals. The first row includes fourinterested audience’ (Scott 2009: 230). Seen from leaf petals with an inverted fleuron at the centre. On this perspective, the difficulty in locating the orig- the bottom of the Hà Trung yoni, there are addiinal or a single story for the carvings on the Trà tional carvings of floral, leaf, and spiral decorations. Kiệu pedestal is not surprising. The recourse to The petals are separated between another inverted Indian texts to comprehend Southeast Asian art is fleuron carved on the top rim. This fleuron motif found in most scholarly interpretations, but how also appears on the Trà Kiệu pedestal. A second, round yoni with the inscription of those texts were understood in Southeast Asia Prakāśadharma (C. 97) from Mỹ Sơn is also comis not certain. Even if there was some certainty about the identification of a group of characters parable stylistically to the liṅga-yoni of Trà Kiệu in terms of a specific text, we cannot know if the (Trần 2018: 128). The yoni is carved with an inlocal interpretation was consistent with the Indic scription on the centre of its body, which dates to textual one. As far as I know, there are no extant the 7th century (Fig. 9.4). The inscription states, Cam representations of the Kauṇḍinya myth or ‘this [kośa] for [the liṅga of] Vāmeśvara, installed life scenes of Kṛṣṇa and the visual representations by Sri Vikrāntavarman, the lion among kings, will depicted on the Trà Kiệu pedestal would be unique endure as long as the world exists’ (Golzio 2004: if we identify them as such. Based on inscriptions 29). The kośa is a metal cover for the liṅga that and other visual arts related to the Rāmāyaṇa in can be added and removed, and is thought to be Campā, there is sufficient visual evidence that the installed as an offering after lustration rituals (Guy Trà Kiệu pedestal depicts characters from the 1998, 2018: 89). This suggests that various things Rāmāyaṇa, but which specific stories and scenes were placed on liṅgas and the placement of objects could also change over time.8 Except for stylistic from the epic remains subject to debate. In addition to conflicting interpretations connections, there is little evidence that the inscripabout the story carved on the Trà Kiệu pedestal, tion of Prakāśadharma is related to the Trà Kiệu the scholarship is not conclusive about its date. pedestal and the liṅga-yoni. The base of the yoni Two time periods are proposed in the literature with inscription carries decorative designs of lotus for the object: early 7th-century or a much later petals, with one large single petal alternating with one, closer to the 10th century. For instance, Henri two doubled petals around the rim. Trần argues Parmentier and Trần Kỳ Phương date the pedes- that the ‘round-shaped yoni decorated with large tal to the 7th–8th century, while Jean Boisselier, doubled lotus petals was introduced as a new art Emmanuel Guillon, and Anne-Valérie Schweyer form in the late seventh century’ (Trần 2018: 128). place the pedestal in the 10th century. Using recent scholarship, archaeological findings, and careful 8. This is true for sculptures of different cultures includvisual analysis, I will argue that the Trà Kiệu ped- ing Greece, India, and China, where the original objects may have been clothed, painted, or surrounded by flowers estal dates closest to the 10th century. among other things for their intended purposes.

09_CreSouthV1_8P_11Feb22.indd 309

11/2/22 3:48 PM

310

Mya Chau

of central Vietnam, archaeologists discovered narrative reliefs depicting the Rāmāyaṇa on the foundations of a 10th-century temple, namely one of three temples at the archaeological site of Khương Mỹ. Cecelia Levin (2008) argues that the narrative reliefs from the Khương Mỹ temple have artistic connections to the Rāmāyaṇa reliefs in South India and Java.10 The foundation bases from the Khương Mỹ temples include scenes of the sighting and pursuit of the golden deer, the abduction of Sītā, and the confrontation between Rāvaṇa and Sītā (ibid.: 88). Levin sees architectural connections between Cam and Javanese temples. She states, ‘the assignation of Khương Mỹ to the early Classical phase of [Cam] art [10th century] was made on the overall basis of architectural ornamentation and particuFig. 9.4: Yoni Pedestal with the inscription of lar stylistic links shared with Classical Javanese Prakāśadharma (C. 97), Campā, 7th century. monuments’ (Levin 2008: 90). She also observes © Trần Ky Phương an association between Campā’s matrilineal society and Sītā depicted as a powerful figure, rejecting Rāvaṇa’s advances, at Khương Mỹ. Her observaLiṅga-yonis were constructed in different shapes tion of the importance of the matrilineal culture is and sizes with varying decorative motifs from the bolstered by Trần Kỳ Phương and Rie Nakamura’s 7th to the 9th centuries. There is a circular hole article about the dualist cult of Cam society. They in the yoni and the museum has recently inserted suggest that two sanctuaries, Mỹ Sơn and Pô Nagar, what it sees as the correct Trà Kiệu liṅga. However, ‘reflect certain characteristics of the cosmological the hole does not descend into the square base. dualism’ defined as a Cam male and female dichotWithout the liṅga-yoni, the base is an autonomous omy (Trần and Nakamura 2012: 208). Inscriptions and icons found at two Cam sites object, and there is no other Cam yoni with images suggest that from the 8th to the 13th century, Mỹ from the Rāmāyaṇa. On the basis of the above considerations, I Sơn was built for the God Bhadreśvara (benevsuggest the pedestal should be studied separately olent Śiva/Mountain/Father) and Pô Nagar for from the liṅga-yoni, without excluding the possi- the Goddess Bhagavatī (Pô Nagar/Sea/Mother) bility that another deity or multiple objects may (ibid.: 274). The authors conclude that Mỹ Sơn and be originally paired with the decorated base. A Pô Nagar symbolized the female and male realm, 9th–10th-century date is the most likely as the respectively, thus revealing an emphasis on Cam subject and decoration of the reliefs can be com- cosmological dualism. In volume two of this edited pared to prominent 10th-century artworks related collection, Roy E. Jordaan discusses the possibility to the Rāmāyaṇa in Java, Cambodia, and Campā.9 for the representation of Sītā as Rāvaṇa’s daughter During a 2001 excavation in Quảng Nam province at Candi Prambanan. What is clear is that both cultures in Java and Campā had local expressions and an emphasis on lineage, kinship, and parent9. Unlike the modular 7th-century Mỹ Sơn pedestal, comage, all of which reflect Southeast Asian traditions prised of several small blocks that form a small cella inside practised in the 9th–10th centuries. In addition, a temple, what is highly distinctive about 10th-century Trà Kiệu pedestal, is that it consists of a single raised pedestal with elaborate figurative scenes. Excavation reports show that the pedestal was found split into two parts, and it was reassembled in the Tourane outdoor garden and the Đà Nẵng Museum of Cham Sculpture.

09_CreSouthV1_8P_11Feb22.indd 310

10. On the latest scholarship regarding the Rāmāyaṇa in mediaeval Indian art, see Julie Romain’s dissertation (Romain 2015).

11/2/22 3:48 PM

The Colossal Trà Kiệu Pedestal in Campā

311

Levin has theorized some connections of artistic elements such as temple motifs and foliage found in Campā and Java, which may have a common origin in China (Tu 2018: 84). Following these suggestions, the artistic relationships to consider between Campā and related cultures would include Cambodia and China, as well as Java.

Javanese, as well as less explored areas such as similarities between their military and courtly cultures. Intriguing evidence for peaceful diplomatic interchange is found in the 10th-century Nhan Bièu inscription from Campā that records the historical connection between Java and Campā in the form of a royal officer’s two expeditions to Java. The inscription states,

inscriptions and historical records

He was a favorite captain of the king Śrī Jayasiṃhavarman and had riches equal to his desires. He was named Pov Kluñ Piliḥ Rājadvāra. At the command of the king he went to the capital of Yavadvīpa on a diplomatic mission and obtained credit by the success of his undertaking… Again, at the commands of the king he went to Yavadvīpa a second time and was successful in his undertaking. (Golzio 2004: 112–113)

The interaction between Campā and Java is mentioned in inscriptions, which record, for example, Java’s sea attack against the Cams in the 8th century. An inscription dated to 787/788 CE reads, Then owing to the excess of faults in the Kali Age it (i.e., the temple […]) was burnt by the army of Java coming by means of ships, and became empty, in the year of the Śakas denoted ‘nine, ambara, adri’ [709]. (Golzio 2004: 43)

While we do not know the journey’s objectives, they were considered successful, and it is assumed that Peter Sharrock has noted that the relationship they would have introduced new cultural and rebetween the Cams and the Khmers has ‘too often ligious knowledge to both the Javanese and Cam been characterized as continuous warfare. A close courts. Chinese historical records indicate that the look shows a succession of friendships, and miliCams exchanged Javanese textiles with China. For tary, economic and cultural alliances’ (Sharrock et al. 2018: 111). Similar to the interaction between example, textual sources document that in 963 CE, Campā and Cambodia, there is little detailed an envoy from Campā brought to China differhistory of the nature and dating of the relationships ent types of Javanese textiles such as ‘gu thin silk between the Cams and the Javanese. However, there from the country of Ma-li-yan-luan in Java’, and is evidence of contact, both military and economic. ‘gu thin silk from Sha-wan in Java’ (Wade 2011: Arlo Griffiths states that ‘the ancient name 161). Whether or not these were actual Javanese of the island of Java, and its identifiability with textiles cannot be verified, but what is important various similar-sounding names known from to stress is that the Chinese accepted the textiles as ancient Southeast Asian, Chinese and Arab sources, Javanese. The documents indicated that the Cams, is an old problem, compounded by the fact that the Chinese, and the Javanese operated within the similar-sounding and obviously related names in international circle, either directly or indirectly some modern mainland Southeast Asian languages during the 9th–10th centuries. The Cams prerefer not to the Javanese or to Java in particular, sented the Chinese with Javanese textiles, among but to Malays or the Malay world in general, or to other trade objects, while likely receiving military 11 localities in northern Laos’ (Griffiths 2013: 49). He weapons and horses in exchange. Faxian’s account of the sea passage from India suggests that Khmer inscriptions refer to the island to China in A Record of Buddhist Kingdoms (399– of Java. Griffiths concludes that there were ‘inter414 CE) indicated travels to Java-dvīpa (Java). His national political relations between Khmer, Cam passage reads, and Javanese polities in the late 8th and early 9th century of our era’ (ibid.: 76). The visual and textual evidence has not been comprehensively analysed, 11. Ibid.: 11. The Chinese refer to these as tributes, but but both type of sources will shed some light on such encounters were never one-sided and the Cams would always receive something in return, particularly goods and the warfare exchanges between the Cams and the other textiles.

09_CreSouthV1_8P_11Feb22.indd 311

11/2/22 3:48 PM

Mya Chau

312

After proceeding… more than ninety days, they arrive at a country called Java-dvipa, where various forms of error and Brahmanism are flourishing, while Buddhism is not worth speaking of. After staying there for five months, Faxian again embarked in another large merchantman, which also had on board more than 200 men. They carried provisions for fifty days, sailing on the 16th of April, 414. (Legge 1965: 113)

In another inscription, the Mỹ Sơn Stele Inscription of Prakāśadharma, dating to 656 CE, the king describes himself as Constantly devoted to the Brāhmaṇas, the gods among men, (he was) gracefully attended by enemies subdued by his own prowess; (he was) the cause of Lakṣmī pride; Śrī Lakṣmī, in the fond hope that he was Rāma, son of king Daśaratha, dutifully followed him, and this was very worthy of her indeed! (Golzio 2004: 20)

While Faxian’s stay in Java was due to a storm that drove his ship off course, Java and Sumatra Here, the inscription places the Cam king on the became major destinations for Chinese monks in same level as Rāma from the Rāmāyaṇa. The posithe 6th and 7th centuries to stop en route on their tion of king was often seized by usurpers through voyages to India and study Buddhist texts and learn military prowess. Thus, the heroic Rāma, who was Sanskrit. The doors were open wide for cultural able to defeat all his enemies and rivals, was an ideal interchange. role model in Cam society. The kings often wrote Furthermore, Campā and Java are connected of the ideal leader, frequently invoking the quintesthrough their cultural appreciation and shared sential heroic prince, Rāma. A 12th-century Khmer artistic iconography for the sacred epic, the inscription compares the Cam king to Rāvaṇa, the Rāmāyaṇa. The visual and textual material in the defeated enemy of Rāma. The inscription reads, Rāmāyaṇa reinforces kingship and ‘the develop- ‘Sri Jaya Indravarman, king of [Campā], arrogant ment of the notions of political authority in South- as Rāvaṇa… having an army led by chariots, went east Asia, most notably in the ancient—and emi- to the country of Kambu equal to the heavens in nently “Southeast Asian”—conception of the king order to fight’ (Thompson 2016: 130). Both Khmer as a “man of prowess” endowed with superhuman and Cam inscriptions indicate that the Rāmāyaṇa abilities and worshipped by his entourage in terms was well known by the 7th century and continued of bhakti relationships’ (Acri 2017: 10). There is to be mentioned in later inscriptions. ample evidence that the Rāmāyaṇa was celebrated in Campā. A Cam inscription dating to the 7th century reveals the worship and divine nature of military and temple culture Vālmīki, th