A Buddhist mantra recovered from the Ratu Baka plateau

1. REDIRECT[[]]

by JEFFREY ROGER SUNDBERG

A Buddhist mantra recovered from the Ratu Baka plateau A preliminary study of its implications for Śailendra-era Java

The mantra and its find-spot: Buddhism and the Ratu Baka plateau

The character of the civilization which existed in ancient Central Java has been a matter of some debate ever since the British began to display curiosity toward the antiquities of the islands fallen under their sway by the happenstances of the Napoleonic Wars.1 As those readers conversant with the scholarly

literature on ancient Java are aware, the history of the political life of the island has been particularly stubborn in yielding its secrets and the extant data has been construed in many ways. The basic facts, unhappily never many, have suggested to some scholars that the wonderful stone remnants of Javanese religious artefacts were variously the work of Indian conquerors, refugee Cambodian lords, Malay war leaders, or of local Javanese rulers made good; that

the epigraphic evidence showed one, two, or even three dynasties acting upon the Central Javanese stage; that there was or was not at least one Buddhist dynasty at work in relationship to a Śaiva

JEFFREY ROGER SUNDBERG, who graduated from the University of Illinois at Urbana-Champaign and from the University of Southern California, is an electrical engineer. He is a specialist in VLSI design and high-speed signal integrity. His address is: 2601 W. Broadway Blvd, Tucson, AZ 85745-1787, USA.

1 I wish to dedicate this essay to Lokesh Chandra, David Snellgrove, and Caesar Voûte, three energetic scholars whose work has done much to rescue the relics of Javanese Buddhism from the clutch of dissolving time.

dynasty which it fought, tolerated, cooperated with, or even intermarried; that this dynasty did or did not subscribe to a tantric form of Buddhism. In particular, the vision of contention in the arena of Central Java between two great dynasties – a Śaiva dynasty stemming from their great champion Sañjaya

and a Buddhist dynasty known to history as the Śailendra – has taken a powerful hold on the scholarly historical imagination.2 This paper offers an examination of a neglected gold-plate inscription from Central Java which bears impact on all of these subjects. The subject at hand is a Buddhist mantra

inscribed on gold foil and unearthed from the Ratu Baka prominence, a kilometer’s distance to the south of the famous Prambanan temple complex in Central Java sometime during or just after the Second World War. Its only description in the archaeological literature occurs in the report of its finding in the

journal of the Archaeological Service of the former Netherlands East Indies (Oudheidkundig verslag 1950) and the seemingly single analytical commentary devoted to it was that offered by the late Indonesian archaeologist Kusen (1994:85). The mantra’s present location is unknown. I sought this mantra while

in Java and it sadly seems to have disappeared from all likely institutional repositories without a trace, a casualty of wartime, theft, or neglect. Luckily we are provided with a hand-drawn facsimile by the archaeologist Soehamir, who during the time of the Japanese occupation and subsequent war of

independence had assumed operating control of the Archaeological Service and was responsible for writing its reports during this period. The mantra itself is worth some effort to describe (see Figure 1), as its vajra (diamond or thunderbolt) shape tends to confirm its message. The gold foil has been snipped

into the form of a double trapezoid, the physical dimensions of which are unfortunately lacking in Soehamir’s report. The foil has been inscribed on both the leaves of each side with a short inscription written in a script commonly called Kawi or Paleojavanese. These four inscriptions repeat the basic mantra

oµ †akΔ hūµ ja˙ svāhā3 but with a highly significant 2 See Jordaan (1999) for an account and evaluation of the postwar scholarship on the subject of the ›ailendras in Central Javanese history. 3 I wish to point out the rather frequent occurrence of another and very similar Buddhist mantra, running

paki huµ ja˙, which has been found in a number of Yogyakarta area inscriptions, including the inscriptions of Candi Abang (872 AD), Vihāra (874, a second instance of which has recently been recovered south of Yogyakarta and discussed in an unpublished conference paper by Rita Margaretha Setianingsih),

Borobudur (undated gold foil mantra, treated by Boechari 1976), Alih Tingal (undated but assigned to circa 883 by Damais 1970:49), and Paki Hum Jah (undated). I do not believe that Soehamir mistransliterated the inscription, mistaking the ‘pa’ for a somewhat similar ‘†a’. The facsimile of Soehamir

leads one to believe that his transliteration of ‘†a’ is the correct one, the character opening to the right rather than upwards. His choice of transliteration is supported, as we shall see, by the double-vajra shape of the foil. Furthermore, to reverse the coin, personal inspection of four of the

Paki Hum Jah inscriptions shows that their ‘pa’ is very clearly chiseled; their mantra is not that of the Ratu Baka gold plate. I hope to soon devote a paper, parallel to this one, to the implications of the ‘paki huµ ja˙’ mantra.

A Buddhist mantra recovered from the Ratu Baka plateau 165 variant on one side: within the exaggerated bubble which comprises the vowel ‘Δ’, there are a further two cramped lines comprising the characters

‘panarabwan’ and ‘khanipas’. Figure 1.

Figure 1. Ratu Baka gold plate

The mantra itself was located rather imprecisely by Soehamir, who merely gave its location as found ‘during the excavation of the remains of the big retaining wall in front of the first gate a small gold plate, in the form of two interrelated diamonds [which were] inscribed on either side’. A want of greater precision thwarts a better understanding of the precise implications of the burial location of the mantra, for in front of the first gate stands not a retaining wall but a ramp leading down from it. We must presume in the absence of better information that the mantra was found closer rather than farther from the first gate, possibly in the steps or ramp leading up to it.

The date of the mantra: Java in the late eighth century The gold foil is unfortunately undated; words uttered or written in the context of the Eternities have no business carrying marks of their temporality. However, there is sufficient evidence, both internal and external, to suggest that the mantra may be

safely dated to the period 784-803 AD, for both regal references and a paleographic peculiarity point to this period as the most appropriate for the dating of the mantra. In the first instance, we take the word ‘panarabwan’, itself seemingly devoid of meaning in Sanskrit, Old Javanese, or Old Malay, as an orthographic variant of the name of the king given as the Raka of Panarban in the inscription of Wanua Tengah III (908 AD).4 This latter provides a 784 AD

coronation date for the Raka of Panarban and of 803 AD for his successor, the Raka of Warak, providing a bracket of 18 years for the reign of Panarban. 4 See plate I, face B, line 4. of this unfortunately unpublished inscription. Both Kusen and Boechari made transliterations of this inscription. Readers interested in obtaining the transcription may contact me to obtain a copy.

In the second instance, there are paleographical considerations which tend to support the dating suggested by the reference to the royal name. The form of the ‘ka’ as represented in Soehamir’s facsimile is an archaism and could only come from a relatively early period of Javanese history; so far as I am

aware, no datable occurrence of this can be found in Java later than the 792 AD inscription of Mañjuśrīg®ha, dating to the middle of Panarabwan’s reign and found at the periphery of the temple yard of Candi Sewu, a half hour’s walk to the north of the Ratu Baka escarpment. This ‘ka’, which is ordinarily formed by a single arch divided vertically by a line, as such , has been formed in an archaic manner, by extending two arcs of different diameter from a common terminus on the left-hand side, as such . This latter form is found sporadically in the inscription of Mañjuśrīg®ha along with the more standard form.5

The source of the mantra: Falling at the feet of a yakßa, one is raised in nirvāøa After a substantial amount of directed browsing through various categories of Mahāyāna and Buddhist tantric literature, I believe that it is safe to say that the Ratu Baka mantra inscribed by Panarabwan represents a

variant of the famous h®daya (‘personal or quintessence spell of a divinity’, see note 18) which is to be found in its locus classicus in the second section of the Sarva Tathāgata Tattva Saügraha (The Symposium of the Reality of All Tathāgatas, hereafter abbreviated as STTS), the root tantra or mūlatantra of the category of Buddhist literature labeled in Tibet as the yoga tantra. As the Buddhist scholar David Snellgrove has emphasized, the STTS

inaugurated a new, distinct form of properly tantric Buddhism notable for its appreciation of the salvific power of the vajra and the importance of ritual consecration for the attainment of enlightenment. The second section of the STTS recounts the

5 See the beginning of line 13, åku, for instance. This form of ‘ka’ is also noted in the fifth line of Sukarto Atmodjo’s facsimile (1994) of the unfortunately undated Buddhist inscription of Wutit, found near the north coast of Central Java. Interestingly, an approximate recurrence of this form can

be seen in Holle’s facsimile (1877:4) of the seemingly misdated tinulad inscription of Ku†i. I have not seen this inscription myself and therefore cannot comment on the accuracy of his facsimile. This peculiarity of ‘ka’ is found also in two Sanskrit and Old Khmer Buddhist inscriptions, dated about two

centuries after our mantra and found near Wat Wieng in southern Thailand, very close to the location of the 775 AD stone of Ligor. Because these latter were found closer to what is believed to be the Śrī Vijayan and Malay sphere of cultural influence, questions necessarily exist about the possible origins

and persistence of this form of the ‘ka’ in Sumatra-dominated lands. The fact of its existence on the Old Malay inscription of Mañjuśrīg®ha may be considered as another piece of evidence pointing to this conclusion. For this reason, the important mantra plates unearthed from the Sumatran Buddhist temple of Candi Gumpung deserve the thorough paleographic description and facsimile which has so far been lacking. For this latter, see Boechari (1985).

tale of Trailokyavijaya, the famous story of the taming of Maheśvara and his installation into the maø∂ala by the bodhisattva Vajrapāøi, a legend which is the most important tale in Vajrayāna Buddhism.6 In this context we wish to note the presence of the mantra ‘Hūµ †akkijja˙’, which appears in the STTS as both the great All Tathāgatas Elephant-Goad Spell of Summoning (the manuscript of Yamada (1981:160) offers both sarva-tathāgata-samayāükuśan and sarva-

tathāgata-samaya-vajrāükuśan to characterize this) in the chapter on the Trilokavijaya-mahāma ∂ala as well as the personal svah®daya of the wrathful bodhisattva Vajrapāøi in the Trilokacakra-mahāma ∂ala.7 Both the Trilokavijayamahāma ∂ala and the Trilokacakra-mahāma ∂ala concern Vajrapāøi’s efforts to compose a maø∂ala with forcefully subjugated Hindu deities (Snellgrove 1981:43-51). In order to facilitate understanding of the context in which the mantra

of summoning occurs, I will present an abbreviated version of the story of Trailokyavijaya, cobbled together from the work of the various translations of the story,8 coalescing some dialogue and omitting some details, and accenting elements of the story which I feel most pertinent to arguments which will appear further on in the essay. Needless to say, any defects in the presentation which follows are my own fault and not that of the fine translations used as the basis of the summary; the English-speaking reader is urged especially to consult Snellgrove (1987) and Davidson (1995) for a more complete presentation of the story.

A summary of the tale of the subjugation of Śiva The second section of the STTS commences with an invocation of the Lord of the Maø∂ala, who is accorded 108 great names beginning with Mahāvairocana, Vajradhara, Vajrapāøi, etcetera. After the invocation of this Lord of the Maø∂ala, the story begins for real on the peak of Mount Sumeru,

6 Davidson (1991:199) calls it ‘perhaps the most influential myth of esoteric Buddhism’. 7 Yamada (1981:253-80) describes the entire maø∂ala, with the specific instance of the mantra on page 259. 8 For those readers conversant with Italian, Tucci (1932:135-45) was the first to offer a translation from a fragment of a Sanskrit manuscript he recovered from Nepal. Snellgrove (1981:39-43), working with a ninth-century Brahmi manuscript

discovered by him and Brough in Nepal, provides a translation and a reprise of the translation accompanied by a commentary on selected parts of the story (Snellgrove 1987:134-41). Iyanaga (1985) composed a careful Frenchlanguage summary of the colourless eleventh-century Sung redaction. Shashibala (1986:66-

8) provides an English translation. Davidson (1991:198-202) offers both a substantial summary as well as a more extended translation (1995:550-5) which captures the narrative flavour of the Sanskrit. Linrothe (1999:183-5) again offers a summary as an introduction to an extended analysis of the meaning of the story and its realization in Indian art (1999:178-212). The text itself has been critically edited by Yamada (1981).

where all the Tathāgatas requested Vajrapāøi, Lord of All the Tathāgatas and Universal Ruler, to summon to the Maø∂ala the divinities of his retinue, those who exert control on behalf of the Tathāgatas. Vajrapāøi declined, claiming that there still were untamed beings such Maheśvara (Śiva) and others who were

not yet converted to the Dharma by the Tathāgatas. In response, the Lord Vairocana entered into a samādhi known as Wrathful Pledge-Vajra to Perform the All Tathāgatas’ Exalted Procedure, and enunciated the syllable, ‘H√µ’. At once there emerged from the vajra at the heart of the Lord Vajradhara a manifestation of a variety of fearful, fanged, flaming, frothing Vajrapāøiforms, reciting this verse: Oho! we are the means of conversion, possessed of all great means.

Stainless ourselves, we assume a wrathful appearance so that beings may be converted by these means. Then Vairocana uttered the spell hūµ †akkijja˙, known as the disciplinary Elephant-goad Spell of All the Tathāgatas. By this spell, the wicked lords of all the material universes, preeminently Maheśvara (Śiva) and his group, were all dragged to the Adamantine Jeweled Palace at the peak of Mount Sumeru. Vajrapāøi raised his vajra away from his heart and waved it

while he surveyed the whole circle of the summoned Triple World. He spoke, ‘My friends, come to the teaching of all the Tathāgatas. Take refuge with the Buddha, the Dharma, the Saµgha, so that you may gain the knowledge of the Buddha.’ But Mahādeva (Śiva), the proud Lord of the whole Triple World grew angry and said, ‘Listen you little yakßa, I am Iśvara, Lord of All the Worlds, Creator, Destroyer, Lord of all Life, God of all Gods, Supreme God. Why should I

carry out the order of some low-life yakßa like you?’ Vairocana tried to intervene, to persuade Maheśvara and the whole host of gods of the Triple World, ‘Friends, take the vow of the triple protection before Vajrapāøi, this so-called yakßa, the Great Bodhisattva, wrathful, terrifying and fearful, should destroy the whole universe with his blazing vajra’. Maheśvara ignored Vairocana and decided to intimidate Vajrapāøi by assuming the fearsome, flame-wrapped

form of Mahābhairava, commanding, ‘It is I who am the Lord of the Triple World. You will obey my orders and take my vow!’ But Vajrapāøi, waving his vajra and laughing, said, ‘Come here, you who snack on corpses and flesh, you who use the ashy residue of funeral pyres for your food, as your bed, as your clothing: obey my command!’ After a prolonged and fruitless exchange of more threats and insults, Vajrapāøi returned to consult Vairocana, saying, ‘Well,

Sir, that prideful god is still not paying homage to the Dharma. Now what do I do?’ Then Vairocana recited the great vajra-mantra which has its origin in the heart of all the Tathagatas: Oµ nisumbha vajra hūµ pha†, to which Vajrapāøi added his own vajra syllable: Hūµ! At this, all the devas of the Triple World fell flat on their faces, moaning miserably, and all these

gods except Maheśvara went to Vajrapāøi for protection. Maheśvara himself died and his body lay motionless on the ground. Vajrapāøi preached to the other gods about the necessities of the Buddhist doctrine, and they all were converted to the Dharma. Then Vairocana requested that Vajrapāøi revive Maheśvara, noting that Maheśvara was useless as a corpse but might become a good and worthwhile being if brought back to life. After his revival from death, Maheśvara

was too weak to stand, but asked Vairocana, ‘Sir, what is it that you are trying to teach me? Aren’t you Buddhas supposed to save wicked beings like myself?’ Vairocana responded, ‘You still haven’t done what Vajrapāøi told you to do. It is Vajrapāøi who is the Lord of All Tathāgatas. Your education is his responsibility, not mine.’ Vajrapāøi then interjected, ‘Why don’t you simply do what I tell you?’ When Maheśvara heard Vajrapāøi, he again grew

recalcitrant, displaying his form as the fearful Mahāraudra and saying, ‘I can suffer death, but I will not obey the command of this yakßa!’ Having exhausted all other means, Vajrapāøi spoke the appropriate mantras, and to the roars of laughter of all his retinue, Maheśvara and his consort Umā were dragged nude and feet up before Vajrapāøi, who stepped on Maheśvara with his left foot and on Umā’s breasts with his right. Vajrapāøi then uttered another

mantra causing Maheśvara to batter his own thousand heads with his thousand arms, while all his underlings outside the palace laughed and said, ‘Look at our Lord being disciplined by the Great Bodhisattva!’ Then Vairocana took pity on Maheśvara and spoke the mantra which encapsulated the Compassion of All Buddhas. Immediately the contact with the sole of Vajrapāøi’s foot brought liberation instead of suffering. Maheśvara became the recipient of

consecrations, powers of meditation, liberation, the faculties of knowledge of all the Tathāgatas and magical spells of the highest sophistication, equal to the reality of Buddhahood. His body having thus fallen at the feet of Vajrapāøi, Maheśvara was reborn as the Tathāgata called Bhasmeśvaranirghoßa (Soundless Lord of Ashes) in the universe known as Bhasmacchatra (Umbrella of Ashes), which exists far down below, beyond countless world systems. Then

Mahādeva said: The marvelous wisdom of all the Buddhas! Trampled under the feet of a yakßa, one is raised up in nirvāøa! Then the Great Bodhisattva Vajrapāøi, making special mudrās and uttering special mantras, invited all these gods of the Triple World – Mahesvara, Umā, and the others – to enter into their proper place in the great Vajra-Pledge Maø∂ala of all the Buddhas.9 Vajrapāøi revealed the great maø∂ala to them, and having consecrated them with the Gem Consecration and having placed

9 I borrowed the terminology of Snellgrove (1987:128). Davidson (1991:202) translates this as ‘the great circle of the Adamantine Assembly of all Tathågatas’. in their hands their proper vajra-implements, he consecrated them with the Name Consecration and established them in their proper locations in the maø∂ala to further the welfare of living beings.

Some considerations on the story of the Trailokyavijaya and the Ratu Baka mantra We first offer a note to reconcile the slight differences between the mantra of the Trailokyavijaya and that of the Ratu Baka gold plate. The Ratu Baka mantra has, as the reader will recall, Vajrāpaøi’s hūµ interposed between

the †aki and the ja˙, and is further seemingly couched in the form of a benediction, with the mantra bracketed between an oµ and a svahā. The species of the mantra found on the Ratu Baka therefore sounds less aggressive than the utterances used to bring Maheśvara under Buddhist control; it is as though the Javanese inscription is a hearkening to the glory of Vajrapāøi’s triumph rather than a formal attempt to duplicate the feat of summoning using the proven

efficacy of his mantra.10 This observation will be further extrapolated in an examination of the locational context of the mantra below. Note that it is entirely possible that a source other than the BroughSnellgrove Nepali Sanskrit manuscript of the STTS may be found for the exact form of the Ratu Baka mantra. Indeed, even within the known group of rescensions of the STTS (two extant Sanskrit manuscripts including one dating to the ninth century, a

partial Chinese translation deriving from no less than Amoghavajra in 753 as well as the standard S’ung translation made in 1015, and several Tibetan translations dating from as early as the early eleventh century)11 there is a surprising and considerable variety in the reportage of the mantra, probably relating to the arbitrariness of its syllables. For instance, Tucci’s Sanskrit fragment read hūµ †atki jja˙, while among those variants thrown up by Yamada’s patient collation of the other rescensions are the ††akki within the Sanskrit manuscript itself and †akke in both the Chinese, and all other manuscripts oppose the Brough-Snellgrove instance of only a short ‘u’ in the hūµ.12 Davidson notes that the vast popularity of the Trailokyavijaya myth ensured that it was told and retold in several prominent works of the yoga tantra and anuttara-yoga tantra classes,13 and as well of course in a

10 As I hope to show on another occasion, the Javanese did indeed employ their mantras ‘raw’, in a direct attempt to invoke the Buddhist and ›aivite deities, rather than as mere commemorative slogans.

11 Yamada (1981:5-6) describes his sources, to which we must add the Nepali Sanskrit fragment read by Tucci (1932:appendix A).

12 Tucci 1932:appendix A; Yamada 1981:160 note 5, 200 note 9, 201 note 5, 259. 1

3 Davidson (1991:203) specifies the Vajraśekhara-mahåyoga-tantra, the Candraguhyatikalamahātantrarāja, and the Guhyagarbha-tattvini¢caya as texts which carry or reflect the tale.

now irrecoverable oral form. We therefore find a reflux of this mantra in the thirteenth-century rescensions of the yoga-tantric Sarvadurgatipariśodhana tantra, where it again is used in a context of summoning various divinities under the general auspices of Vajrapāøi.14 We incidentally note the recurrence of the mantra’s signature syllables in the existence of ‡akki-rāja, a yāma who is a fierce attendant of Mahākāla in the anuttara-yoga tantras (Willson and

Brauen 2000:337). One of the most interesting of the mythological figures associated with the mantra under examination is the Japanese Aizen-myōō, the ‘King of Esoteric Knowledge [called] Tinted by Love (or Lust)’, who has been subjected to an extensive review by Goepper.15 This deity was mentioned in the ‘Sūtra of all Yogas and Yogīs of the Pavillion with the Vajra-top’ which is apocryphally attributed to either Vajrabodhi or Amoghavajra and which was first

transmitted to Japan in 805 by Kūkai; the text later became considered as one of the Five Basic Esoteric Texts within the Japanese tradition, although it was seemingly a T’ang concoction with no presently known Indian prototype (Goepper 1993:10). Commentary by Japanese tantric clerics explicitly identifies Aizen with a whole host of the primary yoga-tantric deities: Vajrasattva, Vairocana, Vajrarāja, ‡akkirāja, and Trailokyavijaya. In its treatment of Aizen-

myōō in the ‘Sūtra of all Yogas and Yogīs of the Pavillion with the Vajra-top’, the mantra oµ †akki hūµ ja˙ with variants hūµ †akki huµ ja˙ and huµ †akki huµ ja˙ occurs in various associations with Aizen (Goepper 1993:127), the primary association occurring within Chapter 5 where Aizen is linked with a particular conquest: All irritations of bad character [Endangering] the multitude of Bhikßus of pure conduct, The poisonous and wicked dragons, difficult

to control, Nārāyaøa and Ôśvara (i.e. Vißøu and Śiva), The Four Heavenly Kings, protecting the world, 14 See Skorupski 1983:25. The mantras which are recorded are superficially more attractive to us as one of them places the huµ into the same sequence as the Ratu Baka mantras: Sarvadurgatipari¢odhana records instances of vajravåcå †akki huµ ja˙ ja˙ ja˙ and as well †akki ja˙ ho˙. Furthermore, the text at this point specifies that the two mantras be uttered with the tantric bell and vajra in hand, held to the heart. Unfortunately, these mantras are not employed in the earlier eighth-century rescension of Sarvadurgatipari¢odhana, the version which was translated into Tibetan by ›antigarbha and JayarakßΔta and which was so heavily commented upon by some of the best Indian Buddhist scholars of the time, including Buddhaguhya. Both editions of the

Sarvadurgatipari¢odhana have been published in Skorupski 1983, where the thirteenthcentury version is buttressed with a Nepali Sanskrit manuscript and graced by a translation into English, and the eighth-century version is critically edited by Skorupski’s referencing of the commentaries of Vi¢vavarman. 15 Goepper 1993. I am indebted to Lokesh Chandra for pointing out the existence and relevance of this work to me.

They all will immediately be subdued and will lose their lives. Again, I will preach Rāgarāja’s Heart vidyā in One Syllable: Oµ ‡akki hūµjja˙ (Goepper 1991:15)

It is sufficiently clear that this Sino-Japanese deity has been synthesized from tantric source elements, some of which are close to those which inspired the Ratu Baka mantra, and this synthesis occurred before the time that Kūkai took a copy of the sūtra back to Japan with him in 805, leading us to the interesting observation that these Aizen-myōō innovations must have been nearly contemporary with the writing of the Javanese mantra. Wherever else

instances of the Ratu Baka mantra may be found, it is my opinion that we must argue for a primacy of the story of Trailokyavijaya as found in the Vajra Family material of the STTS as the context of the Ratu Baka mantra – its presence within the STTS placed it in the root text of the yoga tantra tradition but also seems to have been the paradigmatic usage to which all other references point. Iyanaga (1985) spent a considerable portion of his essay tracing

the history of the Chinese translation of the Trailokyavijaya text in the STTS and makes it clear that the three famous translators – Śubhākarasiµha, Vajrabodhi, and Amoghavajra – devoted their attentions to at least summaries of the story, proving that it was considered part of the STTS by 705 AD at the latest. Furthermore Śubhākarasiµha personally brought to China paintings of the maø∂alas of the STTS.16 However much it is desired that scholars with the

relevant linguistic skills will be persuaded to search the extant tantric texts in the hope of uncovering more recurrences of the Ratu Baka mantra,17 it is clear beyond much doubt that the basic tale of Trailokyavijaya was bound up most famously as a component of the STTS, the root (mūla) tantra of the yoga tantra class, and was acknowledged to be an integral component of that text at the time when the STTS was sent from

16 Iyanaga (1985:724-5) notes that Śubhåkarasiµha’s summary of the Trailokyavijaya story in his commentary on the Mahåvairocana sūtra involves Acala and not Vajrapåøi but concludes that the Vajrapåøi story was the original version and first found in the STTS. 17 In light of the research conclusions of Nihom (1995:113-5), which suggests that texts of the Yoga Tantra class were known in ancient Java but that the Vajradhåtu maø∂ala was not, one of the

most fruitful texts to search would be the Vajra¢ekhara-mahåyoga-tantra, mentioned in a previous footnote, which carries the tale of the submission of Mahe¢vara but omits the material about the Vajradhåtu (Davidson 1991:203). Nihom (1995:109-10) has found in the Tibetan recension of the Vajra¢ekhara-mahåguhya-tantra, itself the premier commentary on the STTS, the origin of a certain Balinese Buddhist mantra. Similarly, Matsunaga (1998:5) believed that

the iconography of the Nganjuk bronzes best accorded to a vidhi on the STTS by Ånandagarbha entitled the Sarvavajrodaya, and although coming from at least a century later than our period, it may prove useful in tracing sources. On the other hand, De Jong (1974:479) notes that the Javanese Sang Hyang Kamahåyånan Mantranaya seems to more closely accord with Tibetan rather than with Chinese sources; this might provide another clue about the proper direction of inquiry.

India to China, at least half a century before the date we assign to the Javanese gold foil. We turn now to some elucidation of the meaning and consequences of the hūµ†akkijja˙ mantra, the h®daya of aükuśa, used in the text as both the elephant-goad means of summoning the gods to Sumeru and also as the personal h®daya or quintessence of Vajrapāøi.18 Here we summarize how other scholars have chosen to appreciate this mantra; after all, the elephantgoad

will be an implement sufficiently foreign to most of the city-dwelling readership, much less an elephant-goad in a Buddhist context. The basic idea of this mantra is that it is the mantra of a ‘hook’, ‘goad’ or a ‘prod’ for summoning, for compulsion, for convocation, the utterance of which has compelled

Mahādeva to Sumeru. Iyanaga (1985:669, note 47) notes that the Japanese commentary provides some control over the meaning of the Sanskrit mantra, which could also be considered as ākarßana (magic to draw to oneself) or by vaśΔ karaøa (magic to draw someone to your will). Davidson (1991:200) offers

‘dragged’ as a term suitable to connote what happened to the gods after Vairocana intoned the mantra. This power of compulsion or of summoning is the sine qua non for the Mantranayist, for as Snellgrove (1987:235) observes, ‘All tantras claim the power to coerce divinities, for it is by coercing them into an

image or symbol that one is enabled to worship them and make them suitable offerings, and it is by coercing them into oneself that one is enabled to act with their assumed assurance and thus achieve the objective in view’. There is a further passage, again drawn from the indispensible Snellgrove (1987:222-

), which serves very well to differentiate the use of the aükuśa from other tantric implements: the four door guardians represent the four stages of introducing the divinities into the maø∂ala, which are effected by the mantra Ja˙ Hūµ Vaµ Ho˙! Vajrāükuśa (Vajra-Hook) summons them; Vajrapāśa (Vajra-

Noose) draws them in; Vajraspho†a (Vajra-Fetter) binds them and Vajrāveśa (Vajra-Penetration) alias Vajraghaø†a (Vajra-Bell) completes the pervasion of the maø∂ala by wisdom.

18 Snellgrove (1987:141) discusses ‘magical formulas’, noting that the words mantra, dhåraøΔ, h®daya, and vidyå all connote the same thing. H®daya is routinely translated into Tibetan by a term denoting ‘quintessence’, and may be considered ‘comprehending the essence of a particular divinity’, and in this sense bija or seed is a the more correct term. However, the h®daya is the ‘personal “spell” of a divinity even when this consists of several syllables’. Hūµ †akkijja˙ is the personal ‘quintessence’ mantra of Vajrapåøi.

The mantra, the gate, and the king: Why the Raka of Panaraban is likely a great Śailendra lord Accepting that the preponderance of the evidence shows that the mantra found on the Ratu Baka is to be associated with the Sarva Tathāgata Tattva Saügraha or other yoga-tantric literature as exemplified in the

traces of the story of the submission of Maheśvara, we must turn our attention to the remaining words engraved on the foil, those transliterated by Soehamir as ‘Panarabwan khanipas’. In the first instance, we reprise our argument above that Panarabwan very likely designates a variant orthography of the

title of Raka of Panaraban, the territorial title of the king who ruled from 784-803 AD. One careful critic of an early draft of this paper cautioned about the hypothetical character of the identification of Panaraban with the Panarabwan of the Ratu Baka mantra. I have accordingly maintained in this paper a

strict separation of references of the Raka of Panaraban, the eighth-century king documented in the early tenth-century Wanua Tengah III inscription, with ‘Panarabwan’ of the eighthcentury mantra, as it cannot definitively be proven that the latter is a raka title designating the Javanese king. This said, the

reader should be aware of the strong phonic and orthographic reasons why the identity of Panaraban and Panarabwan should be accepted: the ancient Javanese scarcely distinguished between ‘w’ and ‘b’, being perfectly content to spell ‘Buddha’ as ‘Wuddha’ even in royal Sanskrit inscriptions of high literary

competence, so a later scribal pruning of the name Panarabwan to Panaraban would be entirely in order. Alternatively, the Ratu Baka mantra might itself document the name ‘Panarabban’; the differences between the form of the character ‘wa’ and the character ‘ba’ is that the latter is a slightly dimpled

version of the former and the distinction between the two characters might be lost in the necessities of writing the name in the constricted circle allotted for the name, which would make the later scribal truncation from Panarabban to Panaraban all the more natural. Furthermore, the mutation of

Panarabwan to Panaraban between the time of the writing of the mantra and the time of the writing of the name in the Wanua Tengah III inscription is paralleled by even more substantial changes in the spelling of Panangkaran’s name from the Paøaµkaraøa in the inscription of Kālasan to the Panaükaran of

the inscription of Wanua Tengah III. Considered from another perspective, not only do phonic and orthographical considerations suggest that the identification of Panaraban and Panarabban/Panarabwan be accepted, the architectural milieu also suggests this as plausible; Panaraban’s predecessor the Raka of Panangkaran (reigned 746-782 AD) raised a Buddhist vihāra and the Tārā temple of Kālasan,

easily visible from the Ratu Baka, in 778 AD19 and a large temple to MañjuśrΔ was built around four years later, again within sight of the Ratu Baka heights. Furthermore, the Śailendra vihāra for the Sinhalese Abhayagirivāsins was seemingly constructed on the very same Ratu Baka plateau, at least by the

middle of Panaraban’s reign in 792.20 It is clear that the immediate surroundings of the Ratu Baka hill were a hive of Buddhist building activity at the time that Panaraban came to his throne in 784 AD, and it should be no surprise that our Panarabwan should continue this architectural activity with the erection of the immense gates which are the tallest extant features on the plateau. Finally, an instance of the name Panaraban/Panarabban/ Panarabwan is

unknown from any other source and the circle of people who would be privileged with the right of depositing Buddhist memorabilia on the Ratu Baka hill would also be very small, likely consisting of only the royal family and the foremost of the Buddhist ecclesiastics. That this rare name Panarabwan, a trivial and very plausible variant of the known raka title of a king whose rule was contemporaneous with other Buddhist construction in the vicinity,

should turn up on a gold mantra on a plateau known to support a royal vihāra, is to my mind firm evidence of the identity of our mantra’s Panarabwan with the king Raka of Panaraban. In fact, it is difficult to believe that Panarabwan could be anyone other than the Śailendra king under whose auspices the 792 Buddhist inscription of Abhayagirivihāra was composed. (De Casparis (1961:245, note 12) gives the name of this regent as Samaratuüga). Not only does this

conclusion accord with the coincidence and cotemporality of Panarabwan and Samaratuüga on the heights of the Ratu Baka in the period around 790, but seems to derive also from the nature of the Great Gate and its endowment. Accepting the

19 The interpretation of this inscription has been the source of a great deal of historical contention about whether it designates one king or two. Following the lead of Vogel (1919), Van Naerssen (1947) makes the case that the inscription concerns two kings, a ›ailendra king who remains remote and anonymous, influencing the local king Panangkaran through the intermediary of his ›ailendra court preceptors (see also Bosch 1952:113, note 3). Lokesh Chandra (1995:217) notes that the sixth stanza undoubtedly indicates that Panangkaran himself built the Tårå temple, though at the behest of the guru of the ›ailendras: to him, the overall implication of the inscription is that Panangkaran is the ›ailendra king. 20 The inscription of Abhayagirivihåra has sadly not yet been fully published. Portions of the inscription have been studied by Bosch (1928:62-4) and much more extensively by De Casparis

(1950:11-24). De Casparis (1961) treats the newer portions of the inscription which provide the date and the specific affiliation of the monastery. The transcription presented in Sarkar (1971-72, I:inscription VI-A) has unfortunately confused De Casparis’ admittedly cryptic notes on the placement of the

newer fragments, resulting in an erroneous transliteration which accompanies several flaws in translation as well. De Casparis (1981:73-4) presents a transcription and translation of the first three complete stanzas, on which Lokesh Chandra (1995:11-21) offers a comprehensive review, commentary, and

several valid amendments. I intend to publish a complete study, including a full transliteration, of all the extant fragments on another occasion.

manifest Śailendra suzerainty over the plains of Prambanan during this time, how could the king Samaratuüga allow the king Panarabwan to construct not only the tallest feature of the plateau but also a feature under which the great Śailendra mahārāja would have to pass to access the escarpment and his vihāra? On grounds of royal protocol alone, we must suspect that the Raka of Panaraban was the member of the Śailendra dynasty bearing the consecration name Samaratuüga, and that therefore members of the Śailendra dynasty were folded into the lineage of kings commemorated in the inscriptions of Mantyāsi˙ and

chronicled in Wanua Tengah III. I therefore believe that Panangkaran (reigned 746-784) was the Buddhist king reputed as the ‘Killer of Haughty Enemies’,21 that Panarabwan (784-803) was Samaratuüga, and that Warak (803-827) was perhaps – the evidence is circumstantial and would require many pages to argue –

the child known as Bālaputradeva whose vihāra at Nālandā was benefacted by a king of the Pāla dynasty.22 Returning to the important topic of the meaning of the words retained within the arcs of the vowel ‘Δ’, no satisfying resolution presents itself. Kusen (1995:85), taking as his reading hanipas, seemingly a

derivation of the Javanese word tipas, meaning ‘to store’, argued that the words denoted that Panarabwan ‘stored’ or ‘embedded’ the mantra-plate, or rather, because Kusen apparently envisioned a passive form, that the mantra was stored by Panarabwan. It is seemingly possible – the reader may decide – that Kusen’s reading of a ‘ha’, , is an acceptable interpretation of Soehamir’s ‘kha’, which generally looks like this: . There are however several objections to Kusen’s

21 While other scholars (notably and most prominently De Casparis in his 1950 and 1956 publications) have suggested that the kings listed in the inscription of Mantyåsi˙ of 907 AD were Javanese Hindus, Panangkaran’s Buddhism seems demonstrated by his benefactory land donations on behalf of the vihåra at Pikatan as documented in the inscription of Wanua Tengah III of 908. Interestingly, this Pikatan vihåra was founded by the younger brother of the Rahya¥ta i M∂a¥, who must almost certainly be Sañjaya. Another Mantyåsi˙ king whose adherence to Buddhism seems assured is the Raka of Garung (829-847):

his presence as the anonymous ¢rΔ mahåråja of the Plaosan shrines seems confirmed by the presence of votive shrines contributed by both the famous ¢rΔ kahulunnan of the 842 inscriptions and as well by the Raka of Sirikan pu Sūryya, a high minister who accompanied Garung upon his 829 rededication of the

tax charter for the fields at Wanua Tengah for the vihåra at Pikatan. I hope to be able to document further evidence of Panarabwan’s Buddhism at a later time. 22 In the historical reconstructions of Kusen (1995) and Wisseman-Christie (2001), both of which are ridden with invalidating flaws which would

require a substantial essay to untangle, the Abhayagirivihåra inscription’s evidence of Samaratuüga’s reign is inexplicably ignored. Kusen (1995:85-6), unaware of De Casparis’s findings (1961) on the Abhayagirivihåra inscription and basing his arguments upon PramodavarddhanΔ’s need for outside land-

endowments for her temples, saw Samaratuüga as a Sumatran foreigner with no land-rights in Java, a child of Panangkaran allegedly sired upon a princess of a militarily vanquished ›rΔ Vijayan state. Wisseman-Christie (2001:36), who strangely cites De Casparis’s 1961 preliminary study of the inscription of Abhayagirivihåra and its findings that Samaratuüga was ruling as early as 792, still equates Samaratuüga with Warak (reigned 803-827).

translation that render it exceedingly unlikely, if not impossible. First, the passive form of tipas/kipas should be t/kinipas rather than t/kanipas.23 Second, the subject Panarabwan should come after rather than before the verb in the passive construction. Finally the context of benefaction or benediction

calls for a verb somewhat more exalted, more germane, more grave, more Buddhist or at least more Sanskritic than the rather common kipas/tipas. If the proposal of Kusen must be rejected, no obvious substitute suggests itself. The first word, Panarabwan, could represent either a Sanskrit vocative or an Old Javanese or Old Malay subject, but the second word seems to defy comprehension as either Sanskrit or one of the island languages.24 A mistransliteration of

the second word on the part of Soehamir remains a vague possibility; the letters in his published facsimile are too congested at this point to confirm that the first character of the second line is indeed a ‘kha’, though in the absence of further evidence I think we should credit Soehamir with the ability to look up his akßara on Holle’s chart of specimen forms. Possibly the metal scribe erred; the initial akßara could be a consonant prefixed with the vowel ‘e’

to make a form like ‘wenipas’, but this again seems to lead nowhere. Finally, the fact that the first line, ‘panarabwan’ is written inside the formal confines of the ‘Δ’, while the second line ‘khanipas’ is written in the bubble between the ‘Δ’ and the consonant may suggest a conceptual separation between the two words; could we someday, for example, be apprised of the fact that Panarabwan’s queen was named Dya˙ Khanipas? This must remain mere

speculation in the absence of more data, either from outside sources or from a renewed inspection of the mantra itself. We presume from its context that ‘khanipas’ involves either Panarabwan’s metaphysical beliefs or the mundane details of his construction of the gateway. The former is the more attractive supposition because it appears that Panarabwan, by inscribing his name within the dot of the ‘i’, has infixed his name as a vital component of the sacred mantra. In this way, he is operating within the circumference of the mystic vowel itself, and clearly intends to link himself to the mantra or the cosmic being it points to, albeit in a manner we cannot now understand. A question that can only be considered with the acknowledgement that no definitive answer may ever be found is the question of why Panarabwan chose to embed this particular mantra near the great gate leading up to the Ratu Baka plateau. As has

been examined above, the mantra served as both

23 I wish to thank A. Teeuw and Manu Jayaatmadja for offering constructive consultation on issues of Old Javanese grammar. Any faults are entirely my own. 24 Possibly a scholar with the requisite background in Old Khmer or Old Sinhalese may turn up a plausible interpretation, with fascinating implications for the history of Java. If the word should transpire to be Old Javanese, then the Ratu Baka mantra likely contains the oldest written specimen of that language.

the ‘compulsory’ mantra of the Tathāgatas as well as the personal h®daya mantra of Vajrapāøi, and furthermore that the harshness of this disciplinary mantra seems softened by the adjunction of the benedictory svāhā. Do we therefore see in Panarabwan’s choice of mantra an attempt to coerce cosmic beings

to descend on the Ratu Baka plateau or else an homage to Vajrapāøi? (We may for the sake of thoroughness posit the additional and unverifiable hypothesis that Vajrapāøi was considered a personal deity for Panarabwan and that his signature mantra was placed within every edifice Panarabwan endowed. Much more archaeological luck is required before such a hypothesis could be accepted). Without knowing exactly what lay on the plain of the Ratu Baka at the turn of

the eighth century other than a cremation temple,25 a great double-layered gate, and much farther to the southeast a Buddhist vihāra of the Abhayagirin order,26 we might initially surmise that the gate – so proximate to the location of the unearthing of the mantra – might have some connection with a summoning of divinities through it.27 In this idea of the gate as a channel or conduit of mantrically summoned deities, I am particularly attracted by the notion of the establishment of a maø∂ala of statues, long since disappeared, on the great terrace of the Ratu Baka.28 Lokesh Chandra (1995:219) infers the existence of a large statue of Avalokiteśvara on the Ratu Baka from his study of the Abhayagirivihāra inscription, noting that it would complement the statues of Tārā and MañjuśrΔ which were known to be already positioned on the plain. There is some archaeological evidence for the possibility that

statuary was erected on the primary terrace,29 for in the very paragraph of Soehamir’s report which follows his report of the excavation of the mantra, he notes that an iron sword

25 The funereal character of this open-topped temple was noted by Soehamir in Oudheidkundig verslag (1950:33-7), where he documented the deep, open temple pit with remains of charcoal and ash. 26 The Sri Lankan archaeologist Deraniyagala noted substantial similarities between features of the pendopo at the southeast end of the Ratu Baka plateau and the ascetic monasteries outside the Abhayagiri in Sri Lanka (Miksic 1993:25). 27 The theme of

a summoning recurs in several mantras recovered from ›ailendra-era Java. I hope to take this topic up at some length in a future publication. 28 In light of the efforts by the Khmers to throw off the Javanese from their lands culminating in a well-documented 802 religious ceremony undertaken by Jayavarman II, who was possibly a former hostage of a Javanese king, to effect Khmer freedom from Java, the Sab Båk inscription of 1066 mentions a topic which cannot be ignored. In it, the author mentions that nine statues of Buddha-Loke¢vara were erected long before on the Abhayagiri by the magician

Kamsteü ›rΔ Satyavarman in order that Java cannot oppress the lands of Kambuja. These statues were refurbished by the descendent of the founder after they had fallen into ruin. One speculates whether the hypothesized Ratu Baka maø∂ala was not some kind of merit-payment by the Khmer royal family as a condition for their return to their homeland. Interestingly, this Cambodian Buddhist inscription contains references to the Vajrayåna and provides tantric names for the Buddha. See Prapandvidya 1990. 29 In a response to an early draft of this essay, David Snellgrove noted that he expected that a loose gold leaf

mantra of this type would be associated with a statue rather than a building.

of length 60-70 cm, including handle, ‘of a type which is depicted in ancient Javanese reliefs as an attribute of a god’, as well as an iron dagger-head, were found by a villager in the field near the so-called ‘kraton’ of the Ratu Baka (Oudheidkundig verslag 1950:37). However, considerations of the gate as

a channel through which deities were summoned seem to poorly account for the features of the mantra, which carry the features of a benediction rather than a command. Furthermore, the vajrāükuśa, symbol of the very first summoning rite in the performance of a tantric ritual, is associated with the eastern gate, the proper place to begin the pradakßina of the maø∂ala, while the western gate is associated with the fetter. (As Snellgrove (1987:223) notes, this

requirement is made explicit in the Sarvadurgatipariśodhana). Therefore, it is the eastern rather than the western side of a maø∂ala which is appropriate for the summoning implied by the mantra. Another possibility presents itself, that the mantra was placed in the gate but that this gate was naught but a literal accessory, an adjunct to the cremation-ground character of the Ratu Baka plateau. In this regard, the benedictory character of the mantra begins to

have some meaning: Snellgrove (1987:237) notes that the Vajra Family, of which Vajrapāøi is the chief, possesses the power of not just death and destruction, but also of revivification. Indeed, as Nihom (1994:109) remarks in his comments on the maø∂alas of the yogatantric Sarvadurgatipariśodhana, the goal of a consecration in the maø∂ala of Vajrapāøi is eternal life. In this context, the benedictory nature of the mantra makes more sense and

furthermore might better accord with the westward orientation of the gate and the westward-opening cremation temple: the West is the direction of death and the dying of the sun, and it is Vajrapāøi who might be supplicated to amend the unwanted passing away of a life.30 As for the identity of the deceased who merited the erection of a crematorium upon an elaborately and extensively prepared ground,31 we must suppose that it was a great king, perhaps the dynastic

forebear Sañjaya whose military camp or possibly even his capital lay close to the foot of the Ratu Baka escarpment,32 30 A serious opposition could be raised to this interpretation: the mantra used for the revival of Mahe¢vara by Vajrapåøi in the tale of the Trailokyavijaya is vajrayu˙. However, it is unlikely that Panarabwan found that he could actually implement a do-it-yourself resurrection with this mantra.

In the absence of personal ability to revive the dead, it is possible that Panarabwan fell back on an implication to Vajrapåøi using his h®daya. 31 The reader should appreciate that the present form of the Ratu Baka plateau is a man-made feature leveled by the titanic carving out of a vast flat terrace from a craggy limestone hillock. Such extensive preparation of the terrain is also a feature of the construction of the temples of Barabudur and Prambanan.

In the case of the Ratu Baka, I estimate that a quantity minimally approaching 25,000 cubic meters of callous material was carved out of the bedrock. Luckily for the peasant workers the rock is a reasonably soft limestone. 32 That Sañjaya apparently had a camp (tarub, literally a roof or tent) near the foot of the Ratu Baka hill is mentioned in the inscription of Taji Gunung of 910 AD; the Ratu Baka hill therefore may have had some historical or

spiritual significance to him. See Sarkar 1971-72, II:129. In this

perhaps Panaraban’s predecessor Panangkaran whose Tārā temple lay within sight, or perhaps even Panarabwan himself who was cremated there on the plateau. One suspects that a monastery of the highest renown and fame was endowed on behalf of the merit of the deceased king, and the fragments of the

Abhayagirivihāra inscription would therefore point to Panangkaran as the dead and lamented mahārāja.33 The implications of royal support for this mantra: Vajra-Buddhism in Java The implications of the existence of this mantra for the understanding of the religious, cultural, and political history of ancient Central Java are substantial. While the ramifications are too great to be dealt with in depth here, I

would like to stake out the following four points as some of the most important. First, literature of the yoga tantra class, and with that the maø∂alas associated with the Pañcajina, were probably known at a fairly early stage in Javanese history, at least by 803, probably earlier than 782, possibly earlier than 746.34 Nihom (1995:69-115) raised interesting arguments based on his discovery of odd vestiges of yoga tantra class mantras and maø∂alas on

the island of Bali; to him, they suggested that the entire tradition embodied in the root class mūla-tantra, the STTS, was not known as a whole by the Javanese, but rather the maø∂alas and mantras came to the archipelago and were passed on to Bali as individual components at a time before they had been composited into the STTS. Nihom, furthermore, remarked that their origin may have been ŚrΔ Vijayan rather than Indian. The new research results cannot

negate Nihom’s conclusion but render it very unlikely, as the literary chronicle of Chinese Buddhism clearly documents that relatively mature versions of the STTS, including the second section on the story of the Trailokyavijaya, were being shipped under imperial Chinese patronage from South India and Sri respect, I wish to point out possible parallels with a nomenclature from contemporary India, where the great Påla king Devapåladeva issued at least two

inscriptions dating from six year’s apart (Munghir from the 29th and Nålandå from the 35th regnal years) from his ‘victorious camp at Mudgagiri’, connoting the impression that this Påla ‘camp’ was something more than a transient structure. Sañjaya too may have had a capital at the foot of the Ratu Baka, the reason for all of the great stone construction which was later erected there. 33 Although likely but not definitively deriving from East Java, the

short inscription on a silver plate, transcribed as no. 21 in Damais’ list in the Répertoire onomastique and dated by him on paleographic grounds to between 775-825 AD, cannot but attract one’s attention. In it, there is reference to a bha†ara kakΔ¥ abhaya, which probably means ‘the Lord Grandfather at Abhaya’. 34 It should be noted that 746 was the year that Amoghavajra sailed back to China from his Indian and Lankan text-gathering trip, and as well the year of accession of the Javanese Buddhist king Panangkaran to the throne.

Lanka to China at least thirty years before Panarabwan began to cite one of its prominent mantras. Although he was famously dedicated to the teachings of the STTS, Vajrabodhi’s primary manuscript was lost on the 720 sea voyage to China. This text was so important that his disciple Amoghavajra, who significantly met his mentor (and fellow STTS commentator) Vajrabodhi on the island of Java in 717,35 ventured forth from Chang’an to the South of India in

search of a superior manuscript, apparently finding one in Sri Lanka.36 Seemingly the direct connection between the Indian and the Javanese Buddhists is sufficiently firm37 for us to disbelieve that the Javanese remained in ignorance of the STTS as a whole and composite work, though how and why such a haphazard representation of it remained in modern Bali is a mystery left to be explained. Second, the type of Buddhist ritual and philosophical innovation

that inspired texts of the yoga tantra class is well-described in Snellgrove’s IndoTibetan Buddhism (1987, see particularly the extensive Chapter 3: Tantric Buddhism). The theological and ritual concomitants are the primacy of maø∂ala and mantra; the substitution of Mount Sumeru for the Bo-Tree as locus of the paradigmatic enlightenment-giving ritual consecrations and therefore as the ultimate centre of religious activity and holiness in the Buddhist

universe, the corresponding importance of the rite of consecration into the mysteries of the maø∂ala, and the ordering of the Five Jinas into a set under the primacy of Vairocana. All of these ideas are sealed under the newly formulated symbol of the vajra. Given Panarabwan’s acquaintance with this material at a datably early phase in Javanese history as well as the 782 Śailendra inscription of Kƒlurak, a tantric work of uncertain ideological provenence38

which also explicitly subordinates the three Hindu supreme gods to a Buddhist deity, we may safely say that the grounds are open to find expression of these tantric convictions in the stone temples of Java. For those scholars who have perceived the influence of the Vajradhātu maø∂ala and Vajrayāna Buddhism in general at such Central Javanese monuments as

35 Yi-Liang (1945:321-2) notes that based on chronological considerations contained within the writings of Amoghavajra himself, the two Chinese biographies which hold that Amoghavajra met Vajrabodhi in China must be considered wrong; the two clearly met before their arrival in China in 720. 36

The biographies of Amoghavajra say that he requested instruction by the Lankan teacher Samantabhadra on the doctrine of yoga tantras and collected over 500 Buddhist texts and commentaries on his journey. See Yi-Liang (1945:292). 37 In this context, I wish to remind the reader of Bosch’s paleographic study

(1928:3-16) of the 782 AD inscription of Kƒlurak, in which he concluded that the Javanese specimens of the Siddhamåt®ka script were entirely current in northern India at that time. The inscription, heavily tantric in its inspiration, also mentions an Indian guru. 38 Nihom (1994:70, note 128) has suggested that the yoga-tantric DharmadhåtuvågΔ¢varamaø∂ala may explain some of the references within the Kƒlurak inscription.

Barabu∂ur, the Ratu Baka mantra foil seems to provide support in that the requisite classes of Buddhist texts were known in Śailendra-era Java during a time of the Buddhist constructions on the plains of Prambanan. If further research into the ideological background of Barabu∂ur and other monuments

suggests that there is no connection between the stones which comprise the monuments and a vajra-centric Buddhist inspiration,39 then this itself may be taken as an argument suggesting the earliness of their construction, during the period before the 782 tantric inscription of Kƒlurak. In passing, we note that although its composition is formally dated to the reign of Siø∂ok, there is no strong reason why the constituent verses of the Sang Hyang

Kamahāyānikan were not known from the time of Panarabwan or even earlier: that text combines instructions and consecration rites taken from the Caryā and yoga tantras, and the root text of this yoga class was very likely already known in Java more than a century and a half before Siø∂ok.40 Third, implicit in the choice of mantra used by Panarabwan is a story of some degree of tension between the Buddhist yoga tantrists and the Śaivites whose god was mythically

subdued, murdered, resurrected, and converted. Linrothe (1999:186), writing from a Tibetan context, sees in the tale of Trailokyavijaya a story of psychological progression, noting that ‘As far as the text is concerned, Mahesvara’s “wickedness” consists of his wrongful pride’, and continuing: The punishment that is inflicted on Mahesvara just before he undergoes death and rebirth is another indication that the authors of the text intended to

highlight pride, arrogance, and self-centeredness as the principal obstacles in Mahesvara’s behavior. He and his consort are made to suffer what the prideful fear most: humiliation. They are thrown on their backs, legs in the air, their nakedness revealed. In this degraded state, they are the objects of derision for their former minions and the entire assembly. (Linrothe 1999:187.)

While this may be the lesson of the story in the Tibetan context where there are no Śaivas who might take offense at the story,41 the religious milieu encoun

39 See Klokke (1995) and Snellgrove (1996) for a rather convincing exploration of the faults of interpreting Barabu∂ur as a maø∂ala. Their fundamental observations that Barabu∂ur contains no overt vajra-symbolism and that no known maø∂ala multiplies its Buddhas remain potent. It is indisputable, however, that at some time Buddhists with a vajra-oriented penchant worshipped at the Barabu∂ur, leaving behind at least one gold foil which documents some of their

mantras and dhåraøΔs; for this, see Boechari (1976). It is furthermore not impossible in my view to see in the mudrås of the stupa-encased Buddhas (Huntington 1994:153, Figure 2 terms it a ‘Javanese variant’ of the dharmacakramudrå) a reference to the symbolism of the vajra. 40 The Sang Hyang Kamahāyānikan has been extensively edited, with collocated passages from the Sanskrit, Tibetan and Chinese, by Lokesh Chandra (1995:295-434). 41 Mayer

(1998:308) notes that because of the remoteness of Tibet from the Hindu holy men of India, ‘›aivism remained of a mainly abstract and symbolic value within Tibetan thinking’.

tered by Panarabwan must have been entirely different. All indications suggest that king Sañjaya, considered by the kings of the early tenth century to have been the well-spring of their lineage, was a committed devotee of Śiva. In adopting his Buddhism, and more importantly in scribing his name to a

plaque of anti-Śaivite character, Panarabwan seemingly repudiated in strong terms the religion of his probable ancestor Sañjaya. The implications of this rejection are difficult to determine clearly, but it is hard to envision a scenario by which the Śaiva elements in the court and society were not offended

by the nature of the new religion, its story of Trailokyavijaya, and its willingness to humiliate their high god.42 Fourth, the Ratu Baka plateau itself is likely given over to Buddhism at this early stage in Javanese history; possibly it was a giant plateau devoted to the Abhayagirivāsins, the doctrines they

inculcated, or the rituals they were tasked to superintend. If, as folk-legend has it, the Ratu Baka served the Javanese as a kraton, it was almost certainly not the kraton during the early Śailendra era. However, the king ‘Boko’, who figures in the folk-myth of the founding of Prambanan, must have had

some historical basis because his name appears everywhere in the hamlets around the plateau, where the local district is named Bokoharjo. Interestingly, no village or district of this name is available from the number of ancient administrative inscriptions recovered from the Prambanan area.43 If some later

potentate set up operations on the plateau, impressing his name both literally upon the countryside and mythically upon the monuments, he could only be a rump king, some local despot made good in the distant aftermath of the great collapse of Central Javanese government around 929 AD.

Conclusions and notes on future directions for research At the end of this examination of the Ratu Baka mantra, the study of which has hopefully only just begun, a stock-taking should be performed to assess what has been gained and what is left to be developed. On the positive side, the existence of the Ratu

Baka mantra demonstrates the existence of a deeply personal royal involvement with Vajrayāna Buddhism dating to the period around 800 AD or possibly twenty years before. As a byproduct of this dating and the location of the mantra, strong evidence is supplied which tends to confirm the notion that the Raka of

42 One referee of this paper offered the interesting suggestion that if the Ratu Baka mantra was to be taken as a Buddhist memento mori, that perhaps it was the ›ailendras who buried a Hindu Panarabwan. The anti-›aiva nature of the mantra suggests to me that such a rude choice of text by the ›ailendras would be an insult rather than a boon for the deceased. 43 The table of Kusen (1995) updates the pioneering work of Stutterheim in this regard.

Panarabwan was coronated as a Śailendra king under the name Samaratuüga: this in turn suggests that ancient Central Java was ruled by only a single dynasty

at the time that its greatest Buddhist monuments were built. The presence of Vajrayāna narratives during this period, demonstrably part of a wider contemporary pattern of diffusion of tantric texts to Tibet, China, Korea, and Japan, should serve as support to those scholars who wish to see tantric inspirations for the monuments of Buddhist Java. Elements of these tantric beliefs may have promoted disharmony with Hindu believers. On the side of research waiting to be done, it is hoped that some scholar with the necessary philological talents will undertake a search of manuscripts and commentaries of the STTS and similar tantric works in Sanskrit, Tibetan, Nepalese, Chinese, Khotanese, Bhutanese, or Japanese, looking for other instances of the Ratu Baka mantra (with its interposed hūµ between the †akΔ and the ja˙). In particular, the early material brought by Vajrabodhi and Amoghavajra to China

deserves a survey as a likely source. Similar linguistic attention should be brought to bear on the obdurate word ‘khanipas’ to see if it carries any potential meaning in other languages, particularly Sinhalese or Khmer. Further archaeological and theosophical attention should be given to the gates on the Ratu Baka, what lay underneath them and the reason for their orientation to the West. A thorough paleographic description, including photographs and

facsimiles, of the maø∂ala inscriptions of Candi Gumpung, Sumatra, should also be undertaken. Finally, archival attention should be paid to rediscovering the mantra, or at least exhuming a set of notes which more precisely give the size and find-spot of the gold plate. Acknowledgments A number of individuals have offered assistance to me in the course of researching and writing this paper. In particular, I wish to thank

Julie Gifford, Manu Jayaatmaja, Lokesh Chandra, Bertus Nederlof, David Snellgrove, and Hans Teeuw. The two necessarily anonymous Bijdragen reviewers are to be thanked for a number of suggestions stemming from their close reading of an early version of this paper.

REFERENCES

Boechari, M. 1976 Preliminary report on some archaeological finds around the Borobudur temple, Fifth Consultative Committee Conference for Safeguarding the Borobudur, doc. CC/V/6/76. 1985 ‘Ritual deposits of Candi Gumpung’, SPAFA Consultative Workshop on Archaeological and Environmental Studies on Srivijaya.

Bosch, F.D.K. 1928 ‘De inscriptie van Kƒloerak’, Tijdschrift voor Indische Taal-, Land- en Volkenkunde 68:1-64. 1952 ‘ÇrΔvijaya, de Çailendra- en Sañjayavaµça’, Bijdragen tot de Taal-, Land- en Volkenkunde 108:113-23. Casparis, J.G. de 1950 Prasasti Indonesia I; Inscripties uit de Çailendra-tijd. Bandung: Nix. 1956 Selected inscriptions from the 7th to the 9th century A.D. Bandung: Masa Baru. [Dinas Purbakala Republik Indonesia, Prasasti Indonesia 2.] 1961 ‘New evidence on cultural relations between Java and Ceylon in ancient times’, Artibus Asiae 24:241-8 1981 ‘The dual nature of Barabu∂ur’, in:

Luis O. Gomez and Hiram W. Woodward (eds), Barabudur; History and significance of a Buddhist monument, pp. 47-83. Berkeley, CA: University of California Press. [[[Berkeley]] Buddhist Studies Series 2.] Chandra, Lokesh 1995 Cultural horizons of India. Vol. 4: Studies in Tantra and Buddhism, art and archaeology, language and literature. New Delhi: International Academy of Indian Culture and Aditya Prakashan. [Satapitaka Series 381.] Damais, Louis-Charles 1968 ‘Bibliographie indonésienne: XI. Les publications épigraphiques du Service Archéologique de l’Indonésie’, Bulletin de l’Ecole Française d’Extrême-Orient

54:295-522. 1970, Répertoire onomastique de l'épigraphie javanaise (jusqu’à Pu Siø∂ok ŚrΔ Ôśānawikrama Dharmmotuügadewa; Étude d’épigraphie indonésienne. Paris: Ecole Française d’Extrême-Orient. [Publications de l’Ecole Française d’Extrême-Orient 66.] Davidson, R. 1991 ‘Reflections on the Maheśvara subjugation myth; Indic materials, saskya-pa apologetics, and the birth of Heruka’, Journal of the International Association of Buddhist Studies 14-2:197-235. 1995 ‘The Bodhisattva Vajrapāøi’s subjugation of Śiva’, in: Donald S. Lopez Jr (ed), Religions of India in practice, pp. 547-55. Princeton, NJ:

Princeton University Press. [[[Princeton]] Readings in Religions.] Goepper, Roger 1993 Aizen-myōō, the esoteric king of lust; An iconological study. Zürich: Museum Rietberg. [Artibus Asiae, Supplementum 39.] Holle, K.F. 1877 Tabel van oud- en nieuw- Indische alphabetten. Buitenzorg: Lang.

Huntington, John C. 1994 ‘The iconography of Borobudur revisited; The concepts of ßleßa and sarva[[[buddha]]]kāya’, in: Marijke J. Klokke and Pauline Lunsingh Scheurleer (eds), Ancient Indonesian sculpture, pp. 133-53. Leiden: KITLV Press. [Verhandelingen 165.] Iyanaga, N. 1985 ‘Récits de la soumission de

Maheśvara par Trailokyavijaya d’après les sources chinoises et japonaises’, in: Michel Strickmann (ed.), Tantric and Taoist studies in honour of R.A. Stein, Volume 3, pp. 633-745. Bruxelles: Institut Belge des Hautes Etudes Chinoises. Jong, J.W. de 1974 ‘Notes on the sources and the text of the Sang Hyang Kamahāyānan Mantranaya’, Bijdragen tot de Taal-, Land-, en Volkenkunde 130:465-82. Jordaan, Roy E. 1999 The Śailendras in Central Javanese history; A

survey of research from 1950 to 1999. Yogyakarta: Universitas Sanata Dharma. Klokke, M. 1995 ‘Borobudur: A maø∂ala? A contextual approach to the function and meaning of Borobudur’, IIAS Yearbook, pp. 191-219. Leiden: International Institute for Asian Studies. Kusen 1994 ‘Raja-raja Mataram Kuna dari Sanjaya sampai Balitung; Sebuah rekonstruksi berdasarkan prasasti Wanua Tengah III’, Berkala Arkeologi 14: 82-94. 1995 ‘Identifikasi toponim dalam prasasti Jawa

Kuna abad IX-X dari Prambanan dan sekitarnya dengan toponim masa kini’, in: Analisis hasil penelitian arkeologi; Analisis sumbat tertulis masa klasik, pp. 197-210. Jakarta: Puslitarkenas. Linrothe, Rob 1999 Ruthless compassion; Wrathful deities in early Indo-Tibetan esoteric Buddhist art. Boston, MA: Shambhala Li-Yiang, C. 1945 ‘Tantrism in China’, Harvard Journal of Asiatic Studies 8:241-332 Matsunaga, K. 1998 Some bronze statues of maø∂ala excavated from Java. [[[Self]] published manuscript found in the library of the Lab Arkeologi, Universitas Gajah Madah.] Mayer, R. 1998 ‘The figure of Maheśvara/Rudra in the riü-ma-pa tantric tradition’, Journal of the International Association of Buddhist Studies 21-2:270-310. Miksic, J. 1993 ‘Double meditation platforms at Anuradhapura and the pendopo of Ratu Boko’, Saraswati Esai-esai Arkeologi. Kalpataru; Majalah Arkeologi 10:23-31.

Naerssen, F. van 1947 ‘The Çailendra interregnum’, in: India Antiqua; A volume of Oriental studies presented by his friends and pupils Kern-Institute Leyden to Jean Philippe Vogel on the occasion of the 50th anniversary of his doctorate, pp. 249-53. Leyden: Brill. Nihom, Max 1994 Studies in Indian and

Indo-Indonesian tantrism; The Kuñjarakarnadharmakathana and the Yogatantra. Vienna: Gerold. [Publications of the Nobili Research Library 21.] Oudheidkundig verslag 1950 Oudheidkundig verslag 1948. Bandung: Nix. [Uitgave van de Oudheidkundige Dienst in Indonesië in samenwerking met het Koninklijk Bataviaasch

Genootschap van Kunsten en Wetenschappen.] Prapandvidya, C. 1990 ‘The Sab Bāk inscription: evidence of an early Vajrayāna Buddhist presence in Thailand’,

The Journal of the Siam Society 78-2:11-4. Sarkar, Himansu Bhusan 1971-72 Corpus of the inscriptions of Java; (Corpus inscriptionum Javanicarum); (up to

928 A.D.). Calcutta: Mukhopadhyay. Two vols. Shashibala 1986 Comparative iconography of the Vajradhātu maø∂ala and the Tattva-saügraha. New Delhi: Sharada Rani. [Satapitaka Series 344.] Skorupski, Tadeusz 1983 The Sarvadurgatipariśodhana tantra; Elimination of all evil destinies. New Delhi: Motilal Banarsidass. Snellgrove, David L. 1981 ‘Introduction’, in: Lokesh Chandra and David L. Snellgrove (eds), Sarva-tathāgata-tattva-saügraha; Facsimile reproduction of a tenth century Sanskrit manuscript from Nepal, pp. 5-67. New Delhi: International Academy of Indian Culture. [Satapitaka Series, Indo-Asian Literatures 269.] 1987 Indo-Tibetan Buddhism; Indian Buddhists and their Tibetan successors. Boston, MA: Shambhala. 1996 ‘Borobudur: stūpa or

maø∂ala’, East and West 46/7-4:477-83. Sukarto K. Atmodjo, M. 1994 ‘Beberapa temuan prasasti baru di Indonesia’, Berkala Arkeologi 14: 1-10. Tucci, Giuseppe 1932 Indo-Tibetica. Roma: Reale Accademia d’Italia. [Studi e Documenti 1.] Vogel, J.Ph. 1919 ‘Het koninkrijk Çrivijaya’, Bijdragen tot de Taal-, Land- en Volkenkunde 75:626-37. Willson, M. and M. Brauen 2000 Deities of Tibetan Buddhism; The Zurich paintings of the icons worthwhile to see. Boston, MA: Wisdom Publications.

Wisseman-Christie, J. 2001 ‘Revisiting early Mataram’, in: Marijke J. Klokke and Karel R. van Kooij (eds), Fruits of inspiration; Studies in honour of Prof. J.G. de Casparis, retired Professor of the early history and archeology of South and Southeast Asia at the University of Leiden, the Netherlands on

the occasion of his 85th birthday, pp. 25-56. Groningen: Forsten. [[[Gonda]] Indological Studies 11.] Yamada, I. 1981 Sarva-tathāgata-tattva-saügraha-nāma-mahāyāna-s√tra; A critical edition based on a Sanskrit manuscript and Chinese and Tibetan translations. New Delhi: Sharada Rani. [Satapitaka Series, Indo-Asian Literatures 262.]