The Indian Ocean Itinerary of Buddhaguptanātha’s Namthar
Iain Sinclair
Throughout the sixteenth century, the yogi Buddhaguptanātha (b. 1514) travelled across the Indian subcontinent, the Hindu Kush, and the Indian Ocean in search of enlightenment. Late in life, in the year 1590, Buddhaguptanātha crossed the Himalayas and related his travel stories to the impressionable young Tibetan monk Kunga Nyingpo Tāranātha (kun dga’ snying po tā ra nā tha, 1575–1634). Several years later, Tāranātha rewrote his notes and memories of this meeting into The Liberation of the Great Adept Buddhagupta: From the Words of His Lordship, Unsullied by a Jot of My Ideations—A Genuine Account (grub chen buddhagupta’i rnam thar rje btsun nyid kyi zhal lung las gzhan du rang rtog gi dri mas ma sbags pa’i yi ge yang dag pa’o). As this work locates itself within the Tibetan genre of hagiography or story of “liberation,” the namthar (rnam thar), it will also be referred to here with the short English title Hagiography. In contrast with better known Age of Sail travelogues that focus on trade, conquest and discovery, Buddhaguptanātha’s journey is concerned with special spiritual experiences and where to find them. The Hagiography also happens to contain much reporting on places, people, travel routes and other realia of the period. It is precisely the religious bent of the narrator and his subject that has frustrated modern efforts to get useful information out of the text. The Hagiography’s view of sixteenth-century Asia as filled with tantric Buddhist wonders has been seen as wishful thinking, whether on the part of Buddhaguptanātha or Tāranātha. If we doubt that Nāgārjuna had walked on remote Indian Ocean islands, for instance, or that Vajrayāna priests gave tantric initiations in a Portuguese church of St. Lawrence, what else in Tāranātha’s narrative we can accept?
Previous Studies of Buddhaguptanātha’s Hagiography
Buddhaguptanātha’s Hagiography has received its share of attention in modern Tibetan studies, but no complete study or translation has been published to date. The Hagiography was first noticed in modern scholarship in a short 1894 article by L. Augustine Waddell (1854–1938) that conveyed the gist of the text. Most modern scholarly literature on Buddhaguptanātha refers to an influential 1931 article by Giuseppe Tucci (1894–1984), in which Tucci translated and paraphrased long excerpts from the Hagiography and reflected in some detail on its travel itinerary. Tucci also briefly returned to the text in his later writings. More recently, a complete, annotated English translation of the Hagiography was drafted by the late scholar Hubert Decleer (1940–2022). This draft has circulated privately among some of Decleer’s confrères since at least 2006 but has not, to date, been published. Decleer’s translation is also prefaced with a critical review of earlier studies on Buddhaguptanātha. Decleer referred to a 1993 Tibetan summary of the namthar by the venerable Dzongtsé Jampa Thubten (rdzong rtse byams pa thub bstan)—unseen by this author—and to work in progress on the text presented by David Templeman at a Tibetological seminar in 1995. Decleer contested what he saw as the view that Buddhaguptanātha had “Śaiva leanings” and lacked the Buddhist credentials attributed to him by Tāranātha. David Templeman went on to put his own full translation of the Hagiography at the center of his 2008 doctoral dissertation on Tāranātha’s India-oriented oeuvre. This also remains unpublished at the present time. The Hagiography has, then, been the subject of much erudite commentary and scrutiny, even though the most recent work has not yet surfaced in public.
So far there has been no attempt to retrace Buddhaguptanātha’s route at every step and to determine whether the Hagiography describes plausible long-distance journeys. In this regard, the question of Buddhaguptanātha’s bona fides, touched on in previous studies, is tied up with the veracity of his narration. How trustworthy is the Buddhist lore attributed to people who lived in places where Buddhism had been all but extinct for centuries?[1] In this study, only small inroads can be made into the mammoth literary edifice of sixteenth-century Indic Buddhism constructed by Tāranātha in his Seven Instruction Lineages (bka’ babs bdun ldan), History of Buddhism in India (rgya gar chos ’byung), Life of Kṛṣṇacārya (slob dpon chen po spyod ’chang dbang po’i rnam thar) and Origin of Tārā Tantra (sgrol ma’i rgyud kyi ’byung khungs), to name a few of the titles connected to the Hagiography by their author and subject matter. The aims of this study are confined to reconstructing the part of the itinerary spanning the Indian Ocean world.
Reading Tāranātha’s Histories: Some Background
The task of extracting credible information from Tāranātha’s writings on Buddhism in the Indic world poses certain challenges. Many of his sources are unavailable, many of his claims are unverifiable, and much of what he describes is unknown to others. His writings aim to exploit his position as the sole inheritor of a body of lore and teaching brought to Tibet by Indian yogis at a very late stage of South Asian Buddhism, as David Templeman and others have discussed.[2] Tāranātha maintains that he was conversant in what he calls the vernacular of “Madhyadeśa,”[3] and says that he had “become extremely well versed in Sanskrit” and various Indian dialects, “experiencing only a few difficulties” in his conversations with South Asians.[4] According to an independent account of an Indian pundits’ meeting attended by Tāranātha, the participants conversed in Sanskrit “in a quite excited manner.”[5]
While Tāranātha presents his narration of Buddhaguptanātha’s travels as an accurate, unembellished account, Tāranātha is very much the author and creator of the Hagiography as a Tibetan literary work and a coherent, chronologically ordered narrative. Tāranātha often disclaims responsibility for the narrative’s more far-fetched elements, and asserts that they are not products of his personal reveries or dream states, which feature in some his other writings. If we take Tāranātha at his word, we are led to diagnose problems in the itinerary as cases of misremembering, misredaction, miscomprehension, or erroneous transmission, and not necessarily as mere fantasy—although we cannot yet rule that out. It is, of course, the fantastic elements in the narrative that elevate Tāranātha into the position of direct confidant of a living, powerful Indian tantric guru—the very kind of personality that has driven the development of Tibetan Buddhism since its inception.
Tāranātha describes taking notes of Buddhaguptanātha’s spoken teaching, which were, it seems, later re-redacted to create the Hagiography.[6] With up to twelve years intervening between the initial meeting and its writing up, and the difficulties of keeping track of an orally described, unfamiliar topography, Tāranātha was at risk of misreading his notes or inserting them into the narrative at the wrong place.[7] Cases of travel directions needing to be exactly reversed, for instance, have been noticed by Decleer.[8] As will become clear, problems often come up in trying to map the directions of the Hagiography onto a recognizable historical reality.
The Itinerary’s Place in the Hagiography
In order to provide some context for the analysis that follows, a short précis of the Hagiography will be provided here. Tāranātha begins by narrating the great yogi’s birth into a merchant family in a town near Rameswaram in present-day Tamil Nadu.[9] In his childhood, Buddhanātha, as he was then called, studied Sanskrit with the aid of the Chandaḥsūtra of Piṅgala and the Cāndravyākaraṇa. While still young, he joined a local group of yogis following the tradition of Gorakṣanātha. This group was affiliated to the Nāṭeśvarī panth, which is claimed to have been Buddhist in affiliation. This phase of his life is said to have been partly contemporary with the reign of the Vijayanagara ruler Rāma Rāya (r. 1542–1565).[10]
Buddhanātha travelled with his fellow yogis to various pilgrimage spots in the north and far northwest of the subcontinent: Delhi, Kashmir, Afghanistan, the Punjab, and so on. Here he encountered various local goddess shrines, fakirs, thugs, and yoginīs, and through these encounters enjoyed some meditative breakthroughs.[11] He then returned to South India, travelling through its many smaller districts. The next thirty percent or so of the Hagiography is taken up with the yogi’s maritime journeys. This phase of his life lasted at least eight years. During this period he receives the initiations of Cakrasaṃvara and Hevajra and meets his final guru, the accomplished Buddhist yogi Śāntigupta, who bestows the initiatory name Buddhaguptanātha.[12] Now advanced in years, he continues wandering as Śāntigupta’s disciple.
In the final part of the Hagiography, Tāranātha reflects on his own involvement with Buddhaguptanātha. He recalls the yogi’s journey via Lhasa to his hermitage at Narthang (snar thang). Tāranātha had anticipated that the two would meet in a dream the night before, and when Buddhaguptanātha arrived, Tāranātha immediately asked for all the instruction he could get.[13] The dozens of topics of teaching received from Buddhaguptanātha are then described in the style of a senyik (gsan yig) inventory.[14] Soon, however, Buddhaguptanātha broke off his teaching due to Tārānatha’s “excess of discriminative thought” (rnam rtog shas che ba) and began to make his way out of Tibet. Tāranātha goes on to reminisce about Buddhaguptanātha’s yogic techniques, feats and subsequent career, as related to him by the yogi’s Indian students in Tibet. Tāranātha concludes with verses of praise to Buddhaguptanātha, including his past-life incarnations. In his colophon Tāranātha records his age at the time of writing: twenty-seven years old.
The Indian Ocean Itinerary
This study examines fourteen segments of the Indian Ocean itinerary narrated in Tāranātha’s The Liberation of the Great Adept Buddhagupta, or Hagiography. For easy reference, citations of the Tibetan text refer to the 2008 Pedurma (dpe bsdur ma) edition (Dpe) of Tāranātha’s collected works, volume 34, pages 94–126. The pagination of the Phuntsok ling (phun tshogs gling) blockprint (Phun) is also provided; this extract spans folios 545–557 of the blockprint. Other printings and hand copies of the Hagiography have not, to date, been deemed philologically vital. While it would be ideal to examine the itinerary with the aid of a fully edited text and translation, both Decleer and Templeman aver that they found few variants in the printings they used, and Decleer’s or Templeman’s complete translations may well be published in future.[15] All translations from the Hagiography provided here are the author’s own. Citations of the Tibetan Buddhist canon refer to the Dergé (D) or Qianlong (Q) printings.[16]
1. São Lourenço
The first segment of Buddhaguptanātha’s maritime itinerary is the most difficult to locate clearly in sixteenth-century Asia. Here Buddhaguptanātha sails west from the Konkan coast for an unspecified time. He reaches a place that has the local name São Lourenço and the Indic name Ḍamiḍo Dvīpa. He finds an ecclesiastical community there and studies with them for a year. We are then finally told that they practice Vajrayāna Buddhism. The problem here is that no place then called São Lourenço can be associated with tantric Buddhist activity in the same period. Three names are given for this place:
It is called Island of the Flying (’gro lding)[17] in Tibetan; it is written *Drāmiḍadvīpa (ḍa mi ḍo dwī pa) in Indian language; and in the language of the minor isles of barbarians and borderlands people, it is pronounced São Lourenço (sam lo ran so).[18]
The name da mi ḍo is cognate with the words drāmiḍa, Dravidian, and Tamil. This *Drāmiḍadvīpa is, in other words, the “Tamil peninsula,” the southernmost part of India. In Tāranātha’s History of Buddhism, *Drāmiḍadvīpa clearly refers to the Tamil South.[19] But here, it is reached by a “boat heading west across the ocean,” and such a place cannot be part of the Indian subcontinent.
In Tucci’s reckoning, São Lourenço “may as well be one of the small islands along the Koṅkana country. One may think of Goa itself where traces of Buddhism are to be found.”[20] Tucci also observes that Buddhaguptanātha’s description of the religious community at São Lourenço accords much better with “Christian monks, probably Portuguese,” than with any known Buddhist community. Decleer, however, scorned the idea that Buddhaguptanātha would have been “unable to distinguish” between Buddhist and Christian priests.[21] Here the Hagiography says:
The monastic communities are numerous, but there is no proper adherence to the discipline. The clergy wear black garments, and there is much alcohol consumption and so on.[22]
Tucci infers that these are references to Catholic priests in black vestments and to the drinking of wine at Holy Communion, as might have been encountered on Goa.[23] It has since been pointed out that a church of São Lourenço once stood at Goa, but it has not yet been determined whether this church was operating at the time of Buddhaguptanātha’s travels.[24] The nearby coastal fortifications date only from the beginning of the seventeenth century.[25] Furthermore, the identification of São Lourenço with Goa does not solve other problems in this part of the travelogue. Most critically, Goa’s location does not fit the itinerary, which states that São Lourenço is reached after a westward oceangoing journey departing from the Konkan. Goa might fit the point of departure, but not the destination.
The island of Madagascar, which lies in the Indian Ocean to the west and south of the Konkan, had been known as São Lourenço since the beginning of the sixteenth century. In this period, travel between India and Madagascar was common.[26] If the Hagiography refers to Madagascar here, as Tucci later asserted, the direction of travel should have been given as to the southwest rather than to the west, since Madagascar lies quite far to the south as well as to the west of India.[27] However, the Hagiography never specifies the direction of travel as an intercardinal direction, such as southwest. A Portuguese settlement is known to have started up at Anosy, on the island’s southern coast, in 1540, but it was soon abandoned. By all accounts, it is very unlikely that a viable Christian establishment was maintained on Madagascar—especially one willing to accommodate random travelers from India—at the time of Buddhaguptanātha’s travels.[28]
The description of São Lourenço as a major tantric Buddhist center with named active teachers and traditions in no way fits Goa, Madagascar, or for that matter Macau, which will be discussed later. Most studies have then concluded that the Buddhist milieu of São Lourenço is either sheer fantasy or was located in some other place where tantric Buddhism might have survived. The account of Buddhism in São Lourenço should then be examined more closely. It is said that a local paṇḍita called Sumati gave Buddhaguptanātha the oral authorizations (dbang gsan) of Saṃvara and Hevajra, the latter supposedly according to a tradition of *Padmākara (padma ’byung gnas).[29] And:
Generally on that island the tradition of all four tantra classes is uninterrupted, and the transmission of the very highest [class] in particular exists there,[30] apart from the Kālacakra, [which has been] widespread in India: the Kīlaka, the Ten Furies (Daśakrodha) tantra, a plethora of Herukatantra, Vajrapāṇi, Garuḍa, Mātṛkā, Mahākāla and so on. The clan of the Unsurpassed Hayagrīva—completely nonexistent in India—still exists there.[31]
In the sixteenth century, the Kathmandu Valley of Nepal was the only place where a living tantric Buddhist milieu matching this description could be found outside Central Asia. It is very unlikely, however, that Tāranātha included material relating to Nepal by mistake here. The fundaments of Buddhism in Nepal were known to, and yet of next to no interest to, Tāranātha and most of his Tibetan contemporaries. Tāranātha describes in a short sentence Buddhaguptanātha’s journey out of Tibet via Kathmandu, where he is said to have met “some king” (’dzu ’dzu ’ga’ zhig)—at the time, the ruler was Śivasiṃha Malla (r. 1578–1620)—and distributed his offerings to the poor.[32]
Another possibility is that this part of the account was originally part of the description of “Javadvīpa,” which comes later in the itinerary. That part of the narrative has an odd lacuna, as nothing seems to be said about Java itself there. Some of the subtraditions of Buddhist tantrism supposedly extant at São Lourenço still had visible traces on East Java and Bali in the sixteenth century. A number of them are represented at just one site, Candi Jago in East Java—especially if Javanese gate-guardian figures are taken to represent Vajrapāṇi or the Furies.[33] Nonetheless, these traditions would have existed almost wholly in inert form, as defunct temple icons, vestigial worship traditions and so on, and not as the living religion described by Tāranātha. Receiving Cakrasamvara or Hevajra initiation is not known to have been possible on Java in the sixteenth century, if not for almost the whole of the island’s history.
A reinterpretation of the first segment of Buddhaguptanātha’s maritime journey can now be proposed. The only viable elements here are the Konkan coast as the port of departure for a westward ocean journey and the toponym São Lourenço. While travelling around the Konkan, Buddhaguptanātha may have visited a church of São Lourenço at Goa, if it had been in existence at the time, but at this point the Hagiography might also have confused recollections of two different places called São Lourenço; this issue will be raised again later. It is doubtful that Buddhaguptanātha visited Madagascar, which does not match the description of São Lourenço in key details. The putative Vajrayāna community at São Lourenço seems to be an extraneous element, and should instead be located in South India or the Konkan, judging from the references to Buddhist enclaves there elsewhere in the Hagiography and in Tāranātha’s other works.
Furthermore, at the time Tāranātha was writing, in 1602, he had no way of recognizing an account of Christian liturgy for what it was. Jesuit missionaries had not yet made their appearance in Tibet. It can then also be suspected that Tāranātha embellished the account of Christian priests at São Lourenço on the mistaken understanding that it referred to tantric priests, and added a list of traditions that he believed to be associated with black-robed (gos nag, kṛṣṇāmbara) or blue-black-robed (gos sngon, nīlāmbara) Vajrayāna practitioners.[34] For the purpose of reconstructing the itinerary, the whole account of Buddhism at São Lourenço is gingerly put aside.
2. Śaṅkhadvīpa (an Arabian Sea island)
After a year on São Lourenço, Buddhaguptanātha sailed with merchants to a place called Śaṅkhadvīpa. Its name is glossed in Tibetan as Conch Island (dung gi gling). Pauranic descriptions of Śaṅkhadvīpa and other islands surrounding the Indian subcontinent have often been discussed in modern scholarship, but have not yet been successfully mapped to the topography of the Indo-Pacific.[35] The length and direction of Buddhaguptanātha’s journey from São Lourenço is not recorded, but the next segment of the itinerary shows that Śaṅkhadvīpa was located to the north and west of Sri Lanka and the Maldives.
The Śaṅkhadvīpa visited by Buddhaguptanātha is tentatively identified here as Socotra island. Socotra had long been a destination for Indian mariners, although it had fallen well within the Arabic cultural sphere by the sixteenth century. The island’s coastline can be compared to the shape of a conch shell with the shell’s apex aligned to the island’s eastern tip. Although other islands in the Indian Ocean could also be said to have roughly conch-shaped coastlines, the account of the Hagiography offers additional details that match up with Socotra.
The name of a mountain on Śaṅkhadvīpa is given in Tibetan as Abode of the Arising of Bliss (bde ’byung gi gnas). This an attested Tibetan translation of Sanskrit sukhodaya, which could be a mishearing of *sukhadhāra, or a free translation of *sukhatara; these names are presumed cognate names of Socotra.[36] The name heard by Tāranātha could also have involved a homophony between sukha and/or śaṅkha and sūq, the Arabic word for a market as well as for a market town of the period on Socotra. Buddhaguptanātha observes that medicinal plants grew profusely on this mountain. Socotra is one of the most biodiverse regions on earth, and if Buddhaguptanātha, a professed alchemist, had been there, its flora may have struck him as especially exotic and worthy of mention. The dragon’s blood tree (Dracaena cinnabari), for instance, is endemic to Socotra and today grows only on the slopes of its central mountain.[37] Its red resin has medicinal properties and it has been procured as an Indian Ocean trade commodity since ancient times.
On Śaṅkhadvīpa, furthermore, Buddhaguptanātha “saw many people there coming from the continent called Gajanāsa, having faces with elephant-like noses.”[38] The epithet Gajanāsa, “Elephant-nosed,” naturally refers to Western Europeans.[39] Socotra was briefly occupied by the Portuguese in the early sixteenth century. The Portuguese fort built at Socotra had been abandoned by 1511, but Portuguese ships continued to visit or pass the island in later decades. Buddhaguptanātha’s report therefore ought to place Śaṅkhadvīpa in an area under more direct Portuguese control, such as the Kingdom of Hormuz, but no other place matching the name, features or map position described in the Hagiography has been identified.
3. Pālata (a Maldivian atoll)
After staying at Śaṅkhadvīpa for an unspecified, and probably short, period of time, Buddhaguptanātha sailed south to “an island called Pālata” (pā la ta zhes bya ba’i gling) and then east to Sri Lanka. A ship travelling from the Arabian Sea to Sri Lanka would have reached this Pālata in the Laccadives or Maldives, as Templeman has surmised. Several small islands in the Maldives have names that sound similar to Pālata: Filladhoo, Bileddhoo, Bathalaa. Buddhaguptanātha is presumably referring to one of the larger settlements in the Maldives. Felidhoo atoll, for instance, lies in the middle of the island chain directly to the south of its largest city, Malé. In the early seventeenth century, the initial f of present-day Maldivian dialects was still pronounced with a p sound. Felidhoo’s name was transcribed as Polisdous by French travelers who visited in about 1601–1602.[40] If the Tibetan name Pālata refers to Felidhoo, then its spelling would naturally reflect this period pronunciation.[41] The spelling in the Hagiography may also involve mishearing or miscopying on the Tibetan side, as confusions of ta with da and the dropping of subscribed vowel marks are common errors in the transcription of Indic words. The Hagiography says nothing more about Pālata. As the Buddhist religion once practised on the Maldives had long since subsided into sandy ruins by the time of Buddhaguptanātha’s visit, the Maldives can be ruled out as a location of any of the living Buddhist sites mentioned in the itinerary.[42]
4. Sri Lanka
The Hagiography reaches firmer ground when Buddhaguptanātha comes to Sri Lanka via the Maldives. Here Buddhaguptanātha waxes lyrical about the natural beauty of the island and its thriving Buddhist community, visits well-known sites such as Adam’s Peak (śrīpāduka), and claims to have had an audience with the king.[43] The name of this king is recorded by Tāranātha as rā he shing kha bhan ḍa ri, namely, Rajasinghe I, formerly Tikiri Bandara.[44] He was on the throne at Kandy from 1581 (or in effect from 1578, according to the Portuguese) to 1592. Buddhaguptanātha’s five-year sojourn in Sri Lanka would then have started in 1573 at the earliest and ended no later than 1587, since he continued his travels for some years before meeting Tāranātha in Tibet in 1590.[45] This is all plausible and consistent with what is known about the period.
Nonetheless, Tāranātha could not resist including a story about a meeting with a 700-year-old monk called Yaśākaraśānti in a cave in Kandy. This monk is supposed to have been a student of the Pāla-era ācārya Ratnākaraśānti (slob dpon chen po shānti pa, fl. ca. 945–1040s).[46] Tāranātha, anticipating criticism, hastens to say here that Yaśākaraśānti must have been invisible to “ordinary people.” But what is almost as incredible as the claim of an outlandish lifespan is the assumed knowledge of Ratnākaraśānti’s visit to Sri Lanka. This obscure factoid is reported only in a couple of hagiographies in the Tibetan corpus; it would be surprising if it had been known to Buddhaguptanātha.[47] And as Decleer points out, the name slob dpon chen po shānti pa is liable to be confused with the name of Buddhaguptanātha’s master Śāntigupta.[48] Śāntigupta is said to have had students elsewhere in the Theravāda domain, and two of paṇḍita status, Dharmākṣaghoṣa and Hitānandaghoṣa, are named in the Tenasserim segment of the itinerary.[49] Here an anecdote about a student of Śāntigupta may have been misinterpreted by Tāranātha, perhaps deliberately and for sensational effect. Although the tale of Yaśākaraśanti has no direct bearing on the reconstruction of the itinerary, it represents a clear case of unreliable narration.
5. Ulliṅga (Pulau Lingga)
After his Sri Lanka sojourn, Buddhaguptanātha sailed for over a month and arrived at Ulliṅga, an island that has not previously been identified. The Tibetan transcription u lingga can be read as Sanskrit Ulliṅga, “Up-phallus,” without needing further correction.[50] This place is classified as “a solitary island” (gling phran), an island proper. Ulliṅga is said to take about a month on foot to traverse. It was also a port where merchants set off for the next island, a shipping center another month’s sail away. The direction of travel from Sri Lanka to Ulliṅga is not stated. Sailing to the west and north of Sri Lanka for a month, for instance, would bring a seafarer close to the Strait of Hormuz. However, no island called Ulliṅga, or a related name, is located in the vicinity, and this area does not line up well with the next segments of Buddhaguptanātha’s journey.
Alternatively, a month’s eastward ocean journey from Sri Lanka via the Malacca Strait could reach Lingga, located in the Riau islands of present-day Indonesia. Pulau Lingga sits at a major maritime crossroads, but it had not been a center of trade up to the late sixteenth century, and this is consistent with Buddhaguptanātha’s very brief account of the place. The Riau region had, however, seen rapid growth in maritime traffic since the start of the colonial era. The Johor Sultanate transferred its seat to Lingga shortly after Buddhaguptanātha’s visit. The main port of Lingga, Daik, is within view of Daik Mountain, which gives Lingga its name, and which may have been a memorable sight to a devout traveler from India. The spelling recorded in the Hagiography can be accounted for as a contraction of pu la u lingga or of an eggcorn form such as pu ra u lingga.
6. Amuka (Macau)
Having left Ulliṅga, Buddhaguptanātha reached the next island, Amuka, after another month’s sail in an unspecified direction.[51] Again, this island has not been identified in previous studies. The word amuka means “so-and-so” or “insert name here” in Sanskrit, but in the context of a travelogue, amuka would not mean a place with a name that has to be supplied for some reason. It is more likely to be a phonetic rendering of a native name.
At Amuka, Buddhaguptanātha reports that he was the only Indian travelling at this point in the itinerary. This observation does not clearly determine whether he was travelling far to the east or to the west of Sri Lanka. If he was in the vicinity of the Arabian Sea, he would have had ample opportunity to encounter merchants from Gujarat, Goa, Malabar, and other southwestern Indian mercantile regions that did business there. Further hints as to Amuka’s location are offered by the description of the next leg of the journey:
From there, together with the people of [these] isles [and on] a huge ship of superior craftsmanship with superb features, [Buddhaguptanātha] headed southward together with some five thousand people, arriving after about four months.[52]
The figure of five thousand persons refers to a combined fleet, as long as “we are not confronted with an exaggeration of the narrator or of the writer or with a clerical error,” as Tucci remarks.[53] Amuka was then a place where large oceangoing boats could be built or assembled. This place is also said to be located some four months’ journey by sea north of the next destination, which has been identified by other scholars as the Swahili coast.
If Ulliṅga is Lingga, then Amuka’s location is to be found in a shipbuilding region in or nearby Southeast Asia.[54] Manila is one place that lies within a month’s sail from Lingga and about four months’ sail from East Africa. A trans-Pacific shipbuilding industry involving indigenous and Chinese artisans was instituted at Manila by Spain’s rulers in the early 1570s.[55] Large vessels built there were, however, used for the Manila–Acapulco galleon trade, not on the Indian Ocean route that Buddhaguptanātha seems to be describing. Ayutthaya is another maritime center of the period that potentially fits Amuka’s position in the itinerary, but Ayutthaya was under siege for much of the 1580s, and if Buddhaguptanātha had been there, he ought to have said something about its Buddhist community, as he later does on his visits to nearby Tenasserim.
In the late sixteenth century, Macau was known to the Portuguese by various names similar to Amuka: Amaqua, Amaco, Amakau and so on. These toponyms are cognates of the name of the oldest seafaring shrine on Macau, the A-Ma Temple, which is called aa maa gok 阿媽閣 in Cantonese.[56] It is generally accepted that this temple was established in 1488, and that it was synonymous with Macau for the Portuguese, who established themselves on the island in the mid-sixteenth century. A slight mishearing or a graphic confusion between a ma gu and a mu ka could account for the minor difference in spelling. Tāranātha might also have hypercorrected his notes here, because the word amuka, “such-and-such,” occurs several times in Sanskrit mantra formulas in his collected works.
When other factors are considered, it seems likely that Macau was the place visited by Buddhaguptanātha. The Ming Ban on Chinese maritime trade had been lifted in 1560, leading to a surge of traffic in the South China Sea. The 1575 Portuguese embassy to China reported witnessing “unbelievable” (yncreyble) shipbuilding activity in Fujian. One of its members claimed that the emperor had offered the Portuguese as many as five hundred ships to combat piracy in the seas around the Philippines.[57] In this context, Buddhaguptanātha’s claim of a fleet carrying five thousand people gathering for a southward journey becomes much more credible. Even so, the figure of five thousand persons is very large, and ought to be corroborated by shipping records for the period. This figure should otherwise be understood as referring to some other quantity, such as the full population of the township at the time.
A church of São Lourenço had been standing in Macau since 1575. This is then one place where Buddhaguptanātha could have had an opportunity to encounter the Catholic mass and to associate it with the name São Lourenço. And Chinese crews would, arguably, have been more open to taking a tantric Buddhist holy man on board than Arab, Malay, or Siamese seafarers of the period. Even so, it is not clear whether the whole of the fleet mentioned in the Hagiography would have been owned and operated by Chinese traders, or whether, as the aforementioned Portuguese reports might suggest, some of the vessels were being transferred to other owners on their outbound journey.
7. Jhamigiri (Zanzibar Archipelago)
After about four months’ sail south from Amuka, Buddhaguptanātha arrived at a place he calls Jhamigiri. Tucci’s argument that the name Jhamigiri refers “unmistakably to the country of the Zanj, that is, Zanzibar” has been widely accepted.[58] Although Tucci later changed his mind and identified Jhamigiri with Jambi in Sumatra, Sumatra does not match key details of Buddhaguptanātha’s account, especially the island’s small size and remote location.[59] The syllables dzha mi appear to convey a word cognate with the Persian word zangi, which underlies toponyms for Eastern Africa in many languages: Malay janggi, Javanese jenggi, Cantonese zang kei and Hokkien chng kî 僧祇, among others. Tucci added that the giri element of the name “generally means a hilly country,” and indeed the Hagiography refers to specific mountain features of Jhamigiri. A combination of janggi and giri would be an allowable word formation in Malay and Javanese as well as Sanskrit, although the toponym *Janggigiri is not known to be attested in any language. As Buddhaguptanātha later leaves this area for Indonesia, he may have adopted Malayo-Javanese terminology at this point in the journey.
Jhamigiri’s features can be correlated with various places in the Swahili coastal islands. Buddhaguptanātha’s report of three hill ranges running north to south matches the topography of Zanzibar proper, that is, of Unguja island. These hill ranges are said by Buddhaguptanātha to separate two nondescript towns located seven days’ journey apart.[60] At least two substantial settlements had existed on Zanzibar in the sixteenth century: Unguja Ukuu in the southwest and Fukuchani in the north.[61] The two towns lie about thirty-five miles from each other in a straight line. Buddhaguptanātha managed to stay there for a year, practicing rejuvenative therapy with musk (kla pa), which could have been obtained from the introduced civets (Viverricula indica) on Zanzibar.[62]
The Hagiography adds, “At the summit of one of the hills the golden glow of a small rock formation shines even at night.”[63] As Tucci realized, the glowing rocks on the mountaintop were in the caldera of a volcano.[64] Although Unguja island itself has no volcanoes, there is an active volcano at Mount Karthala on Grande Comore in the Comoro Islands, located about two hundred nautical miles away. The only other persistently active volcano in the western Indian Ocean is the remote Piton de la Fournaise of Réunion, located more than eight hundred nautical miles away. The landscape of Grande Comore is dominated by the volcano cone. The Hagiography, however, describes the glowing mountaintop in the same passage as the three hill ranges and two towns of Jhamigiri. Tāranātha’s text does not support the interpretation that the volcanic mountain is found on a different island. If this segment of the itinerary takes place in the Swahili coast region, the Hagiography must have conflated two distinct localities in the same area.
Buddhaguptanātha goes on to observe that Jhamigiri’s inhabitants live by foraging: “As there are no fields or the like, there are tree trunks, five times human size, [which allow] for taking a single fruit-seed; a single household’s subsistence is always dependent just [on that].”[65] These trees were presumably coconut palms. Although coconuts were a novel sight for the Portuguese who first noticed them on Zanzibar in 1498, a South Indian such as Buddhaguptanātha should not have found them to be remarkable in themselves.[66] He seems to be emphasizing the centrality of coconut products in the local diet—a topic of interest to other Age of Sail travelers—if not also the height of the locally established palm tree species.[67]
The Hagiography also asserts that stone images and temples of the “Victor,” the Buddha, were worshipped in regular festivals at Jhamigiri. No trace of premodern Buddhism has, at present, been found in the Swahili Coast area. The report of an image worship festival taking place in an Indian Ocean island is not in itself implausible, however. It could be true for other candidate localities in this maritime region. Nonetheless, this detail keeps us from confidently identifying Jhamigiri with Zanzibar, or in any case, casts doubt on the reporting of this part of the itinerary. Buddhaguptanātha goes on to discuss a site called Nāgārjuna’s Throne (klu sgrub phebs pa’i bzhugs khri), which purportedly goes back to Nāgārjuna’s lifetime in about the third century CE. This site and its associated legend is described in more detail in Tāranātha’s Seven Instruction Lineages.[68] There it is described as a boulder-like temple in which a stone image seemingly seated on a throne can be viewed through a crevice. A rock formation corresponding to this description has not yet been identified with a particular site.
8. Poṭala Minor (an Indian Ocean island)
At the end of his year at Jhamigiri, Buddhaguptanātha sailed east with some merchants heading towards “some small islands” (gling phran ’gar). After a month he reached the renowned Poṭala. In Buddhist lore the mountain called Potala or Poṭalaka is known as the abode of the savior Avalokiteśvara. The Poṭala of the earlier Buddhist scriptural tradition has been identified as the Pothigai Hills in Tamil Nadu, which have long been regarded as sacred in local Hinduism and Buddhism.[69] In the later scriptural tradition of the Gaṇḍavyūha, and in tantric Buddhism, Poṭala came to be known as “a mountain rising above the sea.”[70] The location of this maritime, insular Poṭala has not yet been determined. Tāranātha calls it “Poṭala Minor” (po ṭa la chung ba), a quasi-scriptural term, in order to clarify that it is located in the mundane world (mi’i ’gro ba’i spyod yul gyi) instead of the realm of the visionary.[71] Here Tāranātha is referring to visionary journeys to Poṭala of the kind described in his other works and by his Tibetan peers.[72]
Buddhaguptanātha locates Poṭala a month’s sail east of Jhamigiri and in a place where Java lies “a great distance to the north.” There is a problem in the itinerary here. If Poṭala was an island lying far to the south of Java, it would be in an extremely remote location in the Antarctic Ocean or off the Western Australian coast. If the “great distance” from Java is confined to about four hundred nautical miles, the Cocos Islands or Christmas Island would be in range. But neither was known to have been populated before the seventeenth century, whereas Poṭala is reported to have a small population.[73] Up to the late colonial era, Christmas Island was avoided by seafarers—even those from Java—because of its treacherous coast and lack of human habitation.[74] Furthermore, none of the islands south of Java can be reached in a month’s sail from Africa, nor do they have any of the features ascribed to Poṭala Minor. There is no prospect of finding Poṭala’s mountain, for instance, on low-lying Indian Ocean atolls such as the Chagos Archipelago or the Cocos Islands. But if the direction of travel is changed from north to east, or if we understand a direction more easterly than northerly, then more promising candidates for identification with Poṭala Minor are opened up.
Tucci at first identified the Poṭala of the Hagiography as Madagascar, but later changed his mind.[75] Here, again, Madagascar does not fit many aspects of Buddhaguptanātha’s description of Poṭala: its size (it can be circumambulated), population (sparse, non-urban) and location (due east of the Zanzibar Archipelago). When Tucci later identified Jhamigiri with Jambi on Sumatra, he then suggested that Poṭala might be Borneo, as it lies nearby and to the east of Jambi.[76] Again, however, there are too many discrepancies. Borneo is a major landform with a large population, not one of a group of “small islands,” and it is much closer to Jambi than a month’s sail.
David Templeman has instead proposed that Mount Poṭala is to be found among the Comoro Islands.[77] This fits better with the reconstructed itinerary; the Comoros do lie east—i.e., southeast—of the Zanzibar Archipelago. Grand Comore has the highest peak in the Western Indian ocean, Mount Karthala (al Qirṭālah). This name might lend itself to assimilation with Poṭala. Nonetheless, Buddhaguptanātha’s separate account of Jhamigiri already seems to refer to Mount Karthala and its volcano.
A ship sailing due east from Zanzibar for a month would be in the vicinity of the Seychelles Islands. On the largest island, Mahé, there is the Morne Seychellois, a mountain 905 meters high. Just as Buddhaguptanātha is said to have circumambulated Poṭala mountain, today it is possible to walk around the foothills of Morne Seychellois. The Hagiography states that Poṭala island has a number of natural beauty spots or “self-created shrines” (ri bo rang byung gi lha khang) to various deities. The shrines to deities named here—Maṇibhadra, Bhṛkuṭī, the Asuras’ cave, the Acacia Forest Tārā, Mañjuśrī, Brahmā, Viṣṇu, Śiva—are stereotypically associated with Poṭala in tantric Buddhist and Tibetan texts, but there are no indications that such texts were known to Buddhaguptanātha. In particular, the itinerary’s “famous spot of Ārya-Mañjuśrī” (’phags pa ’jam dpal gyi gnas su grags pa) is elsewhere associated with Poṭala only in an obscure pilgrimage handbook that survives in Tibetan translation, the Potalai Lamyik (po ta la’i lam yig).[78] Tāranātha then appears to have added this detail, if not others, into the itinerary. According to the Hagiography, Mañjuśrī’s “famous spot” is situated in “jungle at the face of the sky, and water flows straight down from there.”[79] There are a number of mountain waterfalls on Mahé; waterfalls are also found on other Indian Ocean islands. Buddhaguptanātha further describes the mountain as having a “summit with a crystal peak, inconceivably tall and wide.”[80] As no mountains are known to have a large crystal on their summit, this detail ought to refer to a pyramidal peak shaped like a scepter quartz crystal. Again, this feature could describe Morne Seychellois, as well as other Indian Ocean island mountains.
Buddhaguptanātha also discusses the sparse population of Poṭala island. The Seychelles are not known to have been populated at the time of his travels. However, the Seychelles’ apparent lack of habitation in the sixteenth century is regarded by archaeohistorians as an anomaly, since most other islands along the old sea route between East Africa and South and Southeast Asia have preserved traces of human settlement going back to the first millennium.[81] According to the Hagiography, on Poṭala:
All the people of that island at present have no religion of any kind—insider, outsider [Buddhist or Hindu], Mohammedan and so on. There are no villages, except for a few. […] At the threshold of a house, there is a profusion of magically effective devices, not found on the Indian subcontinent.[82]
Previous studies concur that these “devices” or “designs” (’khrul ’khor) are yantras, or protective magical diagrams, as fits the context. The Tibetan word tha (mtha’), here translated as “threshold” (*paryanta, *āntika), means a perimeter or entranceway. Here Buddhaguptanātha may be referring to Swahili doors, some of which are traditionally carved with talismanic designs.[83] As protective designs on the exteriors of houses are found in many island cultures—Sumatra and Borneo, mentioned earlier, are examples—this feature does not fix Poṭala’s location at one place in the Western Indian Ocean islands. If Swahili doors are being described, this detail of the narrative would fit better in the account of Jhamigiri/Zanzibar and its towns. But the observation that the inhabitants had no organized religion would not fit Zanzibar, which was within the Arabic cultural sphere by the sixteenth century.
One of Buddhaguptanātha’s anecdotes about Poṭala has an intriguing and previously unnoticed parallel in the Age of Discovery literature. As Tāranātha reports:
On this island, he saw many people from elsewhere [who] were able to cover the whole body with [their] ear[s].[84]
On Magellan’s first voyage around the world, a pilot from Maluku is recorded by Antonio Pigafetta (c. 1491–1531) as having made a very similar statement:
There was an island […] the men and women of which are not taller than one cubit, but who have ears as long as themselves: with one of them they make their bed and with the other they cover themselves.[85]
These hominids were said to live somewhere in the Alor archipelago of present-day Indonesia, which is thousands of miles away from Zanzibar, and on the opposite side of the Indian Ocean.[86] The pilot’s story of island hominids with body-length ears is nonetheless so similar to Buddhaguptanātha’s that it indicates a common origin. Pigafetta’s account of Magellan’s voyage became widely known in Europe after it was published in the 1550s, but Buddhaguptanātha does not report interacting with the Europeans who are mentioned in passing in his travelogue. It is more likely that this story remained current among Southeast Asian seafarers, with whom Buddhaguptanātha may have been travelling on the journey from Poṭala to Java. One of Magellan’s Malay crew members, Enrique of Malacca alias Panglima Awang, left the expedition in Cebu in 1521.[87] But he left too early to have heard the Maluku pilot’s story, which Pigafetta recorded in his journal in 1522. The sixteenth-century lore of the hominids with body-length ears then seems to have circulated mainly in the Maluku Islands area if not also in neighboring parts of Indonesia and the Philippines.[88] The transmission of this lore to Buddhaguptanātha is another indication that some of his fellow travelers came from this region.
Considering that the features of Poṭala fit Southeast Asian islands better than the mostly uninhabited Western Indian Ocean islands, and that the next destination seems to encompass Sumatra as well as Java, Buddhaguptanātha may have been describing a place in the island chain parallel to Sumatra’s west coast, perhaps in the Mentawai islands or the Nias archipelago. Here we find sparse animist communities, longhouses carved with protective designs, mountains, waterfalls, and a greater possibility of Buddhist vestiges—if that aspect of the itinerary is regarded as credible. But in order for Poṭala to be identified with one of these Sumatran islands, the travel directions have to be emended even further. Instead of sailing from Zanzibar east for a month and north for “a long time,” the traveler would need to sail east for “a long time” and north for about a week. Although it is quite possible that Tāranātha’s text has confused the travel directions in this way, given that similar problems have been diagnosed elsewhere in the Hagiography, we have little choice but to follow the narrative as it stands until more information about Poṭala’s location comes to light.
9. Java/Sumatra
Following his stay on Poṭala island, Buddhaguptanātha sailed to the “island of Java,” which is named as such (dza ba dwī pa). Most of this segment of the itinerary, however, seems to describe neighboring Sumatra better than Java. The possibility that notes on Java proper were redacted into the wrong point in the narrative has been discussed in connection with São Lourenço. Sumatra was known by various names in the period, including—among Europeans—“Java Minor.” Buddhaguptanātha later mentions visiting the two “Golden Isles” (suvarṇadvīpa) of Southeast Asia, which usually include Sumatra.[89] The description of Javadvīpa therefore needs closer scrutiny:
From that [Poṭala] island the ship travelled a great distance northwards; upon arrival, he went to so-called Javadvīpa, Barley Island. On that island were many congregations of the Saindhava monastic order, with whom he did not stay. In the middle of a small lake there is a small island named Vanadvīpa, Forest Island. A place blessed by the ācārya Saroruhavajra there has a stone exterior and a square temple interior. A naturally formed stone statue of the two-armed Hevajra sits in the middle. In one cave there are many books on mantra. Even the [Hevajra] Tantra of Five Hundred Thousand [Verses] is said to be preserved [there], but [the area] is notoriously very volatile; there is no way to see [the books].[90]
As has been discussed in connection with the Poṭala island segment, Buddhaguptanātha’s crew was instead most likely sailing “for a great distance” east, not north to this locale. If so, they would have arrived at the Sunda Strait or the Western Sumatran coast. Landfall in Sibolga Bay would have put him within close reach of the Padang Lawas area that seems to be described here.
A community of monks practicing conservative Buddhism, which Tāranātha calls Saindhavas (sendha pa), is not known to have been in existence in Indonesia in the sixteenth century or for a long time beforehand. The epithet Saindhava, “of Sindh,” is now accepted to have designated the Saṃmatīya order of monks, whose last members disappeared in the thirteenth century together with any non-literate memory of them. Tāranātha then appears to have imposed his own, anachronistic nomenclature—probably drawing on earlier Tibetan historical writing—onto Buddhaguptanātha’s stories of meeting monks who would now be identified as Theravādins.[91] While Tāranātha’s view that conservative or “Saindhava” Buddhist monks resided on Java in the sixteenth century is expressed in some of his other works, it has no known factual foundation.[92] Such conservative monks, as described in Tāranātha’s second-hand anecdotes, are more likely to have been Siamese or Cambodian, but no suggestions are offered here as to how they might have been misidentified as Javanese.
There is, however, a chance that the reference to the Hevajratantra of Five Hundred Thousand Verses (Tib. rgyud ’bum phrag lnga pa, Skt. Pañcalakṣatantra) comes from Buddhaguptanātha himself. His guru Śāntigupta wrote a brief commentary on the Hevajrapiṇḍārthaṭīka, which in turn cites a Pañcalakṣatantra or Hevajratantra of Five Hundred Thousand Verses as a prime authority,[93] although Śāntigupta’s commentary does not itself refer to the Pañcalakṣatantra.[94] The claim that this never-seen text was stashed in a cave in the Sumatran highlands is simply unbelievable, although it is possible that Buddhaguptanātha had heard that manuscripts or artefacts belonging to Sumatra’s last Buddhists were hidden away in such a place.
The only known statue of the two-armed Hevajra found in Indonesia was once enshrined in the Hindu-Buddhist temple complex stretching across Padang Lawas in North Sumatra.[95] This area is located less than a hundred kilometers from Lake Toba and Samosir Island, which is one of the world’s largest lake islands, and is probably the place designated by Buddhaguptanātha’s name Vanadvīpa. The “Forest Island” epithet, in this context, refers not to the island’s sparse tree cover but to its tall, dense forest surroundings. Samosir was described in such terms by a Indian visitor later in the colonial era.[96] No Hindu-Buddhist temples have been found on Samosir, however. As many stone funerary monuments are spread around Samosir island, it would be seem to have been a fitting place for a Hevajra/Heruka temple, but it would be mere speculation to venture that an image of the deity was emplaced there. To date, the only two-armed Hevajra image found in the whole of Indonesia is the unique statue from Padang Lawas. The description of a “square temple interior” apparently refers to a cave or cliff-face temple, but if it is understood (with a slight change to the wording) as a temple with stone or brick construction and a square inner sanctum, it would designate premodern temple structures on Padang Lawas such as the Biaro Bahals or Biaro Si Pamutung. On the available evidence, it would again seem that recollections of two different but geographically close sites have been combined in Tāranātha’s text. The site’s stated association with Saroruhavajra (mtsho skyes rdo rje), likewise, remains uncorroborated. If this association is not a mere figment of Tāranātha’s imagination, it might refer to some aspect of the iconographic program of Saroruhavajra’s seminal Hevajrasādhana.[97] This sādhana has been preserved in several Sanskrit witnesses and might therefore have been seen at some point by Buddhaguptanātha.
10. Destinations in the Bay of Bengal region
In the final maritime segments of his itinerary, Buddhaguptanātha travelled to various places clustered around the Bay of Bengal’s coastline. As the destinations in this part of his journey are closer to the Buddhist heartland and more recognizable, they will be covered more briefly here. After leaving Java, so called, Buddhaguptanātha returned by ship to Sri Lanka and sailed on to the Konkan coast and Malabar. There Buddhaguptanātha met with South Indian Vajrayāna practitioners who introduced him to his guru Śāntigupta. The profusion of names, places, and traditions given by Tāranātha leaves the impression that a genuine religious milieu is being described, but again, too little of this has been corroborated to discuss further here. The yogi then travelled on foot—as Tucci understands the itinerary—to various well-known places: Jharkhand (dzā ri khaṇḍa), Jagannatha at Puri (dza gad nā tha), Bengal (bhaṅga la), Khasarpaṇa at Puṇḍravardhana (puṇḍa bharta ru kha sarpa ṇi), Tripura (ti pu ra) and Devikoṭa, which Decleer identifies with the Kamakhya shrine of Assam.[98]
Buddhaguptanātha then crossed into mainland Southeast Asia. He sauntered through Rakhine (ra khang), Hariphunchai (ha ri bhanydza), Bagan (pu khang) and Bago (pal gu), all located in the vicinity of present-day Myanmar.[99] Here he is said to have met two disciples of his own guru, one “at the monastery of the Great Stupa (*mahācetiyavihāra) of Haribhuñja,” probably Wat Phra That Hariphunchai in Lamphun, northwest Thailand, which has such a stupa.[100] They are also said to have taught Buddhaguptanātha mantra techniques (gsang sngags), which would mean paritta and other apotropaic magic practiced in the Theravāda milieu, if these figures had not also been identified as students of Śāntigupta. Contacts with tantric practitioners there are known outside Tāranātha’s world; a monk from Hariphunchai is mentioned in connection with a short hymn of guru devotion translated by Shalu lotsāwa (zhwa lu lo tsā ba, 1441–1528).[101]
11. Dhanaśrī (Tenasserim) / Dhānyakaṭaka (Amaravati)
Travelling by boat again, Buddhaguptanātha went to a dvīpa area called Dhanaśrī and to the famous Dhānyakaṭaka stupa. Tāranātha gives the impression that they are two names for the same place, whereas each name in fact designates a different place located on opposite sides of the Bay of Bengal.[102] Dhanaśrī, in the understanding of Tucci and others, is the Tenasserim Division of Myanmar.[103] It is now called Tanintharyi, and it incorporates the northern part of the mountain range called Tanah Seri in Malay. Similar names for the region were used in the sixteenth century.[104] As “a great many monastic communities were on this dvīpa” (gling ’di nang dge ’dun shin tu mang zhing), Buddhaguptanātha was evidently in mainland Southeast Asia. However, the narrative turns abruptly to the stupa called Śrīdhānyakaṭaka (dpal ldan ’bras spung), which is located in present-day Amaravati, Andhra Pradesh. The itinerary is either missing a statement to the effect that Buddhaguptanātha sailed west from Dhanaśrī to Andhra, or, more likely, simply confuses Dhanaśrī with the name Śrīdhānya, as he does in his History of Buddhism in India.[105] Since the next segment in the itinerary returns to Bago, it seems that Buddhaguptanātha had remained on the Myanmar coast. Tāranātha then seems to have redacted a note about Śrīdhānya into the wrong part of the itinerary.
The Dhānyakaṭaka stūpa was known to Tibetans who were versed in the Kālacakra tantric system, including Tāranātha. Buddhaguptanātha accurately describes the stupa’s distinctive double-rail architecture, which was documented later in the colonial era. The claim that the stupa takes a whole day to circumambulate must, however, be another exaggeration. Buddhaguptanātha adds:
there is a large town to the east with a great bazaar teeming with merchants from various countries such as China, Europe and India.[106]
If Amaravati is the point of reference, the town referenced here could be Machilipatnam, which is located on the coast downstream from Amaravati. As Dutch and English traders had not yet been established in India at the time, Buddhaguptanātha would again be referring to the Portuguese. Nonetheless, the Portuguese rarely visited India’s east coast during the period in question. Alternatively, if Buddhaguptanātha was still in Tenasserim at this point, he could have been referring to Dawei (formerly Tavoy), the present-day capital of the region, or to Myeik (formerly Mergui), a port located further to the south and east, which Europeans had been visiting since the early colonial era. In any case, the account of Dhānyakaṭaka would be better placed in the itinerary after Buddhaguptanātha’s final ocean journey, which took him to Triliṅga, that is, coastal Andhra.
12. Bago and minor islands in the northern Bay of Bengal
Travelling by boat from Tenasserim (if not from Śrīdhānya), Buddhaguptanātha returned to the coast of Bago. Here he journeyed across the “so to speak ocean of the Poṭala.”[107] This is probably Tāranātha’s term, as Poṭala island itself is not mentioned in this segment, although another unnamed and undescribed island outcrop is. Buddhaguptanātha is evidently still in the north of the Bay of Bengal region. Bago’s location is consistent with the report of “nutmeg, cloves and so on growing” there. According to the Hagiography, the area was “held by the Franks” (phreng gis ’dzin pa’i gling zhig dang), i.e. the Portuguese. Bago only came under overt Portuguese control at the end of the sixteenth century, so it is not clear whether this remark applies to Bago itself or to somewhere nearby.[108] Tāranātha may have added this remark while he was writing up his notes; since his encounter with Buddhaguptanātha, he had had at least six more meetings with Indian travelers, some of whom were the yogi’s students and became Tāranātha’s co-translators.[109]
Buddhaguptanātha is then said to have visited various dvīpa-type locations in the vicinity, about which no further information is given: “Sādhadvīpa; both the greater Golden Isles and the so-called lesser Golden Isles; Sūryadvīpa; Candradvīpa; Sarvadvīpa.”[110] The two “Golden Isles” areas are presumably the Malay Peninsula and Sumatra. The toponyms *Sādhadvīpa and *Sarvadvīpa have not been identified.[111] Sūryadvīpa was used as a name for the Bengal Delta in one tantric context.[112] Candradvīpa is the classical name for the southern coastal fringe of the Delta and especially of the Barisal Division of present-day Bangladesh. The next segment of the itinerary takes place nearby.
13. Gaṅgāsāgara (Sagar Island)
At the end of his travels through the Bengal Delta region, Buddhaguptanātha visited “the small island of Gaṅgāsāgara,” present-day Sagar Island. This is a place of some importance in India’s sacred geography, being located at the confluence of the ocean and the Ganges River. It is the venue of an annual bathing festival devoted to the goddess Gaṅgā.[113] The itinerary nonetheless makes no reference to the ingrained Hindu and Pauranic character of this spot. Gaṅgāsāgara is instead described as a site sacred to tantric Buddhists, where one
encounters the marvel of the ḍākinī-donated sixfold bone ornaments of ācārya Kṛṣṇacārya blazing and so on at night, and an extraordinary tumult.[114]
Tāranātha expands on the association with Kṛṣṇacārya in his dedicated namthar of this famous master. He recounts a long story about a historic icon on Sagar Island that was crafted in the form of Śiva to disguise its true identity as Avalokiteśvara, and which, having been consecrated by Kṛṣṇacārya, was later moved to the island for safekeeping.[115] It is conceivable that a story with these elements circulated among Nātha yogis as late as the sixteenth century. Tāranātha claims that Kṛṣṇacārya’s (Buddhist) yogic tradition was still active during his lifetime, and had support for these claims. He translated previously untranslated manuscripts of works attributed to this master and recorded the oral tradition putatively associated with them in the Life of Kṛṣṇacārya.[116] As for the story about the icon, a Śiva temple is reported to have stood on Sagar Island until the early nineteenth century, but these reports are unclear.[117] The image at the ancient Kapil Muni Ashram on the island has been described as a “shapeless block of stone.”[118] The site’s identity and position in the itinerary is, in any case, not in serious doubt.
14. Unnamed small islands in the Bay of Bengal
On his last recorded maritime journey, Buddhaguptanātha boarded a boat at Sagar Island heading for Jambūdvīpa—that is, toward the peninsular and southern part of India. The boat is said to have stopped en route at an unnamed group of small islands. There are, however, no islands located in a direct line between Sagar Island and India’s southeastern coast; the ship may have gone off course. A ship bearing due south from Sagar Island would head toward the Andaman Islands with an easterly wind or current.[119] Buddhaguptanātha describes his location:
On a deserted island, there is a rose quartz cave and also a sandalwood forest; being very wild, not even a single leaf could be removed.[120]
In the sixteenth century, sandalwood proper (Santalum album) had not yet spread widely beyond its place of origin, East Nusa Tenggara, or its few naturalized habitats in South India. Buddhaguptanātha may have been referring to the wild populations of sandalwood that have since been discovered in the Andamans.[121]
This Bay of Bengal island was also “renowned as the place of Āryāvalokiteśvara Siṃhanāda’s rousing of the thought of awakening.”[122] The legend being referred to here has not yet been identified. Since the canonical narratives of Siṃhanāda’s quest for enlightenment are not known to have survived in written form in medieval South Asia, this oddly specific identification is again most likely Tāranātha’s. Here Tāranātha may have wanted to flesh out vague reports of a Buddhist tradition surviving in the region. Legends of a visit by the Buddha to the Andaman and Nicobar Islands still reached Chinese seafarers in the fifteenth century, although they no longer mentioned the temple that was previously said to have stood there.[123]
On this unnamed island, finally, Buddhaguptanātha had a fortuitous meeting with the yoginī Dinakarā, a fellow student of his guru Śāntigupta. He then returned to eastern India at Triliṅga (present-day coastal Andhra), dedicated himself to following Śāntigupta, and sauntered around many more lands, but did not travel across the ocean again.
Figure 1. A tentative reconstruction of Buddhaguptanātha’s Indian Ocean itinerary, starting from the Konkan coast, c. 1570–1585. https://doi.org/10.6084/m9.figshare.26419987.
Conclusions
The foregoing analysis of the places named and described in Buddhaguptanātha’s Indian Ocean journeys can now guide a reconstruction of his itinerary. As many toponyms are still only provisionally identified, and various flaws in the text have surfaced without being fully resolved, the present reconstruction (Figure 1) is put forward as an initial hypothesis.
At the start of his maritime travels, Buddhaguptanātha’s position is uncertain, beginning with his port of embarkation in the Konkan, which may have been Goa, the only place in the area identifiable with São Lourenço. However, no place called São Lourenço in the Luso-Asian world can be connected to the misplaced and possibly partly fantasized account of a thriving Vajrayāna community. His next destination, Śaṅkhadvīpa, provisionally identified as Socotra, was a place of only minor importance for Indian seafarers in the sixteenth century. The crew of Buddhaguptanātha’s boat may have planned to rendezvous with other traders at Socotra; alternatively, they may have arrived by accident on their way to the port towns of the Omani coast. In any case, Buddhaguptanātha reversed course at Śaṅkhadvīpa. After arriving in Sri Lanka via the Maldives, his itinerary becomes more firmly anchored in place and time to a five-year stay occurring between the years 1573 and 1587.
The next segments of his journey—conjectured to span Colombo, the Malacca Strait, Lingga, and Macau—conform to a known Portuguese maritime route linking the Indian Ocean with the South China Sea. However, Buddhaguptanātha at no stage admits to travelling on a vessel operated by the Portuguese. At Macau he seems to have embarked on a ship built if not crewed by Chinese seafarers. His four-month outward journey from Macau to the Zanzibar archipelago and the Comores would have taken advantage of the easterly trade winds, passage through the Sunda Strait, and the South Equatorial Current of the Indian Ocean. The return journey to Java via Poṭala—tentatively identified with a location in the Seychelles or the West Sumatran Islands—would, in turn, have been assisted by the South Equatorial Counter Current. This route could have ended up anywhere from the northwestern tip of Sumatra to the Sunda Strait end of Java. His port of disembarkation may have been in Sibolga Bay, if his journey is understood to have continued into Padang Lawas and the Lake Toba area. Buddhaguptanātha may also have kept travelling along the coasts of Sumatra and Java for some time.
Buddhaguptanātha’s subsequent travels around Southeast Asia and the Bay of Bengal took him to destinations that are easier to recognize. Nonetheless, the itinerary as a whole remains hard to square with a historical geography whenever fantasized elements intervene. The story of the Poṭalaka pygmies has been traced back to sailors from insular Southeast Asia and, in this context, to Buddhaguptanātha’s presumed encounters with them at some point. Certain other fantastic details—Nāgārjuna’s Throne, the rumor of the Hevajratantra of Five Hundred Thousand Verses in “Java,” the full day needed to circumambulate the Dhānyakaṭaka stupa, the glowing bone ornaments at Sagar Island—could have been supplied by either Buddhaguptanātha or Tāranātha. Ultimately, Tāranātha is responsible for repeating if not inventing a number of tall tales in the Hagiography: Ratnākaraśānti’s seven-centuries-old student in Sri Lanka, the stereotyped faith-scape of Poṭala Mountain, the “Saindhava” monks of Java, the legend of Siṃhanāda Avalokiteśvara in the Andaman and Nicobar islands, and so on.
The presence of occasional fabulized or inaccurate elements in medieval hagiographies of all kinds is not unusual, and does not warrant the lazy conclusion of rejecting everything in their narratives. Tāranātha’s Hagiography has been found to contain much period detail that becomes recognizable as such with due attention to its historical contexts, and which could only have been conveyed to him in the manner he claimed. Problems in the narrative that were not worked out in previous studies can now be accounted for as simple redactional mistakes or as genuine if garbled reports of otherwise obscure locales. Without Tāranātha’s keen ear and literary talents, we would not have been able to appreciate that Buddhaguptanātha was one of the most widely travelled Indians of the early colonial era.
1. Such doubts have been expressed in Tibetan as well as Western scholarship; for some references see Templeman 1997, 959.
2. Templeman 2008, 47–92.
3. On this language, see Roesler 2018, 358–362, for conversation fragments preserved by Tāranātha’s peer Taktsang repa Ngakwang gyatso (stag tshang ras pa ngag dbang rgya mtsho, 1574–1651) in South Asia, which Roesler calls Hindi or Hindavi (i.e. Hindustani).
4. Templeman 2008, 212; see also Templeman 1997, 960.
5. Templeman 2021, 56.
6. However, the Hagiography’s account of its own redaction is understood differently by Decleer 2006, 40; see also Templeman 2008, 277.
7. The Buddhagupta’i rnam thar was completed in 1602, according to Templeman 2008, 1, 151.
8. Decleer (2006, 14–15), for instance, notes that the direction of travel to Uḍḍiyāna (here clearly associated with the Swāt Valley) should be corrected to so that it is located east, not west, of the Afghanistan-Hindu Kush region. Templeman (2008, 137), however, takes the Hagiography at its word here.
9. Dpe 94: rasme shwa ra (read: ramme shwa ra) zhes bya ba dang nyi ba indra ling ga zhes bya bai’i grong khyer chen po. Cf. Phun 532 l. 3; Decleer 2006, 8; Templeman 2008, 250. The nearby “great town of *Indraliṅga” has not yet been identified. Rajendracholiswarar Temple in Ilaiyankudi, located a few miles from Rameswaram, has as its presiding deity a liṅga apocryphally installed by Indra.
10. Dpe 95: rgyal po rā ma rā dza. This king was first identified by Tucci 1931, 686. In general, the dates in the Hagiography refer to the reigns of particular kings.
11. This part of the itinerary is discussed by Tucci 1931, 688–690, and translated in full by Decleer 2006, 12–21, and Templeman 2008, 253–260.
12. Buddhaguptanātha is, of course, to be distinguished from the fourth-century Buddhagupta and the eighth-century Buddhagupta or Buddhaguhya whose works are recorded in the Lhan dar ma.
13. Dpe 119, cf. Decleer 2006, 38; Templeman 2008, 273–274. For further discussion see Templeman 1997, 959–960.
14. Many of the teaching topics listed by Tāranātha here can be correlated with texts that he translated. Some provisional identifications of these texts based on their front and back matter are given here (cf. Dpe 120, Phun 567 l. 1): gsang ’dus ye shes zhabs lugs/ mi bskyod pa lha bcu dgu/ (unidentified, treated as separate topics in previous studies), sems kyi phyag rgya’i dbang/ (Q4845), dpyid thig gi byin rlabs/ (unidentified), bka’ babs drug ldan gyi khrid chen mo/ (unidentified), dza lan dha ra’i do ha dang man ngag/ (Q4652), sems kyi phyag rgyas man ngag gzhung bshad/ (Q4845, if sems dpa’i "), yi ge bzhi pa’i man ngag/ (Q4846), rang byin gyis rlabs pa’i byin rlabs rnams dang gzhung gi bshad pa/ (Q5180), phyag rgya bzhi snying po/ (unidentified), kyai rdo rje’i rdo gsa rim ku sa la’i man ngag dang/ (Q4701), yan lag drug pa/ (Q4703), ye shes ’char ba/ (unidentified; Decleer 2006, 40, n. 63, takes this to be a generic topic), sgron me gcig pa/ (Q4700), dran pa gcig pa/ (Q4783), ma hā mā ya’i man ngag[/] (Q4784), phag mo keng rus la zhugs pa dza lan dha ri’i lugs/ (unidentified), phag mo rngon pa’i gdams ngag[/] (Q4684), phag mo sna tshogs mdog/ (Q4685), rnal (Dpe: rnam) ’byor ma dkar dmar sngo gsum/ (unidentified), ngang ba’i rdo rje bdag med ma/ (unidentified), rdo rje phag mo ca co sgrog ma/ (unidentified), mi g.yo ba’i rdzogs rim a ti shas mdzad pa/ (Q4896), dpal mchog ḍom bi pa’i de nyid bzhi pa/ (if de nyid bdun pa, Q4653), nag po chen po’i sgrub thabs gnyis (Tāranātha is credited with the translation of at least four Mahākāla sādhanas, Q4966–4969, and perhaps Q5170 and Q5174).
15. Both translations are based on the Phun tshogs gling blockprint. Decleer (2006, 2–3, n. 1) provides a preliminary list of corrections to the text of this blockprint. Templeman (2008, 237) also refers to an unidentified blockprint received in his personal correspondence.
16. For the Dergé printing, see Ui et al, 1934; for Qianlong, see Otani 1965.
17. A similar etymology is given in Sarvabuddhasamāyoga 1.7: ḍā ni nam mkhar ’gro ba zhes (D367; for the Sanskrit text, see Negī 2018, 143), which associates the syllable ḍā with the word ḍākinī. Buddhaguptanātha is said to have had access to the Sarvabuddhasamāyoga (Sangs rgyas mnyam sbyor); see Decleer 2006, 29–30; Templeman 2008, 265, 274. The supplied etymology may, alternatively, connect the words drāmiḍa and drāva, the latter meaning “flight.” See Turner 1966, 378–379.
18. Dpe 104: rgya mtsho la nub phyogs su gru btang nas byon pas/ ’gro lding gi gling ste/ rgya skad du/ ḍā mi ḍo dwī pa zhes par phebs/ ’di la kla klo dang yul mtha’i mi dang/ gling phran pa rnams kyi skad du/ sam lo ra na so zer gyin gda’. Cf. Phun 545 l. 3; Decleer 2006, 22; Templeman 2008, 260.
19. Here Drāviḍa country is discussed in the same context as Sri Lanka, and as a region in India’s South; Chimpa & Chattopadhyaya 1970, 207, 331.
20. Tucci 1931, 692.
21. Decleer 2006, 23–24 n.33.
22. Dpe 104: dge ’dun mang po yod kyang ’dul ba’i lag len dag po med de/ dge slong rnams gos nag gyon pa phal cher chang ’thung ba sogs yod par ’dug go. Cf. Phun 546 l. 1; Decleer 2006, 23; Templeman 2008, 260.
23. Tucci later fleshed out his thinking on this a little; see Tucci 1951, 187–188 n. 1.
24. Mallinson 2019, 11 n.85.
25. For the (long defunct) church’s location near the Fortaleza da Aguada, and its dates of construction, see Lopes 2020, 118–124.
26. On the number of shipwrecks recorded near Madagascar in the period, for instance, see Larson 2007, 350 n. 13.
27. Tucci 1951, 187–188 n.1.
28. The Portugese in seventeenth-century Goa claimed that settlers on Anosy had been massacred in 1548 (Oliver 1886, 7), but these claims have been doubted (Larson 2007, 351). There is a report of a Christian missionary’s death on the island in 1585 (Oliver ibid.), but again, no indications of a Christian establishment functioning there at the time.
29. A short treatise titled *Pañcasamaya (dam tshig lnga pa), attributed to Padma ’byung gnas (back-Sanskritized as Padmasambhava in previous studies), is located in the Hevajra section of the Tibetan canon (D1224/Q2353). It does not refer by name to the Hevajra tantric system and its contents are largely shared with the gaṇacakravidhi of Abhayākaragupta (D2491/Q3317).
30. The same claim is repeated (and associated with *Drāmiḍa) in the History of Buddhism in India; cf. Chimpa & Chattopadhyaya 1970, 332.
31. Dpe 105: bde mchog dang/ dgyes rdor gyi dbang gsan nas/ dgyes rdor gyi rgyud la bshad pa’ang zhib rgyus su gsan/ dgyes rdor ’di slob dpon padma ’byung gnas kyi lugs yin cing/ sbyir gling de na rgyud sde bzhi ka’i bka’ ma chad pa dang/ khyad par bla med ches dar dus ’khor ma gtogs rgya gar na yod pa phal cher de na yod/ phur bu dang khro bo bcu’i rgyud dang/ he ru ka’i rgyud mang po dang/ phyag na rdo rje dang mkha’ lding dang/ ma mo dang nag po chen po sogs dang/ rta mgrin bla med kyi rigs rgya gar na med pa mang po yang de na yod pa dang. Cf. Phun 545 l. 4; Decleer 2006, 23; Templeman 2008, 260.
32. Dpe 124: de nas bal yul mthil du phebs/ ’dzu ’dzu ’ga’ zhig gis zhabs tog (Dpe: rtog) dang ’bul ba cher byas pa/ de nyid du ’phongs pa thams cad la stsal. Cf. Phun 571–572; Decleer 2006, 43 (who interprets ’dzu ’dzu, Newar juju, as Newar gubhāju); Templeman 2008, 279.
33. On Cāmuṇḍā (a mātṛkā goddess), Mahākāla (previously misidentified as Bhairava), Hayagrīva, Garuḍa and so on at Candi Jago, see Reichle 2007, 95, 163, 199.
34. The Tibetan Buddhist canon, for instance, contains a dedicated section on the Blue-black-robed Vajrapāṇi (’phags pa lag na rdo rje gos sngon po can), D2200–2214. Tāranātha is credited with translating a related text: gGnod sbyin gyi sde dpon chen po gos sngon po can gyi sgrub thabs, Q5164.
35. Wheatley 1961, 178. In one Purāṇa, Śaṅkhadvīpa is said to be a stupendous 100,000 leagues in extent.
36. An Indic name cognate with Sanskrit *Dvīpasukhada is said to underly the Greek name Dioscorida in the Periplus; Schoff 1912, 133. Waddell (1893, 58) back-translates bde ’byung gi gnas as *Maṅgalasambhava; Tucci (followed by Templeman) proposes *Śambhusthāna; Decleer (2006, 24) back-translates *Sukhākara. None of these names has been identified with an Indian Ocean locale.
37. Al-Okaishi 2021, 4.
38. Dpe 105: gling de na/ ga dza (Phun: da) nā sa zhes bya ba’i gling nas ’ongs pa’i mi gdong mi la/ sna glang po che lta bur yod pa mang po gzigs. Cf. Phun 546 l. 4; Decleer 2006, 24; Templeman 2008, 261.
39. There is no need to wonder “whether these were people suffering from elephantiasis,” as Decleer (2006, 24, n. 34) does. Templeman (2010) recalls that “My own nickname while working in a Tibetan refugee orphanage in the late 1960’s was sna khug ring po, or ‘Long nose.’”
40. Barclay 1735, 468.
41. The name Felidhoo comprises words cognate with Sanskrit phāla, “hoe, shovel, blade,” i.e. “paddle” (the Dhivehi word fali is also pronounced fallā or fāli in some Maldivian dialects, according to Maniku 2000, 73), and dvīpa or Prakrit diva, “island.”
42. Contrary to, for instance, Decleer 2006, 24; Templeman 2010, 5–7.
43. Here Tāranātha gives a śloka in praise of Sri Lanka in Tibetan transcription; cf. Waddell 1893, 59.
44. Dpe 106: rā he shing kha bhan ṭa ri; Phun 547, l. 6: sā he shing kha bhan ṭa ri.
45. Templeman (2008, 147) uses a different point of reference that has Buddhaguptanātha meeting Śāntigupta in 1575 (itinerary no. 10), at least two years after leaving Sri Lanka.
46. Dpe 105: kan ḍa la zhes bya ba’i yul zheg gi phu/ nags ’thugs po zhig gi dbus na brag phug ’dug/ de na sngon slob dpon chen po shanti pas singga li’i gling du phebs pa’i dus kyi dngos slob/ ma (read: ya) shā ka ra shanti bod skad du grags pa’i ’byung gnas zhi ba zhes bya ba/ dgung lo bdun brgya yang lhag lon pa zhig bzhugs pa. Cf. Phun 547 l. 3; Decleer 2006, 25; Templeman 2008, 262.
47. For the story of Ratnākaraśānti’s visit to Sri Lanka, see the Tibetan translation of the Caturaśītisiddhapravṛtti of Abhayadatta, the Grub thob brgyad cu rtsa bzhi’i lo rgyus, Q5091, Robinson 1979, 61–62.
48. Decleer 2006, 25.
49. Dpe 111: ha ri bhunydza’i […] dharmā kṣha gho ṣha (Phun: dharma kṣha gho ṣha) […] bal gu’i yul du bzhugs pa’i bar he tan dhagho ṣha (Phun: bzhugs pa’i heta ndan gho ṣha); cf. Phun 544 l. 5;, Decleer [2006] Decleer 2006, 31; Templeman 2008, 267; and further discussion below (segment 11).
50. The alternative spelling or hypercorrection Umāliṅga, which goes back to the Shambhalai Lamyik (Shambha la’i lam yig) of the Sixth Paṇchen Lama (1738–1780), is noted by Tucci 1931, 693 n.1, and by Decleer 2006, 36. Waddell 1893, 59, spells Ugalinga (Ugraliṅga?).
51. Tucci (1931, 693) instead understands that Ulliṅga and Amuka (“Amuga”) both lie within a month’s sail from Sri Lanka. Decleer and Templeman read the itinerary differently: Ulliṅga and Amuka are each a month’s sail apart. Cf. Dpe 106, Phun 548 l. 2: sing ga la’i gling nas da bar zhag sum cu lhag tsam re rgya mtsho la gzings gtong dgos gnyis dang/ gling gnyis po de la’ang zla lam tsam re yod gsungs.
52. Dpe 106: de nas gling de dag gi mi dang lhan cig gru bo che bzo khyad chos phun sum tshogs pa zhig tu/ mi stong phrag lnga tsam dang lhan cig lho phyogs la gru bskyod (Phun: bskyed) de byon pas/ zla ba bzhi tsam na. Cf. Phun 548 l. 3; Decleer 2006, 26; Templeman 2008, 262.
53. Tucci 1931, 693. As Decleer (2006, 26) notes, “the text does not necessarily imply that all five thousand navigated in a single boat.”
54. Templeman (2008, 262) reads: “they [the islanders/locals] made a mighty vessel,” that is, that the ship was made at Amuka.
55. Valdez-Bubnov 2019, 561–562.
56. Dohardt 2023, 42–45.
57. Valdez-Bubnov 2021, 79–80.
58. Tucci 1931, 695.
59. Tucci 1951, 187–188 n. 1. For the volcano of Jhamigiri, Tucci may have been thinking of Sumatra’s Mount Kerinci, which is, however, located over 400 kilometers inland from Jambi.
60. Dpe 106: grong khyer gnyis kyi bar na ri lho byang du gsum tsam zhig ’dug pa’i gcig…; Cf. Phun 548 l. 5; Decleer 2006, 26; Templeman 2008, 262–263.
61. Abraham 2023, 124.
62. Cf. Decleer 2006, 12, who offers the synonym gla rtsi; Templeman 2008, 253.
63. Dpe 106: …gcig gi rtse mo na mtshan mo ’od ’phro ba’i gser gyi brag chung ngu zhig kyang snang gsung. Cf. Phun 548 l. 5; Tucci 1931, 693; Decleer 2006, 26; Templeman 2008, 263. The word brag is translated by Tucci as “cave.”
64. Tucci 1951, 188 n.1.
65. Dpe 106: zhing la sogs pa med cing/’bras ’bru re btab pa la sdong po mi ’grang lnga tsam yod pa/ khyim re’i ’tsho ba thams cad ’byung ba tsam yod pa. Cf. Phun 548 l.4. The present translation follows the intepretations of Decleer 2006, 26, and Templeman 2008, 262.
66. Pearce 1920, 57–58.
67. See e.g. Mills & Ptak 1996, 62.
68. For the longer legend see Bka’ babs bdun ldan 1, translated in Templeman 1983, 7–8.
69. Vörös 2022, 107, 112.
70. Vörös 2022, 108.
71. The expression “on Poṭalaka mountain of a small ocean island” (rgya mtsho’i gling phran po ta la ka’i ri la) is used in the scene-setting of a Buddhist sūtra, the Theg pa chen po’i mdo chos rgya mtsho, D255/Q921.
72. Templeman 2008, 146; 2021, 48–49, n. 13.
73. Henderson et al 2022, 54–57, 117–125.
74. Henderson et al 2022, 128–151.
75. Tucci 1940, 4–5 n. 8.
76. Tucci 1951, 188 n. 1.
77. Personal communication, May 5, 2010, and March 21, 2021.
78. Q4577. The relevant part of the text was summarised by Tucci 1951, 185.
79. Dpe 107: nam mkha’i ngos la nags tshal dang de nas ’bab pa’i chu dngos su ’bab pa; cf. Phun 550 l. 2. Decleer translates: “a forest adjacent to one side of the sky, but with below a waterfall that is actually coming down from there” (2006, 27); Templeman: “it was located in the heavenly forests and that water showered down” (2008, 264). The Tibetan phrase nam mkha’i ngos elsewhere translates Sanskrit ākāśatala (Rim pa lnga pa, D1802/Q2667, 49b). It might indicate a vertical cliff face on a mountain (which “‘faces the sky”’ horizontally), or mean that the waterfall is enveloped by the jungle canopy.
80. Dpe 108: rtse mo shel ma rdo’i brag ri rgya dpangs bsam gyis mi khyab pa yod pa. Cf. Phun 550 l. 3; Decleer 2006, 27; Templeman 2008, 264.
81. Anderson et al 2018, 32–33.
82. Dpe 108: gling de’i mi kun la da lta phyi nang kla klo sogs grub mtha’ gang yang med/ […] ’dzam bu gling na med pa la/ las las grub pa’i ’khrul ’khor mang pos khyim gyi mtha’ bsrung ba ’ba’ zhig ’dug. Cf. Phun 550 l. 5; Waddell 1894, 59 (who instead understands “towns containing Buddhists, Muhammadans and Toitankaras”); Tucci 1931, 649; Decleer 2006, 26; Templeman 2008, 264.
83. Purdy 2020, 237.
84. Dpe 108: gling ’dir gling gzhan nas ’ongs pa’i mi rna bas lus thams cad g.yog thub pa mang po gzigs. Cf. Phun 551 l. 1; Decleer 2006, 27; Templeman 2008, 264.
85. Translated in Cachey 2007, 116–117.
86. See Cachey 2007, 178 n. 415. For some inevitable drawing of parallels with Homo floresiensis on nearby Flores and other tales of pygmies in the region, see Forth 2008; Henderson et al 2022, 90–92.
87. See Cachey 2007, 60.
88. According to Pigafetta, Maluku was known in Cebu, Butuan and Sarangani in the present-day Philippines. Cf. Cachey 2007, 60, 78, 80.
89. Collingridge 1895, 130, 198, 209.
90. Dpe 108: gling de nas gru btang ste thag ring po byang phyogs su byon pas/ dza ba dwī pa zhes bya ba/ nas gling du phebs/ gling de na nyan thos sendha pa’i dge ’dun gyi sde mang du yod cing/ de rnams su ma bzhugs/ de na mtsho chung zhig gi dkyil na gling chung zhig yod de/ de’i ming ni/ ba na dwi pa ste/ nags gling zhes bya ba yin/ de na slob dpon mtsho skyes rdo rjes byin gyis ba rlabs pa’i gnas/ phyi brag ri nang lha khang gru bzhir yod pa/ /dbus na kyai rdo rje phyag gnyis pa’i rdo sku rang byung bzhugs shing/ phug pa gcig na gsang sngags kyi glegs bam mang du yod/ rgyud ’bum phrag lnga pa yang bzhugs zer te/ shin tu ’tshub par grags pas lta nus pa ni mi ’dug gsung. Cf. Phun 551 l. 1; Decleer 2006, 27; Templeman 2008, 264.
91. Here I thank John Newman for pointing out some of Tāranātha’s potential Tibetan sources for the term sendha pa (email, March 17, 2024), including the works of Sa skya paṇḍita (1182–1251).
92. See for instance Seven Instruction Lineages 8, translated in Templeman 1983, 95. The story is told here of “Javanese” monks refusing to respect Buddhaguptanātha’s tantric guru.
93. See Vajragarbha’s Hevajrapiṇḍārthaṭīka verses 1.5, 1.33 and so on, Sferra 2009, 444, 446. In the Tibetan translation (kye’i rdo rje bsdus pa’i don gyi rgya cher ’grel pa, D1180/Q2310) it is called ’bum phrag lnga pa’i rgyud.
94. Śāntigupta’s *Hevajrapiṇḍārthaprakāśa is preserved in Tibetan translation: Ngyes pa rdo rje’i lus kyi don rab tu gsal pa, Q4697.
95. Reichle 2007, 133–135.
96. See Sadananda 1938, 3: “The large island of Samosir entrenched all round by sky-reaching trees that grow on high hills.”
97. Tibetan translation: Kye rdo rje’i sgrub thabs kyi mdor bshad pa dag pa rdo rje sgron ma, D1237/Q2366.
98. Decleer 2006, 31, n. 45.
99. This segment is also analyzed in Ray 1936, 85–87.
100. Dpe 111: ha ri bhanyzda’i gtsug lag khang mchod rten chen po. Cf. Phun 554 l. 4; Decleer 2006, 31; Templeman 2008, 267.
101. D1179 …ha ri pu nydza’i yul gyi shākya’i dge slong chos kyi nyi ma zhes bya bas…, catalogued by Ui et al 1934, 191, under the title Bla ma dam pa la bstod pa.
102. In some later writings (Rgya gar chos ’byung, Slob dpon chen po spyod ’chang dbang po’i rnam thar) Tāranātha distinguishes between Dhanaśrī and Dhānyakaṭaka, but he may not have grasped this distinction when he wrote the Hagiography.
103. E.g. Ray 1936, 86.
104. For instance, the port of Dehnāsrī, named in the Mīrāt-i Aḥamadī of 1571, is identified with Tenasserim, see Bayley 1886, 18.
105. Here he refers to “Śrī Dhānyakaṭaka caitya in the island of *Dhanaśrī” (Chimpa & Chattopadhyaya 1970, 192, 208–209).
106. Dpe 111: shar du grong khyer shin tu che ba/ rgya nag dang/ phreng gi dang/ rgya gar la sogs pa’i yul tha dad pa’i tshong ba shin tu mang pa’i tshong ’dus chen po ’dug gsung. Cf. Phun 555 l. 2; Decleer 2006, 32; Templeman 2008, 267.
107. Dpe 112: po ṭa la zer ba’i rgya mtsho’i do ri gzhan gci. Cf. Decleer 2006, 32; Templeman 2008, 267. In his History of Religion in India, Tāranātha states that Poṭala is reached by travelling across the ocean from Dhanaśrī/Śrīdhānya (Chimpa & Chattopadhyaya 1970, 192).
108. Dpe 113: pai gu’i gling dang/ dza ti dang li shi sogs mang po skye ba/ phreng gis ’dzin pa’i gling zhig dang. Cf. Phun 555 l. 5. Decleer 2006, 32, takes these three phrases to refer to three different locations, whereas Templeman 2008, 267, construes two modifiers of phe gu’i gling.
109. Templeman (2008, 151), summarises all the meetings with Indians recorded by Tāranātha.
110. Dpe 112: sādha dhī pa (sic) dang/ gser gling chen po dang/ gser gling gi ming can chung ba gnyis/ /surdza dhī pa/ tsandra dhī pa/ sarba dhī pa ste/ gling chung chung de rnams su yang zhabs kyis bcags. Cf. Phun 555 l. 6; Decleer 2006, 32; Templeman 2008, 267.
111. Waddell (1893, 69) gives Sādadvīpa for Sādhadvīpa (Sādhudvīpa?) and Sarpadvīpa for Sarvadvīpa. Decleer (2006, 32) spells Sadhadvīpa and Sarmadvīpa (Śarma-?). These variants all appear to involve typographical errors.
112. According to a work related to the Śaiva Kubjikāmata, Puṇḍravardhana (present-day northwestern Bangladesh) is said to be located in Sūryadvīpa; see Dyczkowski 2009, 540.
113. See, for example, Somerville 1931, 166–168.
114. Dpe 113: …slob dpon nag po spyod pa pa’i rgyan drug mkha’ ’gro mas byin gyis brlabs pa de mtshan mo me ’bar ba sogs cho ’phrul dang ’tshub shin tu che ba cig ’dug pa de yang mjal. Cf. Phun 556 l. 2; Decleer 2006, 32 (who translates ’tshub as “occurrences”); Templeman 2008, 268.
115. Templeman 1989, 32–33, here translating from Tāranātha’s Slob dpon chen po spyod ’chang dbang po’i rnam thar.
116. See Mallinson 2019 for references to debate on this subject, and again Templeman 1997, 959. The works attributed to Kṛṣṇacārya are Q5180, Q5181, and perhaps Q5134. Buddhaguptanātha is elsewhere said to have stayed at a shrine built by Kṛṣṇacārya (grub chen nag po spyod pas bzhengs pa’i gtsug lag khang du) at Devikoṭa (Phun 110; Decleer 2006, 30; Templeman 2008, 268).
117. Playne (1917, 462) states that “a Sanskrit college for Pandits and a shrine of Siva […] were diluviated in 1842.” Long (1850, 538) states that it was the Kapil Muni Ashram that was washed away.
118. Somerville 1931, 168.
119. Ships in the Indian Ocean that lose their way were sometimes said to land in the Andamans. The thirteenth-century Chinese travelogue of Rukuo Zhao states that the Andamans are sometimes reached by mariners who are blown off course (Hirth et al 1911, 147–148). So too does Fei Xin, writing in the early fifteenth century, see Mills & Ptak 1996, 62.
120. Dpe 112: gling stong pa zhig na nor bu dmar po ’dra ba’i rdo’i phug pa dang/ tsandan gyi nags tshal yang yod pa/ shin tu gnyan pas ’dab ma re yang gcod mi nus pa zhig yod pa…; Cf. Phun 556 l. 5; Decleer 2006, 33; Templeman 2008, 268.
121. Bheemalingappa et al 2015, 72.
122. Dpe 112: sngon phags pa spyan ras gzigs dbang phyug sengge (Phun: sanggai) sgras thugs bskyed pa’i gnas su grags pa. Cf. Phun 556 l. 5; Decleer 2006, 33; Templeman 2008, 268.
123. Mills & Ptak 1995, 63; Mills 1970, 125.