Michael Willis, The Archaeology of Hindu Ritual: Temples and the Establishment of the Gods

presented in the previous chapters and provides a new chronological framework for the changing patterns of trans-Indian Ocean trade over the 700–800 years up to the sixth century CE. In summary, this is a timely book which is rich in detail and clearly set out and expressed. It will be essential reading for all students of eastern Mediterranean history and of Indian and Indian Ocean studies. It is rather more a book for the specialist than for the general reader, who may be a little confused by the complexity of abbreviated ceramic terminologies – RPW, LR1, Dressel 2–4, torpedo, RW, CRSW, and so on – although they are all given their full names when they first appear. The significance of the research could well have justified a more substantial volume with more and higher quality illustrations, although the ones included, mainly maps, line illustrations, and listings, are good enough for their purpose, and a more substantial book would have cost a lot more, which would have put it out of the reach of most students and scholars.

This is an extraordinarily ambitious and exciting book. Both in terms of the questions it seeks to answer and the manner in which it assembles evidence, it is innovative and provocative, and points towards a mode of doing historical studies that we must furthermaking space for the possibility of contact among literary, epigraphical, art historical, and archaeological materials in our analysis of the past. The particular context that Willis has chosen to explore is northern India in the period of the imperial Guptas, of the fourth to sixth centuries; he seeks not only to trace the emergence of new religious practices and institutions in this eraas indicated by the book's subtitlebut to show how these were intimately linked to the activities and the ambitions of the kings of the Gupta dynasty.
The book is divided into three lengthy chapters. The first focuses on Udayagiri, a rocky hill in central India where a number of shrines and sculptures have been carved into the stone of the site over many centuriesincluding the famous relief panel of Viṣṇu in his incarnation as Varāha, the boar, rescuing the goddess Earth. With so few Gupta period inscriptions here (there are only three), much of Willis's analysis depends upon an examination of the site itself, especially the orientation of the cave temples and relief sculptures with reference to seasonal astronomical phenomena, as well as the interpretation of the iconography of the relief sculptures. Willis contends 'that Udayagiri was reworked under Candragupta II (circa CE 375-415) to articulate a revitalised form of early Hindu kingship in which the ruler was envisaged as a paramount sovereign (cakravartin) and supreme devotee of Viṣṇu (paramabhāgavata)' (pp. [2][3]. As the site of the Gupta's royal consecration ceremony (rājasūya), of the memorialising of that solemn Vedic sacrifice, and of the annual celebration of the rainy season festival for Viṣṇu, it was a ritual centre critical to the constitution of 'Gupta kingship and the socio-religious culture of the Gupta age' (pp. 10-11, 67). Willis argues that Candragupta II is represented in the relief sculpturesas a devotee kneeling below the sculpture of Viṣṇu Nārāyaṇa asleep on the coils of the serpent Ananta (p. 35), as the devotee worshipping Varāha, and as the god Varāha himself (pp. 58-59, 64-65; Varāha is also said to represent Candragupta's father Samudragupta). Willis suggests that Vaiṣṇava imagery was further mobilized, and linked to the image of the Gupta king, in a temple at Udayagiri, now entirely ruined, which he is certain enshrined Viṣṇu's footprints as an important object of worship (pp. 74-75). Vaiṣṇavism as the 'official religion' (p. 71) of the state was expressed at Udayagiri through the world-possessing, world-protecting, and cosmological imaging of Viṣṇu, with the representation of the Gupta king's intimate relationship with Viṣṇu as a public declaration that these rulers 'stood close to the supreme power that guided the world' (p. 78).
The character of Gupta kingship continues to be a theme in the second chapter, where the focus shifts to a consideration of the role of the ruler as granter of land to Brahmans and templesa role that is part of the dharma of kings but also one that allows for the king to expand his authority through the establishment of new estates in the hinterlands, benefiting his treasury, developing a strong political network, and encouraging 'a stable and complex social structure (i.e. the varṇa-system)' (p.159). But the central argument in this chapter concerns the establishment of Hindu deities in the 'public' context of the temple and the institution of forms of worship in that setting. Willis maintains that temple pūjā arises as the result of the transfer of Vedic rituals from the domestic setting to the temple environment. 'Priests have carefully and deliberately moved sacrifices from the domestic environment to the temple and attracted funding to support these rites in the new location' (p.111). Temple pūjā was 'modelled on the reception of guests in the domestic setting'as described in the Vaikhānasasmārtasūtra and gṛhya textsand it is the priests of the Gupta period who were responsible for this development (p. 112). The Gupta period is thus witness to 'a new institutional reality: a deity established in a temple controlled by a priest holding endowments endorsed by the state' (p. 119).
The main discussion of the role of priests is in fact found in the third chapter, where the temple context is very little in evidence. The focus is, rather, the Gupta court, which, in Willis's view, was dominated by the 'spiritual triumvirate' of the royal chaplain (purohita), sacrificial priests (ṛtvij), and spiritual guide (ācārya), each of whom 'played a part in the constitution of the king's religious and political identity' (p. 8). Despite the extreme paucity of references to these figures in the epigraphical record, Willis finds it possible to conclude that the Gupta kings' sacrificial priests would have been Maitrāyaṇīyas, and that their ācāryas belonged to the Bhāgavata sectarian group. Brahmanswhether of an orthodox Vedic stripe or serving as preceptors and temple priestsmay be said to have been co-opted to the Gupta imperial project (p. 245). But in fact it was they who actually had the upper hand, especially those priests 'whose special knowledge ... [and] vision transformed the religious and political life of India', as Willis tells us in the book's final sentence (p. 246). And this seems to return us to what the book's 'Introduction' has promised: 'proof' in the pages of this book that 'religious ideology and religious institutions structured the political and economic relationships of the Gupta age rather than vice versa' (p. 9).

Methods and models
Willis's effort in this book is to depart from 'the static, desk-bound forms of analysis that have so far governed the study of Indian inscriptions, sculpture, built environment, and landscape' and instead to 'place literary sources "on the ground" in actual places and specific religious, political, and ritual contexts'with Udayagiri as 'the crucible in which I forged this method and ... from Udayagiri ... I have taken it across early Hindu India' (p. 3).
The main sources for Gupta history are stone inscriptions and copper-plate charters. There are two major published collections of these, both of whichconfusingly enoughare counted as volume 3 in the series Corpus Inscriptionum Indicarum (CII). The first CII 3 appeared in 1888 under the editorship of J. F. Fleet, and is entitled 'Inscriptions of the Early Gupta Kings and their Successors'. Fleet has edited and translated a total of 81 inscriptions in this volume, of which 23 are 'Gupta inscriptions' in the sense that they commence with a eulogy of a Gupta king or record an act or an order of the king; the remaining inscriptions in the volume are assigned to Parivrajakas, Vakatakas, and other rulers of the early medieval period. Starting in 1935, D. R. Bhandarkar began the project of re-editing Fleet's volume, and incorporating Gupta inscriptions which had been discovered since Fleet's time, to produce an updated CII 3. Bhandarkar excluded all those inscriptions that could not with certainty be considered 'Gupta inscriptions', but otherwise made few changes to the readings and translations produced by Fleet; he also composed a lengthy introduction, covering Gupta cultural, social, and religious as well as political history. After Bhandarkar's death in 1950, the project was carried forward by G. S. Gai and B. Chhabhra, and the revised CII 3 finally appeared in 1981.
The total number of Gupta inscriptions (excluding seals and coins) that are found in one or the other of the two volumes of CII 3, or in a few cases in other sources upon which Willis has drawn, is 43. Nine are copperplate grants, of which five were found in the village of Damodarpur in West Bengal. Eleven are inscriptions on images, including six Buddha images and three Jain images. Twothose inscribed on the Allahabad pillar and the Mehrauli iron pillarhave been moved from their original locations. Only eight of the Gupta inscriptions appear to be in the exact place where they were first inscribed: three at Udayagiri, one at the Great Stupa of Sanchi, the Junagadh Rock Inscription, and three pillar inscriptions which are presumed to be associated with temples of which little or nothing remains. Finally, there are 13 Gupta inscriptions on pillars or stones which may be at the same site where they were originally engraved, but which have been displaced, buried as rubble or reused in other building projects. Given that one of the (very commendable) aims of the book is 'to reconsider the wider archaeological context of inscriptions' (p. 5), Willis has a problem. It may be possible at Udayagiri to situate the text of an inscription 'in the ritual landscape for which it was written' (p. 4), but even here, the contents of the inscriptions provide little aid in developing the meanings of the site that Willis has in view, considering that none of the three inscriptions concerns an image or shrine of Viṣṇutwo record the excavation of cave temples dedicated to Śiva and one records the setting up of a Jain image.
In the virtual absence of 'on the ground' archaeological contexts for the Gupta inscriptions, Willis has employed alternative approaches to the task of reconstructing the 'ritual landscape' of the Gupta period. One is to examine the inscriptions issued under the names of kings in territories neighbouring that of the Guptas, such as those published in Fleet's CII 3. The relationship of the 'subordinate' rulers to the Gupta king is presumably illustrated by the inscription in Cave 6 at Udayagiri, where the Sanakanika ruler is said to meditate on the feet (pādānuddhyāta) of Candragupta (p. 57, see also p. 71).
But we find this expression as a genealogical epithet (e.g. Kumaragupta is said to have meditated on the feet of his father Candragupta in CII 3(Fleet).13, CII 3 (Bhandarkar).31); so the Sanakanika king may be saying that he regards Candragupta as his father, rather than using this phrase to express an attitude of reverence toward the divine and thus an understanding of the Gupta ruler as a god, as Willis argues (p. 58). 1 'The key question is whether subordinate kings were effectively independent... [or] are their activities to be seen as a useful gauge of the ideology and culture of the imperial court', and Willis strongly favours the second option, maintaining that subordinate rulers 'conformed to court protocol and religious ideology' and 'replicated and reinforced the norms' of their Gupta overlords (p.157). Thus, in what seems something of a circular argument, the records of these kings may be used as evidence for cultural and religious developments that were initiated and influenced by the Guptas.
Another way in which Willis seeks to elucidate the 'ritual landscape' of the Gupta period is through an examination of non-inscriptional literature of various kinds. The argument for the relevance of this literature to the Gupta context is made in a rather extreme waya truly surprising number of works are said to have been composed, redacted, or circulated in Gupta times and in the Gupta court. Among these are the present version of the Mahābhārata and its Nārāyaṇīya section (from the fourth or fifth centurypp. 224, 335, n. 299), the Arthaśāstra and Nītisāra (both of the fourth centuryp. 62), Manusmṛti (fourth centuryp. 205), an early version of Yājñavalkyasmṛti (p. 206), the Nāṭyaśāstra (p. 283, n. 70), the Viṣṇudharmāḥ (p. 31), and probably a number of the Purāṇas (p. 79). And if there are texts that clearly do not emerge from a northern Indian context, as in the case of the Vaikhānasasmārtasūtra, then they still can be used as 'evidence by analogy' (p. 118). Works like these seem to be cited more often and quoted more extensively than are the inscriptions, and more often than not appear to have priority over the inscriptions. Although the inscriptional texts are in meagre supply and, even when preserved in their entirety, are often terse and obscure in their meaning, it seems a bit dangerous to understand them through the lens of the literary texts.

Issues and absences
In his introduction, Willis informs us that 'from the midfourth century, kings granted land to two types of institutions: (a) tax-free estates held by brahmaṇas learned in the Veda and (b) tax-free estates held by or for temple gods' (p. 7). This is misleading in two ways: firstly, because virtually none of the Gupta inscriptions record royal grants (although some of them are royal orders confirming the transfer of land purchased by others) and, secondly, because gifts made to Buddhist and Jain institutions and shrines are more in evidence in the Gupta corpus than are gifts to temples dedicated to Viṣṇu and Śiva. In terms of this second issue, although the focus of the book is on Hinduism, it seems strange to treat ritual and institutional developments in isolation, ignoring the shared religious cultureand even shared religious space, as at Udayagirithat is evinced by the Gupta inscriptions. Willis indicates that Jainism was marginalized in Gupta times, at Udayagiri being 'relegated to a cluster of caves at the edge of the northern hill' (p. 231). And he characterizes the encounters between Buddhists and Hindus in terms of conflict, involving 'not only competition for patronage but a profound clash of institutions, beliefs, and social ideals' (p. 290, n. 129).
But the commonalities with respect to modes of worship and understandings of imagesand particularly in the matter of establishing permanent endowments for religious institutionsseem in fact to outweigh the contentions. Gregory Schopen's examination of passages in the Mūlasarvāstivāda-vinaya which treat the donation of perpetuities to Buddhist monasteries for their maintenance (and the associated issues of contracts and loaning on interest)and the terminological and conceptual affinities with inscriptional evidence from as early as the first century and with the Dharmaśāstra literaturehas led him to conclude that the Vinaya text was redacted in a Buddhist community 'deeply embedded in the larger Indian, brahminical world' at a time when 'a new type of social institution' was emerging -'the fully institutionalized, permanently housed, landed monastery'. 2 In the book under review, a comparison of Gupta inscriptions that are linked to Buddhist (and Jain) worship and institutions with their 'Hindu' counterparts might have provided us with a more 'on the ground'sense of the relations among these communities.
'The establishment of the gods' (pratiṣṭhā) was not only a ritual action undertaken by Brahman priests (as Willis suggests, p. 168), but by heterodox and Hindu lay donors. 3 Another aspect of the shared religious culture of Buddhists, Jains, and Hindus is the use of certain materials in worship. Willis provides a learned and fascinating account of the meanings of bali, caru, and sattraas temple offerings mentioned in Gupta period inscriptions whose origins lie in the domestic Vedic contextbut he does not speculate on why these elements are so frequently combined and intermingled in worship with flowers, incense, perfumes, lamps, and so on, and what the origins of these offerings might be. Friedhelm Hardy is someone who has taken up this question, and describes these elements of sensuous worship as part of a 'folk religion' which is documented in early Buddhist and Jain texts, as well as in the Harivaṃśa; Hardy argues that there was a conscious effort on the part of Hindu sectarian religious elites to incorporate these offerings into orthodox worship. 4 Willis seems to provide an opening into a discussion of the commonalities among the earliest forms of shrines and of 'public' worship in his analysis of a number of relief sculptures, including ones from Amaravati and Sanchi (pp. 114-16, 194). This is a very welcome inclusion of art historical materialsof which, given Willis's background and expertise, one would have expected a good deal morebut does not give rise to a discussion of how the activities at a Buddhist site (with its tree, shrine, and seat engraved with the Buddha's footprints) might differ (or not) from those at the Hindu site (tree, shrine, and altarfor Vedic offerings?).
Finally, there is the question of the priests and preceptors of Gupta times, who loom exceedingly large as men of influence and agents of change in Willis's analysis, but can scarcely be found in the inscriptions. There are exactly two men who can be identified as temple priests, as a result of our finding in sixth-century copperplate charters their petitions for land grants that would allow them to perform services (balicarusattra), make offerings, and carry out repairs to the temple (CII 3 (Bhandarkar).47 and CII 3(Fleet).25 and 29). It is hard to regard these figures of the late Gupta period, apparently residing in the hinterland, as 'key ritual actors of the Gupta age', responsible for creating such cultural products as 'the living images of early Hinduism and the supporting structure of sacred lore and religious practice' (p. 168). The inscriptions mentioning men termed ācāryas are a little more numerous, although there are none that refer to the Gupta royal preceptor. So to understand the role of the preceptor in Gupta times, Willis is compelled to turn to a set of inscriptions engraved on three Jain images commissioned by the Gupta ruler Ramagupta according to the instruction (upadeśāt) of a Jain monastic teacher (CII 3 (Bhandarkar).5). Willis considers that the word upadeś a refers to 'the special instructions the preceptor imparted to the king in the course of his initiation into the worship of the images' (p. 231). But the instruction of a Jain ascetic to his lay disciple to set up an image is commonly met with elsewhere, although the word upadeśa is not usedinstead we find nirvartana used in early inscriptions from Mathura, 5 and in Gupta times a woman is instructed (prajñāpita) by her preceptor to establish a Jain image (CII 3(Bhandarkar).18)suggesting that what is at stake is more pragmatic and mundane than the 'transmission of sacred knowledge' (p. 232). Willis nonetheless expands on this interpretation and concludes that 'a salient feature of the ācārya's religious role was his control of images and the relationship of devotees with those images' (p. 235) in the Hindu context, echoing the title of the book and underscoring the significance of priests and preceptors in the age of the Guptas.
One may with pleasure and interest take note of the points of interpretation and the synthetic arguments presented by Willis, even if one disagrees with him. But too often one is brought up short by a gratuitous attack on the views of one of Willis's forebears or colleagues in the study of Gupta materials. Joanna Williams and Sheldon Pollock are targets, but nothing can match Willis's hostility toward D. R. Bhandarkar (on the other hand, Fleet is described as a man of genius). There are quite a number of straw men with whom Willis wants to pick a fight, and his categorical statements and tone of certitude are at times quite off-putting. These lapses seem out of place in a book of such grand scope and reach.
How much has Willis succeeded in finding a place 'on the ground' for the Gupta rulers and priests, and the political culture and religious worlds to which they were (he claims) responsible for giving rise? The task Willis has set for himself is clearly an impossible one, and the questions he asks are the Big Questionstoo vast and complex to be answered by any single scholarbut one can only admire and applaud him for aiming so high. The success of his bookinspiring, important, challenging, and at times maddeninglies in its undoubted ability to stimulate discussion, to make us revisit received ideas and to review the primary sources for ourselves. And for this we are indeed indebted to Michael Willis. deathwhich receives much discussion in Willis's book, is too large an issue to treat here. But I must take exception to the useas 'supporting evidence' for the identification of the king with Viṣṇu in the form of Varāhaof the royal titles of the Pallava ruler Narasimhavarman inscribed on the base of a Varāha sculpture at Mamallapuram (p. 64). To regard these titles as 'descriptive labels', identifying the deity as the king, is a grave misunderstanding. It is a shame that Willis makes so little use of the inscriptions of the early Pallavas (conveniently collected in T. V. Mahalingam, Inscriptions of the Pallavas (Delhi, 1988)), who were the contemporaries of the Guptas in South India. Surely a comparison of the patterns of patronage and royal eulogy among the Pallavas of the fourth to sixth centuries with those of the Guptas would have illuminated what was distinctive in the Gupta context and moderated the image presented by Willis of the Gupta origins of so much of India's