From Bauddha Deśa to Śākta Pīṭha: Re/Locating the Hindu Goddess Tārā

Abstract

The incorporation of the goddess Tārā into the Hindu pantheon appears to have begun around the turn of the first millennium, a couple of centuries after her first mentions in Buddhist sources. The earliest Hindu texts concerned with Tārā tend to acknowledge this through a narrative wherein the Vedic sage Vasiṣṭha must travel to ‘Greater China’ to learn from the Buddha how to propitiate the goddess properly through the violation of brāhmaṇical purity codes for which Indian tantric traditions are infamous. Over time her ‘foreign’ associations faded, narratives linking Tārā to sites in Assam and Bengal became more prominent, and her worship drew closer to regional Hindu orthopraxy. This essay tracks the latter stages of that process especially through a reading of early modern ritual manuals in Sanskrit before considering a more recent revival of interest in the Hindu Tārā’s Buddhist connections as shown in Bengali sources and fieldwork.

If the crowds are not too thick, and they are not yet completely covered by flowers, one can see near the entrance to the sanctum sanctorum of the eponymous Tarapith temple a gilded pair of feet—a stylised representation of stone ‘footprints’ left by the goddess in the nearby cremation ground, much as the gleaming metal murti on public view in the temple is said to be a reflection of the self-arisen stone icon secreted away in another part of the complex (Fig. 1). The goddess Tārā left the original stone feet behind, we’re told, after appearing to either the legendary sage Vasiṣṭha or the nineteenth century ‘mad saint’ Bāmākṣyāpā, and conflict between or conflation of these origin stories is itself emblematic of two ways of accounting for the sacrality of the location, alongside the separate claim that Tarapith is, in fact, one the famous 51 (or 108) śakti pīṭhas: a ‘seat of power’ blessed by the pupil (tārā) of the goddess Satī’s third eye during her famous dismemberment in the purāṇas.

These three divergent, though not entirely exclusive narratives—that Bengal’s most sacred site of devotion to the goddess Tārā derives its primacy from the activities of the ṛṣi Vasiṣṭha, or of the sādhaka Bāmāksyāpā, or from its status as a śakti pīṭha—also have different implications for how people conceptualise the Hindu Tārā in relation to her more internationally revered Buddhist counterpart, particularly since the version of events privileging Vasiṣṭha strongly implies that this form of the goddess is in some sense an import from some predominantly Buddhist area called Mahācīna or ‘Greater China’ on the other side of the Himalayas. Presenting Tarapith as a śakti pīṭha, by contrast, lends the goddess and her ‘seat’ a pan-Indic if not properly purāṇic pedigree, while the figure of Bāmākṣyāpā holds more local Bengali significance.

Traditionally trained scholars do not seem to have taken much interest in locating this Mahācīna, but Indologists since at least the 1918 publication of John Woodroffe’s Śakti and Śākta¹ have attempted with questionable success to pin down where it might have been, and with it the putative sectarian origins of the goddess Tārā herself. Vasiṣṭha also finds the Buddha in the rather less ambiguous ‘Cīna’ (China) rather than ‘Mahācīna’ in some versions of the story, but fortunately, we can avoid wading too deeply into geographic debates here. Suffice to say that this Mahācīna appears to refer to someplace beyond the brahmanical heartland of Āryāvarta, and that Vajrayāna Buddhism was (or is) prevalent there. Whether it ‘actually’ existed in East Turkistan a.k.a. Xinjiang Province, or on Mt. Wutai; or somewhere in Nepal, or in Tibet (as is popularly assumed); or perhaps only in texts as an essentially symbolic border region like the legendary Kingdom of Women (Strī Rājya) in the vicinity of which it is sometimes said to reside,² is of no great consequence for us now since the story of Vasiṣṭha’s travel there is itself mythic in character.

Neither is it a primary objective of this essay to adjudicate whether the worship of this goddess first arose among Buddhists or Hindus: We shall see that the textual evidence weighs heavily towards the former, but ultimately we are concerned with the views of the latter, many of whom do not at least in this instance treat the distinction as especially meaningful. Nevertheless, some background on this point should help to clarify the stakes. As Philip Almond (2006) and others have shown, early orientalist scholars did not initially distinguish Buddhism from Hinduism, and when they did, the purportedly pure yet occluded traditions of Indian Buddhism they constructed on textual bases often served as foil for condemnations of alleged Hindu decadence and idolatry. Early academic appraisals of Tibetan and Himalayan religion often similarly stressed the idea that the Vajrayāna was essentially medieval Hindu Tantra in Buddhist (or perhaps ‘Lamaist’) garb. Shorn of some of its Victorian ideological underpinnings, scholarly debate over the issue has continued into the twenty-first century. Some of the clearest (and still clearly partisan) stakes concern the supposed wholesale appropriation of Śaiva ritual by Indian tantric Buddhists. This philologically-based conclusion championed by Alexis Sanderson (2009) has been hotly contested by a camp of Buddhologists best represented by David Ruegg (2007), who have posited instead a common religious ‘substratum’ from which all Indians, regardless of formal religious affiliation, could have freely drawn.

Similarly, over the centuries both Indian and Western scholars have fingered various phenomena as fragmentary and/or clandestine continuations of the India’s largely eclipsed Buddhist traditions—most obviously the Buddha avatāra of Viṣṇu alluded to above, and quite famously the Advaita Vedānta of Śaṅkara too, along with various aspects of Trika, or what is commonly termed ‘Kashmir Śaivism’ and a bewildering range of so-called 'folk cults.' There are several difficulties with this approach, not least the growing scholarly suspicion that commonplace reports about the ‘death’ of Buddhism in India may in fact be somewhat exaggerated. Douglas Ober (2023 pp.21–31) points to a number of biases that have heretofore naturalized this narrative, including an overemphasis on monastic institutions and doctrinal orthodoxy, a geographic bias centered on the modern Indian state, and an underestimation of Buddhist pilgrimage networks. If we allow literal and figurative borderlands more space in the frame, the story of Buddhism—and the memory of Buddhism—in the subcontinent looks rather more lively, even during those presumptively lost centuries between the collapse of the major monastic universities following the Ghurid conquests of Northern India in the early 13th century and the dramatic entrances of the Maha Bodhi Society and of Dalit Buddhism onto the 19th century stage.

The present study revisits one of the better-known examples of practices apparently transplanted from a Vajrayāna Buddhist to a Hindu tantric context, but will attempt to mitigate, as much as possible, the theoretically and politically fraught implications of foregrounding a search for origins (including the frame of ‘syncretism,’ which amounts to seeking the origins of parts of a thing.) Claims of ownership over particular sites, saints, texts, deities, ideas, and practices inevitably involve pragmatic concerns and rhetorical maneuvering, and it is in these that I hope to find firmer ground.

My own analysis of the various acculturative processes at work in Śākta Tārā traditions is partly synthetic and partly based on underexplored Sanskrit sources from the early modern period, contemporary Bengali materials, and ethnographic fieldwork at Tārāpīṭh. As such, I will take but a moment here to briefly name a few more local processes whose influence on matters at hand I suspect has yet to be fully appreciated. The first is the re-evaluation of Bengal’s Buddhist heritage in the wake of Haraprasad Shastri’s 1897^,Discovery of Living Buddhism in Bengal and his 1916 publication of Hājār Bacharer Purāṇa Bāṅgālā Bhāṣāy Bauddha-Gān O Dohā (Buddhist Songs and Couplets in One-Thousand-Year-Old Bengali Language). The former notably sparked public interest in the cult of the deity Dharmarāj/Dharma Ṭhākur, while the latter announced the apparent apotheosis of Bengali as the eldest of the Indo-Aryan vernaculars. Having thus become a source of regional pride, various studies chased alleged Buddhist ‘survivals’ from the Chittagong Hills to the lyrics and lore of householder Nāth Yogīs, Vaiṣṇava Sahajiyās, Bāuls, and of course various goddess traditions, alongside archaeological efforts to excavate the illustrious history of the medieval Buddhist heartland in the Ganges Delta.³

Buddhist revivalism in Bengal as discussed by Barua had its own impact in turn, as has the Tibetan diaspora with its sizeable presence in Darjeeling, and the growth of middle-class Bengali tourism in ethnically Tibetan/Nepali/Bhutanese areas in the Northern part of the state and in the Northeast. As a result, local historians and guides in West Bengal frequently take the opportunity to celebrate the monastic university at nearby Nālandā, internationally renowned 11^th-century Buddhist teachers like Atiśa Dīpaṃkara, and other Pāla Era contributions to Indian/Tibetan Buddhism, while Bengali books like Tārāpīṭhera Tārā Mā (Mother Tārā of Tarapith) may suddenly sprout roman script references to the legendary Vajrayāna guru Padmasambhava, the so-called Tibetan Book of the Dead (a.k.a. Bar do thos grol), or distinctively Himalayan Buddhist chöd (gchod) rituals (Gupta 2007, p.23). Arguably related effects visible in the broader culture include a prevalence of Tibetan script/imagery in Bengali supernatural thrillers,⁴ following longstanding subcontinental and orientalist associations of Himalayan peoples with occult practices. The reader is, therefore, advised whenever confronted by sources younger than a dozen decades or so to consider the possible effects of the developments or environments just described.

Tārā’s journey to the West

Tārā worship is attested in Indian Buddhist contexts from the seventh century (Landesman 2020, ‘Introduction’) but whether this includes the form current in Hindu circles is a separate question, since Tārā has taken on in many Buddhist contexts a role comparable to the Hindu Pārvatī in that she is presented as the generic Goddess from whom all other goddesses emanate. What people typically think of as the specifically Hindu form of Tārā is more properly known by the names Ekajaṭā (‘Having One Dreadlock’), Mahācīna[krama]tārā (Tārā of the Greater Chinese Method), Ugratārā (Wrathful Tārā), and Nīlasarasvatī (Blue Sarasvati). This goddess is also known among Buddhists by the first three of those names with the literal, word-for-word Tibetan translation of the Sanskrit ‘Mahā-cīna-krama-tārā’ (Rgya nag rim pa’i sgrol ma) suggesting that this particular form was at least initially transmitted from India to Tibet rather than vice versa.⁵ Indeed in the unambiguously Buddhist Tārā Mūlakalpa, an 8^th-century Sanskrit text now extant only in Tibetan translation, there are references to a manifestation of Tārā called Ekajaṭī (synonymous and interchangeable with ‘Ekajaṭā’) who acts as a fierce protectress (Landesman 2020, ‘Layer One’).⁶ Ekajaṭī’s appearance in this earliest text is unique to it, but within a couple of centuries this goddess had more or less settled into the fearsome iconography of the type now most widely associated with Kālī while retaining the blue lotus emblematic of Tārā throughout (as in Fig. 2). As Benoytosh Bhattacharya noted back in 1924, the following description of Mahācīna Tārā found in the 12^th-century Buddhist ritual compendium called the Sādhanamālā remains standard in most Vajrayāna and Hindu Śākta contexts⁷:

The worshipper should conceive himself as (Mahācīna-Tārā) who stands in the Pratyālīḍha attitude, and is awe-inspiring with a garland of heads hanging from the neck. She is short and has a protruding belly, and her looks are terrible. Her complexion is like that of the blue lotus, and she is three-eyed, one-faced, celestial and laughs horribly. She is in an intensely pleasant mood, stands on a corpse, is decked in ornaments of snakes, has red and round eyes, wears the garments of tiger-skin round her loins, is in youthful bloom, is endowed with the five auspicious symbols,⁸ and has a protruding tongue. She is most terrible, appears fierce, with bare canine fangs, carries the sword and the Kartri⁹ in the two right hands and the Uptala and the Kapāla in the two left. Her Jaṭāmukuṭa of one coil is brown and fiery and bears the image of Akṣobhya within it. (Bhattacharyya 1958, pp.189–90)

Gudrun Bühnemann (1996) has shown how a close variant of this passage was directly incorporated into non-Buddhist sources, including the circa 13^th-century Phetkāriṇī Tantra, from which it apparently spread to later East Indian Śākta tantras including the Mahācīnakramācāra Tantra/Cīnācārasāra Tantra/Ācārasāra Tantra, Nīla Tantra, Bṛhannīla Tantra, Tārā Tantra,¹⁰Śaktisaṇgama Tantra, etc. and to the authoritative 16th century tantric digests like Mahidhara’s Mantramahodadhi and Kṛṣṇānanda Āgamavāgīśa’s Tantrasāra. The most important Tārā-centric ritual manuals like Brahmānanda Giri’s contemporaneous Tārārahasya, its commentary Tārārahasyavṛttikā by Śaṇkara Āgamācārya a.k.a. Gauḍīya Śaṅkara (1630), and Narasiṃha Thākura’s Tārābhaktisudhārṇava (17^th century) likewise retain this form.

What sets this case somewhat apart from the general phenomenon Sanderson (2009, pp.240–2) dubbed the ‘reflux’ of Vajrayāna practices into late medieval Śākta Tantra is that unlike, for example, Mañjughoṣa (a.k.a. Mañjuśrī) or most of the other esoteric deities apparently sourced from Buddhist traditions, Tārā has acquired a substantial independent following in Bengal, with a central temple, pilgrimage economy, and popular devotional culture centred on Tarapith. She remains, however, historically and more widely known as the second of the Ten Mahāvidyā Goddesses whose worship constitutes the core of East Indian Śākta Tantra. Here Tārā, along with more widely worshipped goddesses Kālī and Lalitā Tripurasundarī, typically form a triad accorded pride of place above the remaining seven. Interestingly, we may also highlight at least one other Mahāvidyā among the second tier who also shares a Vajrayāna Buddhist pedigree: Chinnamastā.¹¹

As a group, the Śākta cult of the Ten Mahāvidyās appears to have come into its own sometime between the 13^th and 15^th centuries (Shin 2018, p.29); likely well after the Tārā dhyāna quoted above made its way into the Phetkāriṇī Tantra, and—I think not coincidentally—this was contemporaneous with the much-storied ‘decline’ of Buddhism in India, as weakened Buddhist institutions in a religiously competitive environment may have lost their capacity or historical prerogative to mediate devotees’ access to deities like Tārā. There is also some indication that this Buddhist activity receded from view more slowly in Eastern India (including Bengal) than in most of the rest of the subcontinent,¹² which may well be related to the persistence and popularity of the cult of Tārā as evidenced by many of the texts under consideration here as well as a handful of active Tārā temples in Bengal, Bihar, Odisha, and Assam. Tarapith is arguably the most important of these temples now, but it is unclear how long this has been the case. Older texts emphasise connections to Kāmarūpa in Assam, rather, and to Cīna/Mahācīna, with even the Tārā Mūlakalpa and a parallel passage in the Mañjuśrī Mūlakalpa listing Cīna and Mahācīna among the ‘remote countries’ outside of India which are meritorious ‘lands of accomplishment’ (Landesman 2020, ‘Chapter B.8’). Of particular interest for our discussion of Mahācīna are accounts involving the semi-divine Vedic ṛṣi Vasiṣṭha, whom we encountered briefly at the beginning of this essay.

Only Vasiṣṭha could go to Cīna

As a direct son of the Creator Brahmā and royal chaplain to the Raghu Dynasty, Vasiṣṭha appears in the Epics as a veritable custodian of the classical dharmic order.¹³ By contrast with his rival, the ascetic, formerly kṣatriya sage Viśvāmitra memorably characterised by Adheesh Sathaye (2015, p.117) as the ‘transgressive Brahmin Other’, Vasiṣṭha represents the more conservative vaidika Brahmin purohita—or perhaps more specifically the transitional Atharvin purohita whose pragmatic repurposing of Vedic rites on the state’s behalf paved the way for the eventual triumph of Śaiva tantric ritualists at royal courts.¹⁴ This is significant because our sources commonly contain episodes that depict Vasiṣṭha as the ideological counterpoint to Tārā’s association with borderlands or foreigners, either during his travels to Mahācīna for incredulous instruction in local forms of Śākta ritual, or through his curse that Tārā will be venerated by barbarians who employ the impure methods characteristic of the tantric Left.

The latter account occurs in the Kālīka Purāṇa, where Tārā is enjoined by Śiva to expel the inhabitants of the holy land of Kāmarūpa lest too many of them become immortal (Van Kooij 1974, p.169; Gogaoi 2011, p.237). She obliges, but unfortunately Vasiṣṭha was then among their number and pronounces a malediction in retaliation—his curse being presumably meant to explain the famously Śākta region’s general disinclination towards normative vaidika religion. This purāṇic myth evokes similar themes to the more widely reported tantric story by placing the paradigmatic Brahmanical sage Vasiṣṭha in tension with heterodox and geographically marginal yet apparently quite powerful religious traditions. But we should explore the tantric narrative in greater detail at this point as it deals more explicitly with the issues at hand.

The story of Vasiṣṭha’s travels to Mahācīna has often been traced to the venerable Brahmayāmala or/and Rudrayāmala—an ascription that lends the tale an unlikely air of antiquity since, according to Shaman Hatley (2018, p.6), Tārā material appears in but a single Bengali ‘Brahmayāmala’ manuscript apparently unrelated to and considerably later than its first-millennium namesake. Predictably, there is an early modern Bengali version of the Rudrayāmala as well, but Teun Gourdriaan and Sanjukta Gupta (1981, p.47) are ‘uncertain if an original…ever existed’ of the hypothetically older Rudrayāmala ubiquitously quoted throughout tantric literature. A search for the oldest non-Buddhist text even vaguely connecting Tārā with the Himalayas might bring us back to the Phetkāriṇī Tantra’s allusion to the popularity of Ugratārā among northerners (auttariya) (Bharati 1965, p.60) but a recognisable ‘Vasiṣṭha in Mahācīna’ narrative likely originates at least 300 or so years after the Kālīka Purāṇa in the East Indian Śākta tantras composed after the 15^th century. All these accounts, where we find them, hew closely to a common source for not only the iconography of Mahācīna Tārā (as previously discussed) but also the core Vasiṣṭha narrative and general ritual instructions.

The initial setting for this episode is also usually Kāmarūpa, where we find Vasiṣṭha frustrated by his unfruitful attempts to secure the favour of the goddess Tārā despite performing severe austerities and reciting her mantra over the course of several millennia. Just as Vasiṣṭha loses his patience and begins to curse Tārā’s seemingly ineffective mantra (or it is suggested, rendering the mantra/goddess ineffective), Tārā finally appears. She then informs the great rishi of the reason for his lack of success, which is that traditional Brahmanical rituals are unsuited to the task: ‘Only Viṣṇu in the form of Buddha truly knows the method of my worship, and no one else. Ignorant of my essence, you went to a lot of trouble for nothing during this time by practicing a blocked method’ (Mahācīnācāra Tantra 1:43–45). Having been instructed to seek out the Buddha avatāra, Vasiṣṭha sets out for the country of Mahācīna beside the Himalayas, where the Tārā Tantra characterises as both an Atharva Vedic and a Buddhist land (bauddha-deśa) (7: 19).

What the Buddha Avatāra taught

The texts do not comment on why the Buddha avātara resides outside of India in the first place, or how it is that the inhabitants of Mahācīna—a land ‘distinguished by its many Buddhists’ (bauddha-bahulo deśa-viśeṣa) (Tārābhaktisudhārṇava, p.113)—are so well versed in the Atharva Veda. In the first case, I think we can surmise that by the 15^th or 16^th century, the Indian authors of these texts must have thought of Buddhism as something practised beyond the Himalayas, and this association must have been firm enough in their minds to project it into the past. By implication, the memory of Indian Buddhism had faded to such an extent that the Buddhist past really was a proverbial foreign country. On the other hand, they clearly recognised or imagined commonalities with the religious practices of neighbouring Buddhist countries since these are labelled Atharva Vedic, even if this appellation primarily signifies that Mahācīna is a place where powerful magicians reside.

Vasiṣṭha, for his part, is shocked to find the Buddha avatāra surrounded by gods and sages and meditating on the Goddess while drunkenly cavorting with a thousand women. Scandalised, Vasiṣṭha initially objects that this is all anti-Vedic, but having been assured that this praxis is necessary to please Tārā, he requests teachings from the Buddha avatāra as previously instructed. The subject of this Buddha’s discourse is cīnācāra, the ‘Chinese Practice’, which generally consists of a loosely structured late-night pūjā in a charnel ground or similarly secluded place, during which the male adept offers blood sacrifices and spirituous libations to the Goddess while simultaneously also imbibing, feasting, and ‘worshipping’ women. Drawing perhaps on stereotypes about foreigners, Buddhists, and the freedom reputedly found in frontier zones, there is a strong emphasis here on ignoring customary ritual considerations regarding preparation, timing, directionality, and purity/pollution. The following passage from the Nīla Tantra neatly recapitulates the Buddha’s teachings:

While enjoying the pretty women and all the food and so on, engage in mantra recitation. There are no restraints as far as direction and time here, nor any restraints regarding obligatory rites etc. There is no time-restraint in the case of recitation or in worship etc., nor even in animal sacrifice. Here in this liturgy of the great mantra it is said that restraints are according to one’s wish. Nothing unlawful is found, Devī: anything could be great dharma. The method of practicing according to one’s will is upheld here. One should practice with a merry mind thinking “the thing is done,” but satisfied one way or another.

(Nīla Tantra p.37)¹⁵

What might mark this practice as characteristic of Cīna or clearly distinguishable from other tantric modalities like vāmācara and kaulācāra (or perhaps svecchācāra) is difficult to isolate at this stage,¹⁶ and it may be that such distinctions did not yet or do not entirely fit the situation these texts describe. We can, however, observe that the primary emphases here seem to be on mantra recitation and a studied indifference to standard ritual procedures.

Careful readers will have noticed the Goddess (Devī) addressed in the quotation above as well. This is because in the Nīla Tantra this passage is spoken directly by Śiva to the Goddess, whereas in other texts such as the Mahācīna Tantra (which it cites) most of this material is spoken by the Buddha to Vasiṣṭha. It would be tempting to speculate here that the author(s) or editor(s) of the Nīla Tantra excised the Vasiṣṭha/Buddha narrative from their treatment of cīnācāra—one of several brahmanising or at least ‘Indianising’ trends that will occupy our attention for much of the remainder of this essay. Various iterations of the story, for example, aver that having received these teachings Vasiṣṭha made his way home and achieved siddhi by these ‘Chinese’ methods: in older sources, the locale is usually Kāmākhyā on the northeastern frontier of India, whereas in more recent Bengali material, it has shifted farther west to Tarapith. Before stepping away from Sanskrit scriptural traditions, though, it should help to review developments to this point and highlight some similar features in these older strata of sources.

Bringing Tārā back home

The first and most important thing to underline is the effect of incorporating Tārā into the Mahāvidyā group, whose collective debut seems to occur within a couple of centuries of Mahācīna Tārā and cinācāra’s first appearance in the Phetkāriṇī Tantra. Although the early medieval Tārā Mūlakalpa (which would become part of the Tibetan Buddhist canon) also refers to Tārā as a mahāvidyā or ‘Great Mantra-Goddess’ and as Queen of Mahāvidyās (Landesman 2020, ‘Tārā’s Identity in the Tārā-mūla-kalpa’), the fixed roster of Ten Mahāvidyās are an unambiguously brahmanical phenomenon¹⁷ unifying the two most widely observed Śākta tantric traditions in the subcontinent: the cult of Kālī prevalent especially in Eastern India and the Śrī Vidyā tradition centred on Tripurasundarī, which has flourished most publicly in the South.

Between these two pan-Indic goddesses at the top tier of prestige among the Mahāvidyās, we find the comparatively regionally circumscribed Tārā. Why Tārā should be accorded such importance is not readily apparent, but one effect of having been given such a position is that in addition to the general assertion that all the goddesses in the sequence are manifestations of Kālī, who is universally regarded in this theology as foremost among the Mahāvidyās, texts often explicitly insist that Kālī, Tārā, and Tripurasundarī are identical; and moreover, that thinking otherwise results in damnation—a stance continued in Śākta sources like the Tārārahasya (2002 pp.34–35) more concerned with Tārā than the larger group. Over time, there developed a somewhat comparable set of eight goddesses said to be forms of Tārā (Buhnemann 2000, p.31), but these evince no obvious relation to any of the numerical sets of Tārās (most notably eight and twenty-one) known in Buddhist contexts.¹⁸ Rather we find in various Śākta Eight Tārā lists famous names like Kālī, Bhadrakālī, and Cāmuṇḍa; while contemporary visual representations of these goddesses like this one purchased in Tarapith may simply repurpose common commercial images of Kālī for goddesses known by other appellations (e.g. Mahogratārā in Fig. 3).¹⁹

As a knock-on effect of Tārā’s absorption into the Mahāvidyās and assimilation to Kālī, it seems that the scope of cīnācāra, from the time of its first appearance in the Phetkāriṇī Tantra, was expanded from ‘a code of ritual behaviour peculiar to Tārā worship’—as Goudriann and Gupta (1981, p.153) gloss what they translate as the ‘Tibetan method’—to include other deities, thereby diluting (intentionally or not) any lingering associations with borderlands or Buddhism. Ironically, these are the precisely cīnācāra’s most distinctive features: its introduction (as a term) alongside the goddess Mahācīna[krama] Tārā, its subsequent narrative association with the Vajrayāna through the Vasiṣṭha/Buddha story, and its restricted suitability to some subset of tantric deities. It would be one thing, perhaps even expected, if cīnācāra were also enjoined for the Mahāvidyā Chinnamastā at least, since she like Tārā has a well-known Buddhist counterpart, or for Man̋jughoṣa (better known as the bodhisattva Mañjusrī), but this is inconstantly applied. The voluminous 16^th–17^th-century Śaktisaṅgama Tantra (23:1, 38:2) does indeed mention cīnācāra in connection with these deities but the circa 1700 Mahācīnācāra Tantra (2:37–38) presents cīnācāra as obligatory only for the Mahāvidyās Tārā and [Dakṣiṇa] Kālī but optional for Tripurasundarī and Bhairavī.²⁰ Although a putative throughline may be obscured by employing cīnācāra to worship deities with no clear Buddhist associations, it is notable on the other hand that the practitioner of cīnācāra is sometimes instructed, as in the Śaktisaṅgama Tantra (21:40), not to touch a tulasī leaf or invoke the name of Hari. It seems likely that this apparently anti-Vaiṣṇava posture only applies during cīna-style ritual, although somewhat comparable sectarian polemics are not unknown in East Indian tantric texts composed after the time of Caitanya. In either event, it seems a little ironic to attribute these instructions to the Buddha Avatāra of Viṣṇu.²¹

How should we understand cīnācāra, then? The ‘Chinese Method’ is neither consistently restricted to Mahācīna Tārā or any presumably Buddhist-aligned subset of deities. In its broad outlines the Śākta presentation of cīnācāra appears nearly indistinguishable from other ritual modes occupying the antinomian ‘Left’ end of the tantric spectrum, even though Buddhist sādhanas of Mahācīnā Tārā do not specifically enjoin ritualised contravention of liturgical norms (Bühnemann 1996, p.476). And alongside this freewheeling ‘anything goes’ rhetoric, we do in fact find certain lunar days preferred, with particular botanical offerings prescribed and proscribed. Loriliai Biernacki (2008, p.31 ff.) has argued that the worship of women is the truly defining feature of what she glosses as the ‘Kālī Practice’, though that too is well within the extended repertoire of Left-hand ceremonial.

The refusal of obligatory and physically performed preliminary rites stands out a bit more to me, and perhaps too much so for later authors’ tastes, since despite the Phetkāriṇī’s explicit disavowal of timing considerations and twilight sandhyā rites in particular, Brahmānanda 16th century Tārārahasya (2002, pp.22–29) contains multiple reflexes of this classic śrauta ritual, complete with tantricised ‘gāyatrī mantras’ directed towards Tārā and her retinue which must be recited before their root mantras.²² A more overtly brahmanising gesture would be quite difficult to imagine! At the same time, the Śaktisaṅgama Tantra remains curiously invested in connecting cīnācāra to Buddhism—notwithstanding that cīnācāra as such is not a category used in Buddhist sources. The Śaktisaṅgama Tantra²³ does not equate Buddhist ritual as such with cīnācāra, but rather it distinguishes Brahmanical and Buddhist variants of the Chinese Method: The sort featuring ‘five parts’ (pañcavidha—probably referring to the pancāṅga-upāsanā involved in preliminary rites or puraścaraṇa viz. pūjā, japa, tarpana, homa, and feeding brahmins²⁴) is the Buddhist version, while the ‘partless’ is the Brahmanical (21: 3–4). Clearly, the author(s) believed that Buddhists were the stricter ritualists.

Other than skipping puraścaraṇa, what is the Brahmanical cīnācāra meant to entail? Having asked himself to explain how Brahmins proper can practise cīnācāra, the author quickly mentions vegetarian and non-alcoholic substitutes for offending aspects of the ritual before declaring that he has already discussed the practice With Parts and that one ought now to hear about the Partless. The remainder of the chapter establishes the primacy of mantra repetition (21: 22) and a non-dual approach to ritual such that pūjā and tarpana, etc., should be performed mentally, adding that ‘Brahmā, Vasiṣṭha, and other Lords of Sages’ always followed the Partless Path (21: 48–62).

We can see that the Śaktisaṅgama is hardly pro-Buddhist in an unqualified way, and the same is true of the broader Śākta tradition. Much like the Buddha Avatāra in some of the purāṇic sources discussed recently by Clough,²⁵ the Devī is said to manifest delusory forms for the destruction of Buddhists (Bhattacarya 1999, p.86),²⁶ and authors of Śākta padāvali devotional poetry in 18^th-19^th century Bengal, according to Rachel McDermott (2001, p.241), tend particularly to refer to the goddess as ‘Tārā’ with reference to her delusory capacity. However, the Śaktisaṅgama does not accord with the tendency Bharati (1965, p.70–71) identifies as normative in Śākta texts to treat cīnācara as only effective for worldly benefits, but not liberation. Rather, as in the ninth chapter of the Nīla (Sharma 1965, p.37) and others we are told that ‘one with liberation as the goal obtains liberation’ via the Cīna Method, while ‘one with wealth as the goal obtains wealth’.

The Śaktisaṅgama is an idiosyncratic work, which does include some rather surprising Buddhist material such as mantras from the Guhyasamāja Tantra (Bharati 1965, p.75), but as a rule overtly ‘Buddhistic’ elements or imagery would need to be reframed to facilitate their theological integration. Even if one suspects the authors of most of our later texts especially had never discussed esoteric religious matters with a Buddhist, they were likely aware that educated Buddhists would take issue with the animal sacrifices, theistic philosophical positions, and appeals to non-Buddhist scriptural authority that these Śākta texts advocate. Still, after the first round of edits to the Mahācīna Tārā dhyāna incorporated in the Phetkāriṇī Tantra, during which some of the more obvious Buddhist elements like the epithet ‘Tathāgata’ were smoothed out (Bühnemann 1996), there were relatively few heretical nāstika signifiers left standing to require revision.

The main issue, as others have noticed, seems to have been the image of ‘Akṣobhya’ in Tārā/Ekajaṭā’s crown; a common feature in Vajrayāna imagery indicating to which of the families headed by the Five Buddhas (pañcatathāgata) named Akṣobhya, Amitābha, Amoghasiddhi, Ratnasambhava, and Vairocana a deity belongs. The invocation of another member of this fundamental set, Vairocana, during the puja (Tārārahasya, 2011, p.59) might pass without comment, but a miniature blue Buddha Akṣobhya atop the Tārā’s head was enough to raise some eyebrows, as this particular iconographic convention is uncommon outside of Buddhist contexts. Several interpretations have been proposed. The Toḍala Tantra reckons that Akṣobhya is the name of the particular ‘bhairava’ (consort/protector) associated with this form of Tārā (Goudriaan and Gupta 1981, p.81), which seems commensurable with Brahmānanda’s gloss of akṣobhya (‘unshakeable’) as ‘Rudra’, i.e. Śiva (Tārārahasya, 2002, p.40)—even if the very same source later claims (p.47) that Akṣobhya is the name of the ṛṣi who revealed the mantra. Gauḍīyaśaṅkara, in his Tārārahasyavṛttikā (p.21), speculates further that Akṣobhya could be an oblique reference to an indestructible crescent moon (avināśi-candra-kalā) signifying Śaiva initiation, adding that it is also the name of ‘the guru in this tradition who sits in the dreadlocks in the form of the nāga king Vāsuki’. Indeed, due to later iconographic developments on the Hindu front, ‘Akṣobhya’ is often simply signified by or identified as Ekajaṭā’s serpentine hair-tie, presumably fading into generic snakes-and-dreadlocks Śaiva iconography for the less scrupulous observer.

What kind of Pīṭha is Tarapith?

The fate of Akṣobhya is a minor emendation indeed when viewed alongside the most commonly displayed image of Tārā in Bengal, the Tārā of Tarapith. As a reputed reflex of the rather abstract self-arisen stone murti removed from public view but widely understood to represent the goddess nursing an infant-sized Śiva, this Tārā appears to be essentially conical. Three eyes are clearly visible under an enormous and elaborate crown, but any limbs or implements, if they exist, are seen only by the priests who dress the image. The main temple icon is, therefore, most easily recognised by the stylised smear of red completely covering the lower half of her face and outstretched tongue (Fig. 4). Although Vajrayāna-derived imagery remains prominently displayed around the temple and the town, whatever place the older Tārā iconography may have had in popular culture is being rapidly replaced by this popular local form. Visually, it conveys nothing so much as an affinity to Kālī—for whom the tongue is often a visual shorthand—and a fondness for sacrifice, while the backstory pointing towards its ‘secret’ stone prototype anchors the goddess in mainstream purāṇic mythology, since this maternal Tārā is supposed to have nursed Śiva to cool down his blue throat following his consumption of the oceanic hālahala poison (Morinis 1984, pp.166–7). Some local writers in Tarapith, including Prabodha Kumara Bandyopadhaya (2010, p.22), naturally assert that the maternal image of Tārā current there is far older and more representative of who the goddess really is than the essentially Buddhist Mahācīna Tārā/Ekajaṭā iconography in which she temporarily appeared during the reign of the Buddhist Pāla Dynasty (8^th–11^th centuries CE).²⁷

Similarly, if ‘Mahācīna Tārā’ was ever in any sense an actual topotheonym for a deity of a particular locale, that too has been substantially displaced at multiple earlier stages. Sanskrit narratives concerned with this goddess focus their geographic attentions not on Mahācīna, Cīna, Bhoṭa (Tibet), or Nepal, but almost exclusively on Assam, where local legend holds Vasiṣṭha deposited an image he brought from Mahācīna at the Ugra Tārā temple in Guwahati (Shin 2018, p.30). One possible exception—if the migratory handful of ślokas attributed to it could be reliably dated—might be the Guptacīnācāra Tantra (Secret Cīnācāra Tantra) regularly cited by modern Bengali sources keen to promote Tarapith as an ancient śakti pīṭha. Either way, both moves decentre the bauddha deśa of Mahācīna (wherever it might have been) in favour of existing subcontinental Śākta shrines, and while Kamakhya is easily the more famous, locating Tārā in Bengal does literally move her closer to the traditional Brahmanical centre.

In the process, a different kind of distance from older narratives, concepts, and sources of authority accrues as well, such that even the Vasiṣṭha narrative and its intimations of a Buddhist past for the goddess becomes an optional flourish instead of a binding chain of transmission. In several accounts purportedly based on this Guptacīnācāra Tantra, the original stone murti at Tarapith spontaneously appeared when Vasiṣṭha’s tantric practice in the charnel ground beside the site of the temple finally bore fruit. Like those ‘polyvalent padas’²⁸ that originally appeared to Vasiṣṭha—or was it Bāmākṣyāpā?—the statue and the temple also have multiple lives.

Unlike Kamakhya, the status of Tarapith as a śakti pīṭha remains contested even among the faithful. The simple devotional appeal for those who relish living near the site where Satī’s third eye fell during purāṇic times is obvious, but it also represents the possibility of significantly improved prestige, pilgrim traffic, and infrastructure funding for the locality—undeniably powerful motivators that people have begun to experience more over the last decade of state investment in and promotion of the temple town.²⁹ Rhetorically, it is also one of the strongest ‘Hinduising’ moves that could be made with regard to the goddess Tārā, regardless of whether anyone involved consciously thinks of it in such terms.

To date, firm textual support for the antiquity of Tarapith’s śakti pīṭha claim remains somewhat elusive, though, with partisans relying primarily on vague allusions in or questionable attributions to classical sources. One pamphlet from a local ashram (Mukhopadhyay 2000, p.1), for example, cites a verse from Brahmānanda Giri’s Tārārahasya explicitly naming Tārāpīṭha as a śakti pīṭha and attributing the appearance of the stone Tārā mūrti there to Vasiṣṭha’s efforts, but I could find no such verse in the published editions of the Tārārahasya or its commentaries to which I had access.³⁰

The Śivacarita, although a relatively late and demonstrably Bengal-centric work, is another popular source in which Tarapith is identified as the location of Satī’s third eye. D.C. Sircar (1967, p.39) argues that some recension of the Śivacarita must have existed by the middle of the 18^th century since the famed Bengali poet Bhāratacandra Rāya relies on the text,³¹ but the Śivacarita likely did not include Tarapith at that stage since Bhāratacandra—presumably following an older version of the Śivacarita closer to the Pīṭhanirṇaya on which it is thought to have been based—locates all three of Satī’s eyes at Śarkarāra near the modern Pakistani city of Karachi and associates them with the goddess Mahiṣamardinī rather than Tārā (1996, p.41).

Probably Tarapith was not widely considered a śakti pīṭha during the 1750s, and it is possible that the temple itself was still obscure. Even pious writers like Debashis Mukhopadhyaya (2000, pp.14–16), committed to the idea that an ancient Tarapith temple housing the stone icon revealed to Vasiṣṭha had been destroyed in a flood, insist that it was ‘rebuilt’ in the early 18^th century after an interlude of several centuries. Possibly, the Guptacīnācāra Tantra itself (or a few key lines from it) emerged around this time.³² Regardless, to the extent that the identification of Tārā as the presiding goddess of a śakti pīṭha sticks, it provides her with a much more familiar and higher status purāṇic origin story easily capable of eclipsing its comparatively obscure tantric alternative.

On the other hand, the sacrality of the site need not depend on the antiquity of a text, or a temple, or even on the dismemberment of a deity. A cremation ground as lively and storied as the one at Tarapith could easily have put it on the map as a siddha pīṭha, a ‘seat of adepts’, rather than a śakti pīṭha—with a Vasitha/Mahācīna tie-in as an option but not a requirement. A great many sādhakas (Śākta tantric ‘aspirants’) are said to have performed tantric sādhana among those cremation pyres over the centuries, including the celebrated Śākta padāvali poet Kamalākānta Bhāṭṭācārya (1769–1821) (McDermott 2001, pp.116–20). Still more famous and revered especially at Tarapith itself is the 19^th-century ‘Madman of the Left’ Bāmākṣyāpā. To walk through the town of Tarapith today, one might be forgiven for taking Bāmā as the primary focus of local devotion; an incarnation of Śiva with Tārā as his consort. In several ways, Bāmākṣyāpā cut a similar figure to his contemporary Ramakrishna (Banerjee 2010; Ramos 2017), but being of a more overtly Śākta tantric bent his image and teachings have proven slightly less adaptable to the global stage.

He has become instead the subject of several Bengali films (e.g. Joy Maa Tara, Mahāpīṭha Tārāpīṭha, Sādhaka Bāmākṣyāpā) and two teleserials, including the popular ‘Mahāpīṭha Tārāpīṭha’ that began in 2019, in which Bāmā has been cast as something of a social reformer. Although it is impossible to predict precisely how such developments in Bengali popular culture might play out, significantly increased interest in the Tārā of Tarapith as mediated by Bāmākṣyāpā seems more or less guaranteed to expedite existing movement toward Hindu devotionalism at the expense of historical or otherwise esoteric approaches to the goddess that could draw attention to her background among the legendary Buddhists of Mahācīna. Even still, narratives connecting Bāmā and other ascetics or devotees from Tarapith with Tibet are still in circulation. For example, June McDaniel (2009, p.155) interviewed a sādhaka who spoke of one ‘Aghor Baba’ who studied tantra extensively in Tibet but clashed with Bāmākṣyāpā and was forced to leave Tarapith. Elsewhere (2004, p.261) she encountered a devotee of Nīlasarasvatī who we are told specifically followed the Cīnācara Tantra and claimed that his Sarasvatī icon had been originally brought to Bāmākṣyāpā from Tibet. Further south in Nadia district, I myself heard from a priest at an Ekajaṭā Puja an echo of the ‘Vasiṣṭha in Mahācīna’ story in which Bāmākṣyāpā had to travel to Tibet to learn Tārā practice from a Buddhist lama, which suggests that Tarapith and Tibet are still connected somehow in public memory.

Coda: Borrowing From Oneself

In his oft-celebrated demolition of Comparative Religion’s theoretical foundations, “In Comparison a Magic Dwells,” Jonathan Z. Smith somewhat acerbically characterizes an early twentieth century scholar’s position on ostensively foreign ideas found in late antique Judaism as follows: “Jews did not borrow, for what they ‘borrowed’ turned out to be already their own.”³³ While this kind of circular reasoning is certainly problematic within Religious Studies scholarship, it may, by analogy, aptly summarize the general posture of Hindu writers considered here towards the goddess Tārā and her putative Buddhist background in Mahācīna. By way of conclusion, I want to reiterate that I have no reason to believe that all, or even most of the people engaged with the discourse(s) around Tārā over these last four centuries have done so with the express aim of sanskritising, brahmanising, or otherwise harmonising the goddess Tārā with Śākta Hindu norms. Most of the emendations made to texts were probably viewed as ‘corrections’, most additional rituals deemed necessary, and most interpretative shifts have likely been the result of thoughtful people trying to make sense of something that seemed out of place in the tradition as they received it. There are plenty of natural, aesthetic, and historical reasons for cultic centres to rise and fall or move throughout the landscape, including individual charisma and political patronage. I do not imagine that I have exposed an elaborate conspiracy to memory hole the true history and essential character of Mahācīna Tārā, especially given the near-constant and recursive cross-appropriation between Hindu and Buddhist tantric traditions.

At the risk of overextending Bharati’s metaphor (1965, p.319) of the relationship between Buddhist and Hindu tantra as an area of mutual amnesia, I am more interested in what circumstances trigger certain memories or allow them to be recovered, and how people attempt to reintegrate them or reconcile themselves to them. While the presentation of cīnācāra in the Phetkāriṇī Tantra may have been an authorial wink (or a blink) in the direction of its Buddhist source material, by the time the Vasiṣṭha in Mahācīna story appeared in the 16^th century or so, it likely encoded some notion of Tārā as a deity popular among Buddhists in the Tibetan Cultural Area—an idea perhaps surfaced or reinforced through travel between Bengal and Nepal, as has been frequent among East Indian Daśanāmī ascetics like Brahmānanda Giri. For the late medieval/early modern Indian authors or redactors of the Tārā Tantra, then, including the Vasiṣṭha narrative or the epithet ‘Buddheśvari’̄ for the second Mahāvidyā in 7:9 may look less like an act of ‘agonistic appropriation’ than a familiar instance of the old Travelling Abroad for Hidden Knowledge trope; a little proto-orientalism as a treat.

Only after the onset of colonial modernity would associations between antinomian ritual and Buddhist–Hindu exchange be fully weaponised. By this time, the cult of Tārā as observed according to the Cīna Method must have become widespread enough to draw unwanted attention, with Tārā, in particular, directly identified with wine even by the staid and reformist Mahānirvāṇa Tantra (McDaniel 2004, p.105). Swami Vivekananda, whose campaign to ‘root out that filthy vāmācara’ is quite well documented, could then single out ‘the tantrik practices of Tibet’ as the ‘origin and essence of all the Tantras which are now current among us’. Woodroffe, too, while adopting a more detached posture, similarly suggested on multiple occasions (Urban 2003, p.158) that infamous tantric sexual rituals are of Chinese or Tibetan extraction, a theory that Kripal (1995, p.29) rephrases as ‘all that is decadent in tantra came from the Buddhists’.³⁴

Even today, among devotees in Tarapith, a slightly inverted and softer variation on this general theme feeds into tantric apologia, as when amidst his discussion of cīnācāraBandyopadhyay (2004, p.32) assures us that authentic and efficacious Left-hand Hindu ritualists can still be found despite the bad reputation justly attached to their misguided and derivative Buddhist counterparts:

Buddhists tāntrikas during the Pāla era, being unable to understand the deeper meanings within the guru-to-student teachings on the extremely secret true form and symbolism of the Five Ms sādhana³⁵ according to Hindu tantra, modified Hindu tantra a little and created a new Buddhist 5 Ms sādhana. They took up those mostly external rituals by mistake/misconduct, but those able to correctly value that path advanced on the proper path of sādhana.

Interestingly, Bandyopadhyay sets up this parallel en route to an even more surprising claim (2004, p.33 ff., 2010, p.20 ff.) that the Vedic ṛṣi Vasiṣṭha and the Buddhist siddha Nāgārjuna were the same person! The reason for this is as follows: firstly, Vasiṣṭha as we know is described in some of the tantras as having brought the Cīna Method for successful worship of Tārā to India; and secondly, the iconography of the Hindu Tārā is effectively drawn from the Mahācīna Tārā description in the Buddhist Sādhanamāla. The rhetorical fulcrum here is that one of the Ekajaṭā sadhanas presented nearby in the same anthology contains a colophon (Bhattacarya 1913, p.80) claiming to have been recovered from Tibet (bhoṭeṣu uddhṛtam) by Nāgārjuna.³⁶

So, the thinking goes, if Mahācīna Tārā and Ekajaṭā are the same goddess, and Mahācīna at least somewhat plausibly refers to Tibet from whence her worship was brought (or returned) to India, then Vasiṣṭha and Nāgārjuna must be Hindu and Buddhist names respectively for a single sage who retrieved and promulgated these practices. This unlikely but inventive inference demonstrates nicely how Hindu devotees of Tārā continue to engage creatively with the centuries-long project of relocating their goddess from Mahācīna to India.

Acknowledgements

Much of this research from which this article draws was carried out during an extended pandemic fellowship at the International Institute for Asian Studies in Leiden, for which I am grateful. Thanks also to Bradley Clough and Gudrun Bühnemann for helpful comments at an early stage in this project; and to Peter Bisschop, Shaman Hatley, Péter-Dániel Szántó, Arjundeb Sensharma, Biswajit Majumder, Harry Spier, Keith Cantú, Daigengna Duoer, Alex Meng, Or Porath and colleagues in the Leiden Working Group on the Fundamental Problems and Methods in the Science of Religion, and my generous anonymous reviewers for valuable conversations and suggestions that have improved this work. Such flaws as remain are of course my own.

Funding

Funded by Gonda Fellowship at the International Institute for Asian Studies (IIAS) during 2020-2021. The J. Gonda Foundation is part of the Royal Netherlands Academy of Arts and Sciences.

References

Footnotes