Invoking the Goddess along the Southern Silk Roads: a transregional survey of Prajñāpāramitā protective texts

Abstract

This article brings together various textual materials relating to the worship of Prajñāpāramitā in mediaeval Monsoon Asia. In particular, it covers sources from South Asia, mainland Southeast Asia (Angkor), maritime Southeast Asia (Java and Bali), and Southern China (Yunnan). The aims of this critical survey are twofold. On the one hand, the article argues for the importance of ritual language within wider discussions on cultural and linguistic cosmopolitanism within and beyond the so-called ‘Sanskrit Cosmopolis’. These discussions have too often focused on royal epigraphic eulogies, neglecting religious literature. On the other hand, it highlights the key role of Prajñāpāramitā as a protective deity, and the transregional heritage of her invocatory texts. These texts strike one for their traditional choice of imagery and associations, thus reminding of the importance of exoteric traditions within the history of Southeast Asian Buddhism.

Prajñāpāramitā traditions after the eighth century: looking eastwards

Scholars working primarily on South Asian Prajñāpāramitā texts might expect to find copious evidence for such scriptures transmitted or redacted in Southeast Asia, or at least various references to such primary sources. However, this would appear to not be the case. A reference to a *Mandrasena allegedly active in Funan, who translated the Saptaśatikā into Chinese;¹ a Cambodian inscription mentioning the Lakṣagrantha Abhiprajñā;² a Javanese chronicle telling of the king's interest in a Subhūti Tantra;³ or indeed a manuscript case from the same region engraved with the words bharālī prajñāpāramitā ⁴ are all evidence that Prajñāpāramitā scriptural sources in Southeast Asia are far from abundant and often highly problematic. There are of course references to the exchange of Prajñāpāramitā manuscripts as part of diplomatic relations and some sources in Old Javanese, such as the Sutasoma, offer rare and important references to the ‘mother of Buddhas’ (these references will be discussed below).

Instead, what is generously available to scholars are images of the Goddess Prajñāpāramitā, attested from virtually every region of the Sanskritic world, as pioneering surveys by Conze had already started to show.⁵ Some of these images can be linked to royal practices such as the funerary rites of queens, as with the famous cases of Jayavarman VII's Jayarājacūḍamani or of Kṛtanāgara's Rājapatnī. The often idiosyncratic iconography—hardly reminiscent of Pāla standards—offers fascinating examples of multi-armed bronzes or powerful images of a goddess sunk in meditation and wrapped in intricately woven garments.⁶

It is therefore no surprise that the burden of reconstructing a cultural history of Prajñāpāramitā in Southeast Asia—or indeed within the ‘Sanskrit Cosmopolis’—has generally fallen on art historians. Of course, beyond the technicalities of iconography, art historians are methodologically well positioned to write compelling cultural histories, given their awareness of archaeological contexts, patronage and ritual dynamics, and at least the most relevant textual sources. The growing number of studies in this area include Reichle (2007) (Nusantara),⁷ Conti (2015), Multzer O'Naghten (2016), Chemburkar (2022), and Kim (2022) (Mainland Southeast Asia).⁸

In this article, I collect and discuss evidence for a third kind of source that in my view lies somewhat in between the two typologies outlined above, namely scriptures and images. I am concerned with textual sources that function as invocatory texts and are hence particularly close to the personified form of the Goddess and her worship. These can be epigraphic hymns, single dhāraṇīs on various material supports, or stotras and stanzas found across various collections.

Before examining such evidence, however, one needs to address the wider issue of the conceptualisation and characterisation of Prajñāpāramitā's role in post-700 CE Monsoon Asia.⁹ The conceptual framework that has gained most traction among scholars is that of the complex relationship between Prajñāpāramitā and Buddhist ‘esotericism’. Jinah Kim has been instrumental in developing this approach, primarily in relation to South Asian material, whether it be illuminated Pāla manuscripts of the Aṣṭasāhasrikā depicting Vajrayāna deities or indeed passages from esoteric scriptures that embed Prajñā within their rituals, such as those of the Guhyasamāja.¹⁰ She has recently addressed the complex interaction between Prajñāpāramitā and esotericism with respect to the Cambodian Buddhism of Jayavarman VII, building on the work by Multzer O'Naghten as well as Peter Sharrock.¹¹ This approach reminds of the one taken by Green, with much justification, when describing the adaptations of Lokeśvara traditions in the Cham domain, again under the influence of Esoteric Buddhism.¹² A similar approach would also be justified with respect to the history of Prajñāpāramitā in Java. As shown by Acri, Kṛtanāgara's Buddhism was deeply esoteric in its ritual as well as conceptual outlines.¹³ Indeed, the main Javanese texts relevant to the current discussion, such as the Sutasoma, the Advayasādhana, or the Saṅ Hyaṅ Kamahāyānikan, are all conversant with an Esoteric Buddhist milieu, however broadly defined. It would therefore seem reasonable to approach the history of Prajñāpāramitā in the ‘Cosmopolis’ as a chapter in the book of Cosmopolitan Esoteric Buddhism or indeed the circulation of Esoteric Buddhism across Monsoon Asia.

This approach has many merits, not least in bringing together scholars with an interest in Esoteric Buddhism regardless of the specific region of expertise. It could nonetheless be questioned whether the label ‘esoteric’ has now become so diluted as to have lost much of its substance, or whether occasionally it may have even become misleading. The classical example is that of Borobudur. When looking at this massive complex or walking along its finely decorated galleries, do we have the impression of visiting a ‘maṇḍala’ or a ‘stūpa’? I personally would tend towards the latter option, keeping in mind that certain features, such as the choice and disposition of certain deities, do indeed add another layer of meaning—one that is certainly in line with the notion of ‘maṇḍala’.

With respect to the Prajñāpāramitā sources presented below, one may equally question whether they are ‘esoteric’ in any meaningful sense. How many arms on a Prajñāpāramitā statue are needed to make it ‘esoteric’? What kind of mantras or technical terms within a stotra? After reviewing the evidence, I think most of the Prajñāpāramitā-related passages, hymns, spells, and stotras available are instead rather ‘exoteric’ in character. This brings us to a terminological conundrum. Terms such as ‘proto-esoteric’ or indeed its counterpart ‘Dhāraṇī Buddhism’ may have some use, but they have rightly fallen out of favour among scholars. The first assumes a teleology that is often unattested, while the latter attempts to capture and isolate a form of Buddhism that would have been hardly institutionalised.¹⁴ Perhaps more interesting are terms such as ādikarma or ānisaṃsa Buddhism. The former may refer to ‘foundational practices’, often enjoined for beginner practitioners, whereas the latter—which need not be restricted to Theravāda—captures the quest for merit making and protection.¹⁵ The extent to which this exercise in labelling might be useful is of course debatable. In terms of textual evidence, in around the eight century, we see the appearance of shorter Prajñāpāramitā scriptures that often include ‘spells’ (i.e. dhāraṇīs and mantras) and the wider circulation of Prajñāpāramitā invocatory texts.¹⁶ Many of the recurring themes are rather traditional: Prajñāpāramitā represents the ‘mother of Buddhas’, ‘emptiness’, ‘non-duality’, as well as a source of wisdom and insight. The task then is first to assess these minor textual sources and then to discuss the cosmopolitan dimensions that they may showcase.

The Angkorian invocations

The first point about the history of Prajñāpāramitā at Angkor is that it is not confined to the reign of Jayavarman VII, his Buddhist revival, or his project of imperial symbolism (in which the goddess motif plays a prominent role).¹⁷ While associations with a more pronounced Esoteric Buddhist milieu might have taken shape in the twelfth century, and while the actual term prajñāpāramitā may first appear in inscription in this period—epigraphic reference as well as images may well go back to the eighth century,¹⁸ becoming more prominent during the ‘Buddhist turn’ of the tenth century.

Despite this long trajectory, reference to scriptural sources or the Prajñāpāramitā exegetical tradition are virtually absent. The two notable exceptions are the famous Vat Sithor inscription (K111),¹⁹ informing us of the important figure of Kīrtipaṇḍita and his intellectual interests, and the Sab Bāk inscription of Phimai, in modern-day Thailand.²⁰ As is well known, the former mentions an unidentified Abhiprajñā text in 100.000 units, while the latter refers to Prajñāpāramitā in connection with learning.

This lack of scriptural sources is counterbalanced by the rich art historical evidence. While the best-known images of Prajñāpāramitā are the multi-armed bronzes or the possible ‘likeness’ of the Imperial mother,²¹ the iconographic evidence is of course much more complex, including images on caityas, lintels, and clay sealings. The iconographical association of Prajñāpāramitā with various deities has been discussed in detail by art historians, including those with Lokeśvara, Vajrapāṇi, Amitābha, Vairocana, and Hevajra. Some of these associations bear esoteric connotations and help throw light on the character of Angkorian Buddhism in around the turn of the first millennium.

What interests us here are further epigraphic references to Prajñāpāramitā as a deity. A recent overview article by Chemburkar goes into some detail in this regard. It is clear that the available materials are rich and diverse, and cannot be forced into a single category. For instance, there are a number of inscriptions in Old Khmer that mention the installation of Prajñāpāramitā alongside other deities, such as K168, which mentions the goddess as well as the eleven-headed Lokeśvara.²² Other mentions of the installation of the image are those of K696 from the Southern gopura of Bantey Chmar and K266 from one of the towers of Bat Chum (in Sanskrit, using the term divyādevī).²³ Another typology of reference is when certain rulers or teachers are said to be devoted to Prajñāpāramitā, as with the unpublished inscription K1297—speaking of royal bhakti towards Prajñāpāramitā—or with K772 from Prasat Beng, mentioning a Kavindrācārya who ‘upheld’ the Perfection of Insight (prajñāpāramitādhārī).²⁴

The third typology is that of the devotional hymn. Such hymns often open Khmer inscriptions and can be typically understood as a request for protection and benediction over a specific enterprise. For example, K214 from Bantey Neang offers homage to the goddess, characterising her as one who bestows wisdom.²⁵ In the case of K872 from Beng Vien, we also find the term jinasantānakāriṇī (a paraphrase for ‘mother of Buddhas’), as well as other expressions that style Prajñāpāramitā as a mother and luminous creator.²⁶ The theme of motherhood is conveyed poetically in K273 from Ta Prohm (bhaktyā jinānāṃ jananīṃ namadhvam),²⁷ with the same text also appearing on K908 from Preah Khan.²⁸

Even more interesting are those hymns that clearly mention the protective character of Prajñāpāramitā. K225 from Thma Puok mentions the theme of motherhood in connection with a prayer for protection, expressed with the Sanskrit root pā.²⁹ The Vat Sithor inscription contains a similar stanza, employing the term tārī, or ‘protector’.³⁰ Similarly, one of the inscriptions from the Bat Chum towers plays with the alliteration between pātu ‘may [she] protect [us]’ and pātakāt ‘from grave offence’.³¹

I am currently not aware of other Khmer inscriptions that explicitly mention the notion of protection with respect to Prajñāpāramitā. However, two more epigraphs are worth mentioning here. Scholars who study Prajñāpāramitā in the Khmer domain often mention the Vat Kdei epigraph, as this document paints a picture of how a Prajñāpāramitā statue might have been used on the ground. The image of the deity and that of Lokeśvara are bathed into a pond. It is understood that the water from this pond, blessed by the ablution of the images, would have served agricultural functions, as hinted at by the list of fields (kṣetras) that follows.³² The other interesting epigraph is a caption inscription on the pedestal of a multi-armed goddess. It reads vrah rūp(a) vrah prajñāpāra[mi]tā,³³ with vrah usually functioning as a counterpart to the Sanskrit śrī in Old Khmer. This pedestal inscription does not appear to have functioned as a mantra or invocation, unlike the rare case of a spell engraved on the back of a Lokeśvara statue (K1154).³⁴

To sum up, despite the lack of scriptural and commentarial sources, and despite the fact that epigraphic hymns may offer only a limited sample of technical notions and terminology, there is much here to interest the scholar of Prajñāpāramitā traditions. While acknowledging the merit of highlighting the ‘esoteric turn’ of Prajñāpāramitā cults under Jayavarman VII, I think most of the evidence presented above tells a different story. As far as the hymns are concerned, and perhaps much of the story of Angkorian Prajñāpāramitā itself, we are dealing with a goddess associated with ‘motherhood’ and ‘wisdom’ whom one can turn to for protection and blessings.

The simple character of these invocations may have led to their being overshadowed by discussions on scriptural sources or art historical remains. However, these short poetical hymns should be taken seriously, as they throw light on some of the key associations and functions that made Prajñāpāramitā appealing to the Khmers. At the same time, these associations and functions find some correspondence in Prajñāpāramitā discourses and shorter texts that circulated in contemporary Bengal.³⁵ To test the geographical extent of these associations, in the next section, I will review parallel sources from Java and Bali.

First, however, I would like to conclude this section with a side note. The close association between Prajñāpāramitā and Lokeśvara in the Khmer context is detected across a wide range of sources—from epigraphic hymns, to bronze triads, to the outline of certain rituals (see Vat Kdei inscription). However, the two traditions appear to disentangle once we enter the Cham domain (in modern-day Southern Vietnam). There, the cult of Lokeśvara is well attested, as an article by Green has shown with respect to both epigraphy and iconography. For example, the deity is mentioned in what is virtually the most famous Cham Buddhist inscription—the An Thai epigraph (C138) of the early tenth century. There, the esoteric dimension of Lokeśvara's cult is clearly established, as exemplified by its somewhat idiosyncratic doctrine of the three Buddha families (kulas)—Vajradhātu, Padmadhātu, and Cakradhātu. So far, I have been unable to identify any references to Prajñāpāramitā within the Cham epigraphic corpus. The few images from Vietnam that may depict this goddess are certainly intriguing, but I find it challenging to incorporate them within the wider narrative without supporting evidence. Hence, I can just highlight here some of the ongoing art historical discussions.³⁶

Nusantara and the Balinese hymns

The picture becomes richer once we move into maritime Southeast Asia. The creative interactions between the Prajñāpāramitā traditions and local cultures are more easily traced here, as is their remarkable endurance. Once can think for example of the political connotations of Prajñāpāramitā in modern and contemporary Indonesia.³⁷ These are associated with emic notions such as the wahyu, the female power that ‘turns men into emperors’, or with local legends such as those of the Pararaton—all convincingly sketched by Reichle in her monograph study.³⁸ Here, we will be mainly concerned with the circulation of texts and practices within wider transregional networks. Given the amount and complexity of this evidence, I would like to first present an impressionistic portrait of the available sources on Prajñāpāramitā, before concluding with a discussion of those that are most relevant to the current discussions on invocation literature: the Balinese stotra (‘hymns’) collections.

One may start with the Malay peninsula, pointing out its unique place in the cosmopolitan history of Prajñāpāramitā. The reason for this is that the attestation in this region of Prajñāpāramitā both as a scriptural as well as an exegetical tradition is a feature that is otherwise rare for Southeast Asia and mostly associated with Pāla Bengal, Tibet, or China. Although partly damaged, the Wat Meyong inscription (K407), now housed in the National Museum in Bangkok, illustrates this point. The beginning of the third line of the inscription reads ‘[lost text] ramitārccanaṃ’, which can be confidently restored to [prajñāpā]ramitārccanaṃ (i.e. the ‘worship of prajñāpāramitā’), as suggested by Skilling.³⁹ As the rest of the line makes clear, this is in fact a rare reference to book worship and the copying of scriptural texts.⁴⁰ But the most compelling evidence for the presence of Prajñāpāramitā scriptural traditions in this part of Southeast Asia comes with the commentarial work of a teacher called Suvarṇadvīpīya Dharmakīrti—‘Dharmakīrti of the Golden Isles’. Whether we place this figure in Kedah (Malay coast)⁴¹ or, as previously assumed,⁴² in Muara Jambi (Sumatra), the composition of an Abhisamayālaṃkāra commentary is indication that the flagship commentarial traditions of Pāla Bengal had a cosmopolitan dimension too. Or at least, their study was actively promoted on both sides of the Bay of Bengal. The journey to Southeast Asia by Atīśa only adds further weight to these intellectual exchanges in which Prajñāpāramitā exegesis played a key role. Sinclair raises an important question that has yet to be fully answered: Why would a Malay (or Sumatran) intellectual take up the Perfection of Wisdom exegetical tradition? This is not a question that one could have asked with respect to the Angkorian domain, at least given the current state of the evidence. At Angkor, Prajñāpāramitā scriptures appeared only as a minor and perhaps neglectable component of a predominantly Yogācāra/Esoteric Buddhism, at least as far as the Vat Sithor charter and Kīrtipaṇḍita's scholarly career are concerned.⁴³

Upon descending farther South into Java, the first historiographical question is whether we can detect any evidence at all for Prajñāpāramitā cults before the rise of Eastern Javanese polities, especially Majapahit. Indeed, the ‘golden age’ of Javanese Prajñāpāramitā is associated with the latter. For all the splendour and abundance of Hindu and Buddhist heritage from Central Java, there appears to be little that can be brought into direct correlation with this goddess. Reichle does mention certain relevant bronzes⁴⁴ and some depictions on temple walls look intriguing, such as the goddess above the lotus pond depicted over the south-western facade of Candi Mendut. Textual evidence too is scarce and problematic, with the Saṅ hyaṅ Kamahāyānikan and the Advayasādhana representing the two main documents of interest. As for the former, its links with the Adhyardhaśatikā (one of the few Prajñāpāramitā texts with pronounced esoteric connotations) relate not to the text as transmitted in the Indo-Nepalese tradition, but to the wider Paramādya cycle—an enlarged version of the scripture circulating in East Asia.⁴⁵ As for the latter, it contains a fascinating passage that associates Prajñāpāramitā with ‘non-duality’ as well as with the already encountered notion of ‘motherhood’.⁴⁶ The overall ethos of the text as a manual for trainees allows some comparison with the North Indian and Nepalese ādikarma texts in which Prajñāpāramitā is sometimes mentioned.⁴⁷ Such comparisons, however, have yet to lead to the identification of actual parallels.

As mentioned above, the Eastern Javanese tradition is far richer. The Kakawin Sutasoma picks up on similar themes to those of the Advayasādhana passage but is a little more elaborate on the various notions associated with the deity: ‘That is why the Lord Buddha is considered to be the son, incomparable in the world. The divine non-duality has the nature of a father according to the wise in general agreement. Prajñāpāramitā is the mother, who ever abides in yoga and concentration.’⁴⁸

Then there is the cultural world of Kṛtanagara, his imperial and cosmopolitan aspirations, as well as the particular brand of Esoteric Buddhism that came with it.⁴⁹ In many ways, the evidence here is reminiscent of that encountered in the Khmer context with respect to Jayavarman VII: the association of Prajñāpāramitā with queen mothers, the connection with their funerary rites, and the issue of Prajñāpāramitā statues functioning perhaps not as ‘portraits’ but as ‘substitutes’ in the apotheosis of queens. As explained by Reichle, the exquisite Prajñāpāramitā statue in the National Museum of Jakarta is not the only one from this period. More monumental examples are found across a number of places, including Candi Boyolangu near Tulungagung.⁵⁰ This temple may well be the one mentioned in the Nāgarakṛtāgama, in connection with the funerary rites of the Rājapatnī.⁵¹ She is said to have become Prajñāpāramitā ‘as she returned to the realm of the Great Buddha’.⁵² It is in this connection that we find the puzzling potential reference to a Prajñāpāramitā scripture with which the king is said to have been conversant—the Subhūti Tantra.⁵³ Apart from the fact that Subhūti is usually associated with Prajñāpāramitā, there is little else that can help with the precise identification of this text.

Now, as fascinating as the above evidence might be, none of it clearly depicts Prajñāpāramitā as a protective deity to be invoked through prayers, hymns, or spells—such as Mahāpratisarā or Lokeśvara. A potential and rare reference to the installation of Prajñāpāramitā as a deity in an Old Javanese copperplate charter does not throw more light on this dimension of her cults either.⁵⁴ Furthermore, Reichle's suggestion that Prajñāpāramitā statues and sanctuaries might have had a connection with ancestors’ cults has been criticised as unlikely within Kṛtanagara's cosmopolitan agenda.⁵⁵ Even the Buddhist Kuñjarakarṇadharmakathana, which mentions Mahāpratisarā and some mantra-based recitation practices,⁵⁶ does not appear to refer to Prajñāpāramitā.

There are, however, two key sets of sources that offer valuable insights into further dimensions of her cult: dhāraṇī inscriptions and the Balinese stotra traditions.

Inscriptions bearing dhāraṇīs, mantras, and gāthās are an extraordinary source of evidence for the circulation of Buddhism (as well as Brahmanism) across Monsoon Asia. The most notable examples of this genre come not from Nusantara, but from Sri Lanka—a key maritime region to which I would like to devote the following paragraph.

The image of Sri Lanka serving as a bastion of Theravāda has long since dominated the scholarly picture. In recent years, however, there has been a new wave of interest in the diverse religious landscape of mediaeval Sri Lanka, and especially the once flourishing Mahāyāna traditions. This paradigm shift was also inspired by the identification of dhāraṇī inscriptions, such as those discussed by Schopen.⁵⁷ Recently, a dissertation by Powell has taken up again certain inscribed tablets bearing spells associated with the Sarvatathāgatatattvasaṃgraha.⁵⁸ Unfortunately, no ‘spells’ specifically associated with Prajñāpāramitā have been identified so far. These could have complemented an otherwise highly fascinating corpus of evidence, including a well-known Prajñāpāramitā statue at the V&A, the metal tablets recovered from the vicinity of a stūpa at Mihintale, as well as the engraved golden foils from Jetavanarama.⁵⁹ Both epigraphic sets record fragments of the Pañcaviṃśati—one of the standard scriptural sources of the tradition—rather than any dhāraṇīs. One also finds mention of Prajñāpāramitā manuscripts being used as diplomatic props in the exchanges between Tang China and the local courts, such as that of Aggabodhi VI or the one involving General Mizhunna.⁶⁰ It is possible that further research in this area will unearth invocatory texts or spells associated with this tradition.

Coming back to Nusantara, Griffiths and others have made significant progress in the decipherment and publication of ‘spell’ inscriptions from the region. There are two examples that strike me as relevant to the history of Prajñāpāramitā. One is the Sambas foil from Kalimantan, engraved with various spells and stanzas, one of which is also found in the famous Vajracchedikā.⁶¹ The other is the Lokanātha image from Gunung Tua in Northern Sumatra. The pedestal includes a mahāyānistic formula reading anuttarāyāṃ saṃyaksaṃbodhau parināmayati (‘[one] diverts karmic retribution to the attainment of supreme perfect awakening’).⁶² This, of course, is a fairly standard expression, although it does remind one of Prajñāpāramitā phraseology in particular, as Griffiths duly points out. It is hoped that further examples will come to light in the near future.

Thus, we come to the final set of evidence—the Balinese traditions. It has long been clear that some of these materials are intimately related to the history of Prajñāpāramitā. As pointed out by Sanderson in 2003, the ritual procedures for the cleansing of the hands as outlined in the Balinese Sūryasevana manual (a Śaiva manual) show a rather intriguing feature: the ring finger and the little finger are associated with certain mantras, as well as with the deities ‘Prajñā Devī’ and ‘Parimitā Devī’, respectively.⁶³ These appear to be derived from the name Prajñāpāramitā, with the exchange of the short i for the vowel a, as attested occasionally in Javanese traditions.

The collection of stotras and stutis edited by Goudriaan and Hooykaas presents a number of relevant hymns. Some are addressed to the Brahmanical devī or to other deities associated with ‘wisdom’ and ‘insight’. Three hymns strike me with their direct mention of Prajñāpāramitā as well as their wording and content: no. 495, no. 657, and no. 658.

The first hymn starts with an homage to Prajñāpāramitā (namo bhagavatyai varāliprajñāparamitāyai); this is followed by three qualifying expressions, describing her as ‘of unlimited qualities’ (aparimitaguṇāyai), as engaged with her devotees (bhaktavatsalāyai), and finally as ‘replenished with the knowledge of all Tathāgatas’ (sarvatathāgatajñānaparipūrṇāyai). The third qualification is of interest, as it testifies to the association between Prajñāpāramitā and insight that is common throughout the ‘cosmopolis’.⁶⁴

The second hymn is more elaborate, characterising Prajñāpāramitā as the source of happiness in the world (jagatāṃ tuṣṭikāraṇam) as well as an all-pervading and kind (sattveṣu vyāpinīṃ maitrīṃ). Crucially, she is also invoked and paid homage to as a ‘saviour from all perils’ (sarvopadravatāyinīm). The mantra syllables for her invocation are then given as oṃ aḥ and huṃ. Her worship will lead to freedom from the ultimate bondage and all afflictions (asau bhittvā kleśaṃ mahābandhanamuktaye). Thus, in just three stanzas, we have an invocatory text that not only calls on Prajñāpāramitā for protection, but also enjoins the recitation of certain ‘spells’.⁶⁵

The third hymn starts by extolling Prajñāpāramitā's role within the cosmos and as the bestower of the ultimate good (svargamokṣāgrakāriṇīm). Through her, one attains freedom (kleśabandhanamuktaye) as well as all accomplishments (sattvārthasiddhim āsvame). The relevant mantra is given as oṃ-diḥ-śruḥ-tyādi-hṛnmantrāṃ. This ‘heart mantra’ is well known from Indic sources associated with Prajñāpāramitā, thus representing a rare example of an actual ‘spell’ that can be traced transregionally, similarly to the case of Mahāpratisarā.⁶⁶

These hymns present some linguistic issues that should be commented upon. For example, looking at the apparatus of the edition, one notices some confusion between nominative and accusative, masculine and feminine, as well as long and short vowels. Furthermore, a sentence in hymn 657 presents an absolutive form without a main verb. These could be seen as issues related to the manuscript transmission, but could also reflect a form of Sanskrit known as the Aiśa register, or indeed of local adaptations of the Indic language known as ‘Archipelago Sanskrit’.⁶⁷ Here, it will suffice to point out that our quests for hymns and invocatory texts has led us to engage with different registers of Sanskrit, beyond the epigraphic eulogies and kāvya sources that constituted the bulk of Pollockian language analysis.

In conclusion, there is a diverse range of evidence for Prajñāpāramitā in maritime Southeast Asia. Some of the textual passages associate Prajñāpāramitā with standard notions such as non-duality, motherhood, and insight. There are also dhāraṇī inscriptions that show a certain association with Prajñāpāramitā. A series of hymns preserved in the Balinese traditions offer real examples of invocatory Prajñāpāramitā texts, occasionally including actual ‘spells’. The terminology as well as spells are typically in line with those from South Asia and the Khmer domain, thus exemplifying a cosmopolitan dimension of Prajñāpāramitā invocation literature.

Yunnan

There is one last region that should be covered here in passing: Yunnan. This will be undertaken with reference to the Buddhist traditions of the Nanzhao and Dali kingdoms, from the early centuries of the second millennium. The reasons for including this evidence are the following: 1) There is a rough chronological alignment with the Angkorian polities as well as Majapahit. Although Kṛtanagara is slightly later than the evidence presented below, both Dali and Javanese polities had to face the Mongol expansion. 2) The position of Yunnan has long since been described as a ‘crossroad’ between Tibet, China, and Southeast Asia (although specialists on the region might be understandably wary of this term). In particular, the land connections to Southeast Asia and indeed north-eastern India make this region relevant to broader discussions about cosmopolitanism along the Southern Silk routes. A special connection with Southeast Asia may be detectable in the cult of Aculue Avalokiteśvara—famous for the peculiar iconography—with the first term probably corresponding to Sanskrit ācārya.⁶⁸ Furthermore, the presence of texts associated with Amoghavajra indicate further ‘cosmopolitan’ dimensions. But, more specifically, the reason to cover Yunnan here is that 3) Yunnanese Buddhism had a special attitude towards Sanskrit and Indic scripts, which are more common on local documents than elsewhere in East Asia. As is well known, some tombstones from Dali are engraved with the text of the Uṣṇīṣavijayā in Indic scripts, as well as that of other dhāraṇīs.⁶⁹

In terms of the history of Prajñāpāramitā, the manuscript corpus from Dali is the main source of information. At first glance, one encounters the kind of texts expected in a sinitic context. This includes copies of the Da banreboluomiduo jing 大般若波羅蜜多經 (T.220), Xuanzang's Heart Sūtra (which is also attested on some pillars, such as that of Kunming), and the Vajracchedikā. A manuscript witness of the latter text also includes a closing ‘spell’ transcribed in Chinese script, one already known from other versions recorded in the Dazang jing, the Chinese Buddhist Canon. Also Sinitic and yet more relevant to the current argument is the tradition of Prajñāpāramitā as a state protection deity, exemplified in the Renwang huguo bore boluomiduo jing 仁王護國般若波羅蜜多經—a popular scripture in East Asia, also well attested in the Dali corpus.⁷⁰

But the key point of interest is Dali's commentarial tradition on the Renwang jing. These commentaries, of course, abound with spells used to invoke the deity, as with the short texts associated with Amoghavajra (T994–996). Meghan Bryson has pointed out that one sub-commentary even appears to be unique to the Yunnanese tradition.⁷¹ This text contains one dhāraṇī that is of particular interest here and that has recently resurfaced in connection with some rare Pāla and Nepalese MSS. To optimise space, the full text is offered in this footnote.⁷² The presence of this otherwise hardly attested spell in sources from two distant regions is fascinating. The same spell is encountered once more in the Yunnanese Long Scroll—a document famous for its depictions of Buddhist deities in Tibetan garb, including esoteric deities such as Mahākāla and others. As observed by Bryson, the scroll also depicts two dhāraṇī pillars. On the left pillar, one can find the text of the mentioned dhāraṇī painted in Siddham script. This appears to be a unique example for which no parallel can be found across the otherwise large corpus of Chinese dhāraṇī pillars, in stone or on paper. Thus, a Prajñāpāramitā-associated spell attested in the Indo-Nepalese context pops up once again in Yunnan, in an Indic script, particularly in relation to apotropaic practices that are related to the same deity.

A fascinating aspect of the search for cosmopolitan ritual language is that one is not bound by the confines of what were once misleadingly called the ‘Indianised states’. A study of Buddhist ritual language quickly becomes a pan-Asian endeavour. Here, however, we have chosen to limit ourselves to a region in close cultural and geographical proximity to Southeast Asia.

Discussion: on ritual language in the Sanskrit Cosmopolis

The above discussion shows that ritual texts associated with a certain tradition can be attested across various regions. Far from being obvious, this simple fact can contribute to our understanding of knowledge exchange and cosmopolitanism in premodern Buddhist and Brahmanical contexts. Furthermore, given recent trends in the field, the study of ritual texts from a transregional perspective appears all the more crucial. As pointed out by David Gellner, Sheldon Pollock's The Language of the Gods is in many ways a field-defining contribution.⁷³ The book may have enjoyed an even wider impact were it not for certain structural issues, such as being too technical for non-specialists and too ambitious and generalising in its claims for the taste of many fellow Indologists. There has, however, been fruitful engagement with it beyond the many review articles that have since appeared.⁷⁴ In particular, Daud Ali and Thomas Hunter have continued to debate Pollock's ideas, advancing the discussion on the role of Sanskrit in the epigraphy and poetry of maritime Southeast Asia.⁷⁵

Here, I am mainly concerned with one aspect of Pollock's work that has attracted sustained criticism: the reluctance to engage with the religious domain.⁷⁶ By contrast, I argue that religious literature poses some fascinating questions, such as: What role did ritual language play in the spread of Sanskrit beyond South Asia? How can short dhāraṇīs engraved on metal foils or carved on the back of statues complement the discussion on ‘ornate poetry’ (kāvya) of royal epigraphic eulogies?

At the outset, it is useful to note that, while the latter constitute a key section of many epigraphic documents, they are but one of the many components of a typical stone stele or copperplate charter. The opening of such documents is often ritualistic in character and, while many charters might start simply with an auspicious word or symbol, others—such as many Khmer inscriptions—may offer multiple stanzas dedicated to the invocation of various deities. Furthermore, towards the end of such documents, it is common to find ‘imprecatory verses’ that warn of the consequences of not honouring the terms of grants and donations.⁷⁷ These may be verses from the epics or the Purāṇas, as with many Indian charters, or more specific sets of ‘curses’, such as those of the Indonesian charters.⁷⁸ There is thus considerable material within such documents to open a window on ritualistic language, or at least poetic language with a ritualistic function.

Be that as it may, the custom of engraving royal eulogies in Sanskrit represents an impressive cultural phenomenon, in both its geographic and chronological extents. And, while it is debatable as to how deeply Sanskritic ideals would have actually penetrated within local norms and customs, their epigraphic and linguistic expressions are extraordinarily cosmopolitan.⁷⁹ It is thus easy to understand how misleading expressions such as ‘Indianised states’ could arise, given the simple fact that an inscription from Phanom Rung in Thailand, one from My Son in Vietnam, or even one from Borneo can look so disarmingly ‘familiar’—in terms of both language and contents—to the average Sanskritist.⁸⁰

Therefore, the question is: What can be gained from looking at texts that invoke deities rather than eulogise kings? A first point is that this change in focus allows the inclusion of a more diverse set of sources, including what might be inappropriately termed ‘amulet’ inscriptions,⁸¹ spells engraved on statues or pillars, as well as manuscripts recording hymns (when these are available at all). Secondly, this approach has the potential to take us beyond the language practices of royal courts into those of various religious communities. Given the availability of appropriate documents, this could help to unveil further registers of Sanskrit, displaying fascinating variations in terms of grammar and vocabulary. Thirdly—and most importantly—a study of ritual language reaches beyond cosmopolitanism as ‘state formation processes’, highlighting the role of religious agents as part of a broad network of communities that exchanged ideas and materials as a matter of common practice.⁸² Furthermore, studying ritual language means engaging with Brahmanismas well as Buddhism. While the cosmopolitan dimension of epigraphic Sanskrit and indeed of Brahmanism was primarily limited to South and Southeast Asia, Buddhism had a pan-Asian dimension.

There is already a rich body of secondary literature that focuses on religion and transculturality across premodern Monsoon Asia. Some useful contributions by Andrea Acri—displaying a keen interest in cultural historiography—have helped shed light on these dynamics.⁸³ His edited volume has brought together a variety of approaches to the study of Buddhist agents between Sri Lanka and China, with a focus on the mobility of teachers versed in the ritual practices of Esoteric Buddhism, as well as the interplay between knowledge and material culture.⁸⁴ We are now in a better position to appreciate the cosmopolitan dimension of figures such as Amoghavajra and Kṛtanagara, as well as other agents, along a complex web of interconnected centres and multi-hub networks.⁸⁵

When choosing to focus on ritual language, we are selecting yet another angle, which, while perhaps less ambitious in scope, may add further complexity. Of particular interest here is the circulation of dhāraṇīs and associated materials across the Buddhist networks of Monsoon Asia. Among these are the highly influential Uṣṇīṣavijayā and Mahāpratisarā, on which some studies have already been published.

As for the Uṣṇīṣavijayā, one need only think of two prominent examples: extracts are found engraved on the tombstones of Yunnan (often engraved in Indic scripts, not Chinese transcriptions) and Pali/vernacular versions are still circulating today in mainland Southeast Asia. These two examples show the extent to which ritualistic literature—in this case a scripture associated with the cleansing of negative karma and the quest for a long life—can rise to become a remarkable transregional and boundary-breaking phenomenon. Indeed, even limiting the analysis to the transmission of the Uṣṇīṣavijayā in Indic languages and scripts, one is compelled to look beyond the core territories of the ‘Sanskrit Cosmopolis’ (i.e. into Southern China) as well as across the often artificial Mahāyāna/Theravāda divide.⁸⁶

Equally impressive is the story of the transmission of Mahāpratisarā texts, whose recitation is traditionally associated with protection during the perilous process of childbirth. Fragments of the relevant spells can be found on metal plates as well as statues recovered from Nusantara, while the associated deity appears to be mentioned in the Kuñjarakarṇadharmakathanā—an important Buddhist text in Old Javanese.⁸⁷ More importantly, one such fragment constitutes one of the exceedingly rare remnants of Brahmanical/Buddhist heritage from the Philippines.⁸⁸

These two traditions of Buddhist ritual language thus occupy an unparalleled place in the study of textual cosmopolitanism in Monsoon Asia, and abundant secondary sources are already available—some addressing them as cosmopolitan phenomena. One should also mention the cult of Avalokiteśvara/Lokeśvara. Well represented in Cambodian and Javanese sources, Lokeśvara was also seen as the protector of travellers and maritime crossings, and the circulation of materials associated with his cult can be studied alongside those of the above-mentioned dhāraṇī deities.⁸⁹ The transmission of the protective cult of Amoghapāśa—an eight-armed form of Lokeśvara—has been recently studied by Sinclair.⁹⁰

In this article, I focused on the transmission of certain Prajñāpāramitā texts, thus attempting to add yet another tradition of ritual sources to these discussions. It seems plausible that, similarly to the three traditions just mentioned, ideas about protection as well as processes of personification helped the wider circulation of the Prajñāpāramitā tradition, and of some of its texts, across Monsoon Asia. This point, however, should be taken up in further depth in future studies.

Final remarks

It is clear that the circulation of Prajñāpāramitā traditions along the Southern and Maritime routes took many forms and shapes, befitting the complex and ever-changing character of these traditions. While we may at first think of Prajñāpāramitā in the contexts of the Esoteric Buddhism of the period, such as that of Kṛtagara or Jayavarman VII, as well as their imperial projects and the funerals of queen mothers, there are further dimensions to this story. On the one hand, the study of the circulation of Prajñāpāramitā scriptures brings very little reward, except for glimpses into the diplomatic and scholastic dimensions that rarely emerge from the fragmentary evidence. However, expanding the search for Prajñāpāramitā texts to include hymns, stotras, and ‘spells’ proves to be highly rewarding. Many of these texts present invocations to the deity in association with notions and topoi that are remarkably ‘cosmopolitan’. Often, the spells themselves overlap with those known from the Pāla cultural milieu.

What this evidence indicates is that the circulation of Prajñāpāramitā traditions can also be studied as the story of a protection cult and its invocatory texts. In this sense, the story of Prajñāpāramitā is more precisely seen as parallel to those of Mahāpratisarā and Uṣṇīṣavijayā, and of course those of Lokeśvara as well. However, we may wish to conceptualise this—we are dealing not with forms of Esoteric Buddhism such as those associated with Kṛtanagara, but with a looser category that requires a rather different approach. As far as these Prajñāpāramitā materials are concerned, many are remarkably mainstream, if not indeed old-fashioned. This point can be appreciated even more if we consider that, in recent years, the term ‘esoteric’ may have been used too liberally when it comes to describing pre-fourteenth-century Southeast Asian Buddhisms.

More broadly, the evidence presented above reminds us to take ritual language and invocatory texts seriously as a complementary set of evidence in the study of royal epigraphic eulogies that constellate South and Southeast Asia. In this way, a more complete and complex picture of linguistic and cultural cosmopolitanism can begin to emerge.