A Comparison Of Shaiva And Buddhist Tantra
by Christopher Wallis
The main purpose of this paper is to gather in one place evidence that has been piling up for some years in secondary sources, but the full implications of which are not yet apparent to most scholars who are not specialists in the philological analysis of the primary sources for the study of Tantra. The main thesis of this paper is that in the medieval period, Tantric Buddhism (mantranaya, vajrayāna) and Tantric Śaivism (mantramārga) were conterminous, coeval, and co-functional. In fact, I believe the evidence supports the notion that these two were co-functional and conterminous to roughly the same degree as Śaivism was with Brahmanism (vaidika-dharma), circa the 10th century CE, thereby belying the notion that the latter two can be considered two branches of a single “Hinduism” in the period under discussion. The consequence of this is to more firmly establish the invalidity of the use of the term Hinduism as an emic signifier prior to the 11th century, though that is a corollary and not the primary purpose of the present paper. 
While earlier phases of Buddhism have received much scholarly attention, that of Tantric Buddhism has not, for misunderstandings about it and negative valuations of it by European and Indian scholars once proliferated. (Tribe 2000: 193) The latter were the result of the (often unconscious) projection of normative views based on the tacit assumption of the ethical and intellectual priority of European religious values. The (so far partial) relinquishment of such views has been one of the more fruitful outcomes of the deconstruction of culture and value systems by later 20th century philosophy, leading directly to the opening up and reappraisal of the desirability of doing Tantric studies. But the amount of work to be done is daunting. Prominent Indologist and philologist Harunaga Isaacson estimates that there are approximately 2,000 Tantric Buddhist works in Sanskrit, the vast majority of which have not been published in their original language, let alone translated. (Isaacson 1998: 26) He suggests that therefore “a comprehensive history of Indian tantric Buddhism is something that may not be possible within our lifetimes” (1998: 27).
However, it is certainly possible to sketch the broad parameters of the development of Tantric Buddhism in India, at least in terms of its textual corpus. Before doing so, we must approach the issue of what we mean by Tantric Buddhism, for despite extensive scholarly usage of the term, we are still far from a clear consensus as to what “tantric” means in any context. The easiest definition, and the one that refers to the origin of the term, is to call “tantric” the doctrines and practices taught in texts calling themselves tantras (the adjectival form being tāntrika, though this is rarely used in the literature itself except to refer to some of those who follow the tantras). But this will not do, for many texts that contain material we treat as tantric call themselves –sūtras, -āgamas, -saṃhitās, and so on. Though the Sanskrit word tantra can simply mean text (as in the Pañcatantra), it can also mean “system,” as in a system of practice, and this seems to be the sense in which the Tantric texts use it, for the majority of tantras (or āgamas) in the Śaiva or Bauddha context teach a complete cultus, though sometimes they expected that the reader would supplement the instructions given from associated paddhatis or kalpas (ritual manuals) or other tantras. By extension, tantra refers to a scripture teaching a specific system of practice, which is the meaning we see in perhaps the only indigenous ostensive definition of the term tantra, that of the Śaiva Rāmakaṇṭha (fl. c. 950-1000), who writes: “A tantra is a body of teachings, constituting the revealed command (ājñā) of the Lord, which establishes the obligatory injunctions (vidhi) and prohibitions (niṣedha) for the His worship, preceded by the exposition of the special [[[Wikipedia:initiatory|initiatory]]] consecrations of those eligible for the higher and lower aims of human existence [i.e. mokṣa and siddhi].” (tantraṃ ca—parāparapuruṣārthādhikāriṇāṃ viśiṣṭasaṃskārapratipādanapūrva[ka]m īśvarārādhanāya niyatavidhiniṣedhaṃ tadājñātmakaṃ vākyajātam) 
Now, this precise definition is nearly equally functional, I argue, for the Buddhist tantras, if we simply replace “the Lord” with “the Buddha(s)” or “Bodhisattva(s)” (which itself seems to be a revealing fact).  Of course, we must carefully note that in these Tantric contexts, “worship” involves (at least) the actual evocation or summoning of the deity, and usually entails ritually or meditatively becoming that deity. Rāmakaṇṭha’s definition, then, highlights several features that scholars agree are key to the tantric mode of Indian religions: 1) concern with ritual modes of manipulation (of the environment or one’s own awareness), 2) requirement for esoteric initiation (to receive access to the scriptural teachings and practices), 3) a twofold goal of practice: the soteriological and supramundane one of liberation (variously conceived) and/or the mundane one of extraordinary power over other beings and one’s environment, and 4) the claim that these three are explicated in scriptures that are the word of God (āgama) or the Buddha ( buddhavacana).
This would seem to be sufficient definition, except that it is almost wholly concerned with practice rather than doctrine, and our Western prejudices cause us to want to ask “but what did they believe?” Or at least, how did they think about what they did, and what did they consider important? And it does seem to be the case that there is a commonly held vision of reality, if not actual doctrines, across the Tantric traditions, to which we will come. For now, it will suffice to list some of the features of Tantric Buddhism that scholars of the latter have enumerated in their attempt to come up with a useful scholarly definition of their subject. In so doing, they have invoked the notion of “polythetic definition” popularized by Jonathan Z. Smith on the basis of the work of biologists Sokol and Sneath. Anthony Tribe, writing the chapter on Tantric Buddhism in the volume by Paul Williams called Buddhist Thought, enumerates the following characteristics (discussed 2000: 197-202):
centrality of ritual, esp. evocation and worship of deities centrality of mantras visualization and self-identification with deity necessity of initiation / esotericism / secrecy importance of the teacher (guru, ācārya) ritual use of maṇḍalas transgressive/antinomian acts revaluation of the body revaluation of the status and role of women analogical thinking [esp. microcosmic/macrocosmic correlation] revaluation of 'negative' mental states
Now, all of these categories apply equally to Tantric Śaivism, despite the fact that Tribe was not concerned with Śaivism in formulating them, which supports an argument that will be given later about the validity of the scholarly construction of a general category called Tantra. But let us compare this to the definitional list given by another fine scholar, Stephen Hodge, in the introduction to his voluminous translation of the Mahāvairocanābhisambodhi Tantra, where he highlights what Tantric Buddhism added over and above non-tantric Buddhism (2003: 4-5, summarized and reordered here):
alternative path / new revelation lay/householder practitioners addition of mundane aims, achieved through largely magical means special types of meditation that aim to transform the individual into an embodiment of the divine after a short span of time new technologies: maṇḍalas, mudrās, mantras, dhāraṇīs creative contemplation/visualization of deities proliferation of in the number and types of deities centrality of guru and initiation linguistic mysticism rituals from non-Buddhist sources spiritual physiology (i.e. subtle body) importance of the feminine and utilization of “sexual yoga”
We begin to realize how complex this religious milieu is when we see how few things are held in common between the two lists (five items in bold above). And once again, all of these apply to Tantric Śaivism as well. To these two lists, Donald Lopez’s polythetic definition adds only a handful of less crucial features: the symbolism of the vajra, the cultivation of bliss, and spontaneity (sahaja, a technical term).  One much earlier definition of Tantra should be briefly mentioned here: Snellgrove (1959: 138) lists the process, goal, and means as “the evoking of deities and the gaining of various kinds of siddhi by means of various kinds of mantra, dhyāna, mudrā, and maṇḍala.” But this definition is insufficiently detailed.
Now, with the matter of definition in hand (or rather, with its intractable complexity adumbrated) we can proceed to a brief historical overview of Buddhist Tantra. The relevant indigenous categorization of texts, in chronological order of their appearance, is as follows (with the number of texts in the catalogued Tibetan canon parenthetically enumerated, as the Indian canon has not been catalogued yet):
Kriyā (Action) Tantras (450; e.g. the Anantamukhadhāraṇī-sūtra, Amitābha-dhyāna-sūtra) Caryā (Observance) Tantras (8; e.g. Mahāvairocana Tantra) Yoga Tantras (15; e.g. the Sarva-tathāgata-tattva-saṅgraha-sūtra) Yogottara aka Mahāyoga Tantras (37; e.g. Guhyasamāja Tantra) Yoginī aka Yogānuttara Tantras (82; e.g. Hevajra Tantra, Caṇḍamahāroṣaṇa Tantra, Laghusaṃvara Tantra)
We will address briefly the vexed question of dates. Some Buddhist scholars have implied or stated that Tantra begins with Mahāyāna Buddhism; but we cannot conclude that this is the case. Such a perception derives from the fact that the so-called “Action Tantras” (#1 above) were composed from at least the beginning of the 3rd century ce. However, the contents of these texts do not allow us to label them as properly Tantric, for they do not possess a significant number of the criteria outlined in our polythetic definitions above.  Stephen Hodge concurs, saying “[in the] early tantric phase only a few of these elements may occur together in any given text,” (2003: 5) though he nonetheless uses the term “tantric” prematurely. Indeed, the early “Action Tantras” are simply mantra-manuals (mantrakalpa) for achieving worldly ends, whether natural or supernatural. They contain no doctrine at all (Granoff 20; Isaacson 28), no requirement for initiation, and nothing that would impel us to call them Tantric; nor do they call themselves tantras (Tribe 205). They are, I would argue, simply the appearance in the Mahāyāna sphere of the type of techniques called paritta in Śrāvakayāna Buddhism (Tribe 206; cf. Davidson 2002: 117). They are, Hodge says, “for the benefit of unsophisticated ordinary people beyond the confines of the great monasteries” (2003: 7). Davidson argues most strongly that we cannot take esoteric elements in early texts of the so-called Kriyā Tantras as constituting a unified system (= tantra, see above p. 2), for “[t]hey were not considered constituents of a self-contained path and did not contribute a sense of identity within the community . . . The ritual secrecy, the transmission of separate precepts, the intimate connection between master and disciple—these had not come together in a self-aware manner.” (2002: 117) This would come, he says, only in the second half of the seventh century.
The true advent of early Tantric Buddhism is in the late sixth century (Gray 1997: 1). The Caryā Tantras (#2 above) begin at this period, with the other classes of text following rapidly after, with the Yoginī Tantras (#5) appearing starting in the early eighth century (Gray 2007: 2).  We saw above that Rāmakaṇṭha specifically defines Tantra in terms of a two-fold goal, the soteriological goal of highest liberation on the one hand, and the worldly one of power and pleasures on the other; and it is in this period that we have Tantric Buddhist texts epousing both. Isaacson also takes the soteriological element, and the self-conscious articulation of a path to it, as crucial for the category of “Tantric Buddhism” (1998: 27);  and so does Newman, as we saw above.
We also have negative evidence for this dating; I will give but two examples. Śāntideva, a somewhat conservative Mahāyānist who flourished 700-750 ce, only prescribes practices that we might call tantric in five short passages (Hodge 2003: 9). The Chinese monk Xuan-zang (aka Hsüan-tsang) travelled around India until 645 ce and wrote a detailed account which does not mention tantric practices (Hodge 2003: 9). Just thirty years later, two different Chinese pilgrims do refer to tantric practices: Yi-jing (aka I-ching) calls them “new teachings,”  while Wu-hsing (aka Wu-xing) refers to the new and exceptional popularity of the esoteric path in India around the same time (Davidson 2002: 118).
Extensive evidence for full-fledged Tantric Buddhist texts, fitting our criteria above, exists for the period of the seventh century on and need not be presented here. It is in the mid-to-late seventh century that Tantric Buddhists are self-conscious enough to name their path, choosing the term “Vajrayāna” (Tribe 196). This indicated a nascent awareness of the path as a separate vehicle from the Mahāyāna; the emergence of the term Vajrayāna is coeval with appearance of soteriological tantric rites. Up to this point these esoteric practices were referred to as the mantranaya or ’way of mantras‘ as complementary to (and compatible with) the pāramitānaya or ’way of perfections‘ that constituted the teachings of the Mahāyāna (Snellgrove 1987: 122). Even after the designation Vajrayāna began to be used (and note that the variation Mantrayāna followed soon after), Tantric Buddhism was intimately tied to the Mahāyāna, and articulated relatively few original doctrines. As Isaacson comments,
It is important to realize that many [[[Tantric Buddhist]]] texts say little or nothing about doctrine; and when they do, many offer little more than some allusions or simple explanations of well-known ideas and teachings of the Mahāyāna. Tantric Buddhism thus in the main sees itself as having its own identity distinct from non-tantric Mahāyāna Buddhism primarily in the sphere of means (upāyā) . . . (1998: 28)
That is to say, the Vajrayāna distinguished itself as a different vehicle only in terms of the techniques used to reach Buddhahood.  These comments should not of course be taken to indicate a lack of creativity on the part of Vajrayānists, for they were enormously innovative in the area of technique. Moreover, the Vajrayāna already had appropriate doctrines to fit its agenda for practice, in the form of the Yogācāra school of the Mahāyāna discussed above. Its idealist standpoint, that all things are ultimately consciousness and therefore awareness is fundamental to practice, perfectly fit with many Vajrayāna techniques, such as those involving an interiorized ritual performance whereby one identifies with a bodhisattva deity (or one’s Buddha-nature) and reimagines—and therefore recreates—the world in the image of the deity’s maṇḍala, in a sacred mapping of reality as ordered hierarchies of divinity. In the context of Yogācāra principles, vivid visualization led to identification, and identification when strongly established re-shaped reality. In other words, thoroughly believing oneself to be a Buddha brought about Buddhahood, if done in the right (scripturally prescribed) way. It was for this reason that the Vajrayāna could claim that it was a faster path to the goal. On the whole, its relationship to the Mahāyāna might be characterized as parallel to the complementary relationship between Sāṅkhya and Yoga that is envisioned by the author of the Gītā, where they can be translated as ‘theory‘ and ’practice‘ respectively (BG 2.39). 
We may now proceed to the main thesis of this paper, that Tantric Buddhism and Tantric Śaivism are conterminous, coeval, and co-functional. In fact, I believe the evidence supports the notion that they are co-functional and conterminous to at least the same degree as Tantric Śaivism is with Brāhmanism, thereby belying the notion that the latter two can be called branches of “Hinduism,” at least for the early medieval period. The interchange of Śaivism and Brāhmanism is well documented; what is less frequently acknowledged is the extent to which Tantric Śaivism and Tantric Buddhism borrowed freely from one another, creating marked parallelisms primarily in practice, and sometimes in thought as well. They even have synonymous names: the emic name for Tantric Śaivism is the Mantramārga, a parallel with Tantric Buddhism’s Mantrayāna as well as the earlier Mantranaya. But much more convincing evidence will be presented below.
First I will briefly examine doctrinal presuppositions shared in common as well as intellectual exchange and interdependence in philosophical discourse. Then I will demonstrate how both Śaivism and Buddhism formulated a religious hierarchy of truth that included the other as a lower level of revelation. Thirdly, I will touch on the issue of ritual co-functionality, and fourth, on outright appropriation from each other on multiple levels. Finally, I look at the movement towards total assimilation in the late medieval period. All this evidence will further serve to show that modern scholars have not gone entirely astray in their attempts to formulate a interreligious category called Tantra.
We noted above that with very minor exceptions, all of the criteria that Buddhist scholars identified with Buddhist Tantra apply equally to Śaiva Tantra. This is perhaps because Tantra is fundamentally a way of viewing reality and a modality of practice that became unmoored from the specific religious context it evolved in, making it malleable, protean, and exportable. It is a form, not so much a content, and is therefore adaptable and adoptable. Without denying any of the elements in the useful polythetic lists given above, my ostensive (non-polythetic) and non-sectarian definition of Tantra, partially following David White, is: an Indian interreligious movement driven by a ritual practice presupposing initiation, oral instruction from a guru, and micro-meso-macrocosmic correlations, and utilizing mantras, meditative visualizations, and sometimes antinomian means to access and experientially assimilate the divine energy of the (variously conceived) Godhead, in order to achieve power, pleasures, and liberation. This definition, while dense, seems to encompass as much that is essential of what we would like to call Tantra as possible. 
Śaiva and Bauddha Tantra are nearly coeval and closely parallel in the structure of their canons. Tantric Buddhism as defined above began in the sixth century. This is also the era of the earliest scripture of Tantric Śaivism, the Niśvāsatattva-saṃhitā, the bulk of which was composed by 550 ce (Goodall and Isaacson 2007; Sanderson 2006: 152-55). There is a key difference, however: the Niśvāsa represents the irruption into the textual record of a tradition that must have been developing orally for some time, for the text is a vast (4,500 verses approx.) and complex work, treating not only ritual but theology, cosmogony, and cosmology. Therefore it seems very likely that Śaiva Tantra predates Bauddha Tantra, for the contemporaneous Bauddha Tantric texts show none of this complexity, sophistication, and nuance, even if we were to date the Niśvāsa nearly a century later. Sufficient evidence of this to convince the scholarly community has only recently become available, with the publication last year of an edition of the oldest portions of the Niśvāsa (ed. Dominic Goodall et. al., IFP/EFEO, 2015).
The development of the two canons is parallel, with the exception that the Śaiva texts of the Mantramārga are very much concerned with doctrine ( jñāna) for they are largely unconnected from any earlier textual strata compared to the Bauddha scriptures of the Mantrayāna, which as we have seen are relatively unconcerned with doctrine, for all necessary doctrine had already been enunciated by the Mahāyāna.
The first phase of the Mantramārga, that of the Śaiva Siddhānta, is dualist, non-transgressive and does not include worship of the feminine in any substantial way (though it did initiate women); it may be compared to the Caryā and Yoga Tantras in the Bauddha canon, and the texts called “Father Tantras” by the Tibetans, though the textual output of the Śaiva Siddhānta is at least two or three times more substantial (Dyczkowski 1988).  The second main phase, sometimes called the Mantrapīṭha or the Svacchanda-Bhairava sect, introduces mild transgressive imagery and worships the deity and his consort more or less equally; this more or less parallels the Yogottara Tantras, which introduce erotic elements. Finally, the third phase of scriptural Śaivism, sometimes called the Vidyāpīṭha (where vidyā is a feminine equivalent of mantra), is the most highly transgressive: including both mortuary and erotic elements in abundance, it is very focused on worship of the feminine, and the most emphatically non-dual in its doctrine. Thus it is parallel to the Yoginī Tantras, which have all the same features; unsurprisingly, as it is this phase of Vajrayāna texts that is most directly indebted to the Śaiva sources. 
These goddess-oriented texts, on both sides, are usually written in a register of Sanskrit that deviates from Pāṇinean norms and reflects vernacular influence. This does not necessarily indicate that they were written by non-elites, but probably that most of them were composed in marginal regions, outside the brāhmanical heartland of Āryāvarta in the Gangetic plain. This style was in time established as a scriptural register of Sanskrit, and bizarrely, we have evidence of later scriptural authors attempting to imitate the “poor” Sanskrit of these texts. At any rate, their language constrasts greatly with that of the sophisticated exegetes of the scriptures. On both Śaiva and Bauddha sides, the exegetes flourished from the 10th to the 13th centuries, and both sides produced some towering intellectual polymaths in this period (e.g., Utpaladeva and Abhinavagupta; Abhayākaragupta and Ratnākaraśānti). Of course, this was also the period in which Tantric Buddhism was meeting its end in India and transitioning into Tibet, or else, I suspect, these parallelisms (and indeed convergence) would have continued, even to the point of syncretism.
These two Tantric traditions also contain significant parallels in the theory that undergirds their practice. In a 2003 article, Francesco Sferra outlines some of these, citing numerous primary sources. Both traditions emphasize, as we have seen, the prerequisite of initiation and the importance of the guru who grants it (Sferra 2003: 63 and fn12). In both traditions, initiates are enjoined to view the guru as an aspect of the deity and never to speak ill of him (Tribe 198-99). Secondly, Sferra notes that for both these two traditions, liberation means in some sense becoming the deity (ibid.), whether through merging with it, realizing one always was identical with it, or becoming a deity separate but equal to the one evoked. Thirdly, he correctly argues that both traditions, despite their hyper-ritualism, possess what Sanderson has called a “subitist” dimension of spontaneous gnosis: the notion that immersion into our true nature ( svabhāva) is possible at any moment (whether that nature is conceived of as śūnyatā or śivatva).
Sferra states, “Although the concepts relating to the true nature of the self are very different, the underlying belief is the same.” (2003: 67) This might seem a surprising and even an intellectually careless claim, yet scriptural passages can be cited to support it. A number of Śaiva texts describe Śiva, the highest aspect of reality, as a stainless void, a clear transcendent space of pure awareness. And Sferra’s statement above simply means that both traditions assert that our inner being or true nature “lies behind adventitious maculations” (ibid.) and that experiencing it, through whatever means, is the goal. He wonders if there is any real difference between the concept of śūnyatā and the Śaiva one of sarvaṃ sarvātmakam (a question taking us out of philology and into philosophy, which I think Indologists could stand to do more often). He notes that a verse expressing this sentiment, of unknown provenance (though Abhinavagupta attributes it to the arhats), is found in a number of Bauddha, Śaiva, and Vaiṣṇava works (2003: 69, fn 25, 26): eko bhāvas tattvato yena dṛṣṭaḥ sarve bhāvās tattvatas tena dṛṣṭaḥ | eko bhāvaḥ sarvabhāva-svabhāvaḥ sarve bhāvā ekabhāva-svabhāvāḥ || “One who sees one entity as it really is sees all entities as they really are. One entity has the (same) innate nature as all entities, and all entities have the same innate nature as any single entity.” (Or, if you like, replace 'entity' [[[bhāva]]] with 'being'.)
Sferra’s point then, as I take it, is that this principle underlies the various different religious theories about what exactly that innate nature is, and that the different names given to it obscure the parallel thinking at work. He notes further that both Śaiva and Bauddha traditions require a reliquishment of the notions of “I” and “mine,” the vikalpas that veil the truth (2003: 68). Finally, Sferra notes that both traditions entail in their soteriological sādhanas a process of de-identification with one’s conceptual representations of the past and the future and immersing oneself in the consciousness of the present moment (2003: 74-75, primary sources cited); and both allow for the possibility of final liberation through gnosis alone (2003: 77 fn41, primary sources cited).
Exegetes of both traditions cite the other sympathetically, despite the de rigueur attacks on one another (which Sferra argues are generally clichés, stock criticisms, and simplifications); after citing a number of examples, Sferra writes,
"More generally, it is evident that Buddhist doctrines constituted an important standard of comparison for many exponents of Hindu Tantrism, at least in exegetical literature; one only has to think how heavily authors such as Utpaladeva and Abhinavagupta depended on Buddhist logicians just to develop their technical terminology." (2003: 73, fn34)
Which leads us to briefly consider the interdependence of the philosophical discourse of these traditions, which is a feature mainly of their exegetical phase. Davidson presents the notion that in the seventh century Buddhist philosophy underwent a “turn to epistemology,” that is, began to engage in debate using the pan-Indian, intersectarian terms of epistemological discourse rather than the terms and arguments endemic only to the Buddhist tradition. (2002: 102-5) This shift, he suggests, both strengthened and weakened the Buddhist tradition in various ways, for it brought Buddhist thought into a wider sphere of debate, equipping it with tools but compromising its religious capacity for self-validation. The main proponents of this development were Dignāga (early sixth century) and Dharmakīrti (first half of the seventh century), the latter of which was more influential, possessing a highly refined analytical mind and writing in an extraordinarily dense and concise kārikā style of high-register Sanskrit. It was Dharmakīrti whom the Śaivas took on as their principal non-Śaiva interlocutor, as well as Dharmottara, one of his successors. For example, Dharmakīrti is quoted no less than 44 different times in a single chapter of a work by Śaiva exegete Rāmakaṇṭha. 
It has long been noted that ritual is an identity marker in Indian religious culture, and that sectarian boudaries are more likely to be drawn along ritual lines than doctrinal ones (Granoff 2000: 399). However, in an important article, Phyllis Granoff broached the subject of what she calls ritual eclecticism, that is, the permission or even injunction on the part of one Indian religion to perform the rituals, use the mantras, or worship the deities of another (while acknowledging it as other). She presents much interesting evidence there, but does not draw from her evidence a conclusion which I would regard as fundamental and significant, that all ritual eclecticism takes place on the level of mundane pursuits (siddhi and bhoga), with none involving soteriological rituals. I will summarize some of the evidence she presents, together with corroborating sources. My argument here is not that the religious boundaries were blurred in the early medieval period, but rather that the different traditions saw themselves as fully co-functional with each other on the non-soteriological level, Buddhism no less than the rest. This problematizes the common back-projection of Hinduism onto the medieval period.
It is not surprising that the least successful (in terms of numbers of adherents, amount of patronage, etc.) of the early medieval Indian religions was also the most permissive: Jainism, as represented by the Yogabindu of Haribhadrasūri, enjoins its hearers to worship all the gods equally (!), saying
“And so the teacher of religion, who knows that his disciple, like a dumb beast [paśu], is not capable of discriminating with respect to rituals such as the worship of the gods, and who therefore directs him to perform all rituals, to worship all the gods without distinction in order to accomplish some speciﬁc goal, is not to be faulted in any way." Yogabindu 15-16 (Granoff 2000: 403)
The implication of “some specific goal” here is a non-soteriological goal. It seems as though Haribhadrasūri, knowing that Jainism did not offer enough in the way of worldly ritual technology to compete in the religious marketplace, encourages Jains to use whatever means necessary to accomplish their worldly ends. I would propose that this was seen as preferable to losing adherents by denying them such permission. But let us turn back to Buddhism, for it too engaged in such eclecticism, though in a more restricted manner. Granoff explores sections of the Mañjuśriyamūlakalpa, a later Kriyā Tantra class text (and thus a relatively early text of Tantra in general, as defined above) of which she says, “There is nothing inherently Buddhist about the rituals of the [text] or about its goals” (2000: 418). We do see on p. 15 of the text the injunction that “a bodhisattva is told both not to abandon his own mantra nor to let go of the mantras belonging to others,” and on the following page the instruction to “not belittle either laukika or lokottara mantras” (2000: 404) where of course the former refers to non-Buddhist mantras. All mantras, the text declares, of any sect or religion (citing those of Vaidika, Śaiva, and Vaiṣṇava deities) can be utilized, as long as they are brought into a Buddhist context, either by prefixing the seed-syllable (ekākṣara) given by the text or by reciting them before its maṇḍala. (2000: 405) Indeed, the implication seems to be that it is the text’s maṇḍala, conceived as Buddhist, that renders them efficacious. In other words, by integrating them into a Buddhist vision of the world, they become valid.
The Caryā Tantra called Mahāvairocanābhisambodhi-sūtra (its commentator Buddhaguhya calls it a tantra; Sanderson calls it “our first major Buddhist Tantra” at 2009: 101) exhibits a similar attitude, for it instructs the listener to draw the laukika divinities in the outer layer of its maṇḍala (such as Agni, Yama, the Mothers, Indra, Varuṇa, Āditya, Sarasvatī, Viṣṇu, Skanda and so on), implying their ritual worship as well. (Ch. 2, sections 49-55) Such Inclusivismus (cf. Paul Hacker) is attested in a number of other Buddhist sources as well.  Buddhaguhya describes these deities as “Adornments of the Bhagavat’s Inexhaustible Body” and argues that as such (i.e., in a Buddhist context) they are more powerful. (Hodge 2003: 117) Additionally, in its first chapter, the Mahāvairocana presents a hierarchy of nine different types of mind, all subordinate to the bodhicitta. The first seven of these generate merit through moral observances, leading to the appearance of a “spiritual friend” who instructs one in the worship of the laukika divinities, such as the Vaidika (Brahmā, Viṣṇu, Śaṅkāra [sic], etc.) and popular divinities (Nāgas and sages), saying “These gods are great gods: they bestow happiness upon all. If you respectfully make offerings to them, you will become fortunate in all things.” (Hodge 2003: 62-3) The text clearly sanctions such worship for people of lower aptitudes, saying “This is the most excellent refuge of the foolish common folk who wander lost in saṃsāra” (ibid.). Higher than this, the eighth mind denotes those of the Sāṅkhya/Yoga school (the ones who pursue kaivalya), but higher than them of course is the mind that seeks to generate bodhicitta (though that specific term is not used: see Hodge 2003: 72-3).
Śaivism exhibits this electic quality far less often, which is not surprising; since it was the dominant religion of the early medieval period, it did not have to be as accommodating. We see some eclecticism, however, in the influential Śaiva text called the Netratantra (early 9th cen.), which teaches the worship of Amṛteśvara-bhairava, later called Mṛtyuñjaya. The text declares that Amṛteśvara can be worshipped not only in the form of any of the other Śaiva deities (unusual enough, within Śaivism), but he can also be worshipped in the form of Viṣṇu, Sūrya, Brahmā, Gaṇeśa, and the Buddha or indeed “all other deities,” by prefixing their names to Amṛteśvara’s mantra (e.g., OṂ JUṂ SAḤ BUDDHĀYA NAMAḤ). (Sanderson 2004: 245) This surprising assertion is explained by Sanderson as due to the fact that the purpose of this text was to give Śaiva officiants the ritual tools necessary to take on the position of the rājapurohita or king’s royal chaplain, traditionally the province of Atharvavedin brāhmans. The audacious attempt to replace orthodox Vaidika officiants in this crucial role is itself evidence of Śaivism’s success, and Sanderson further presents evidence that they were successful to some extent, at least in the far north (Sanderson 2004) and that Atharvavedins responded by appropriating Śaiva techniques (Sanderson 2007)! As for the inclusion of the Buddha amongst the forms in which Amṛteśvara could be worshipped, Sanderson explains this by saying that the Buddha was one of the deities of the local religious calendar of required observances in Kāshmīr, as documented in the Nīlamata-purāṇa (Sanderson 2004: 254), and that the officiant of the Netra had to be able to worship all such deities on the king’s behalf. None of this is explicit within the Netra-tantra itself, which of course purports to be divine revelation. The worship of the Buddha was also done for the women of the court, as we can infer from the Netra’s statement that the Buddha’s special role is “bestowing the reward of liberation upon women” (ibid.), perhaps because Kashmirian queens in this period were often patrons of Buddhism. At any rate, none of this contextual information invalidates my claim for a degree of co-functionality between Buddhism and Śaivism. Just as Buddhism had its laukika dimension, within which it admitted co-functionality with other deities, Śaiva texts too spoke of a laukika realm, which for it was generally the Vaidika one.  The fact that it extended co-functionality to the Vedic realm was not evidence of co-religiosity however, for just like Buddhism, Śaivism did not admit of any soteriological value to any Vaidika rite.
Finally, we have the negative evidence for ritual eclecticism, in the forms of texts that forbid it. Clearly, such prohibition indicates that the prohibited was being done by some group(s). Weighing in against hybrid practice (śabala karma) we have Śaiva works like the Sarvajñānottara, Tantrāloka (e.g. 4.249-51), Somaśambhupaddhati, and Mataṅga-pārameśvara (CP 2.2-7), practical Sthāpaka-tantras like the Mohacūrottara, and later texts like the Parameśvara-saṃhitā, which warns against vyāmiśrayājin, those who mix together different rituals (Granoff 2000: 420). These are in direct opposition to works like the 12th cen. Dānasāgara, which not only mixes ritual materials from Śaiva, Vaiṣṇava, and Saurya purāṇas (ibid.) but condemns those insufficiently syncretistic purāṇas that have clearly been redacted from sectarian Tantric Śaiva scriptures (e.g. the Liṅga-purāṇa). On the side of the syncretists we also have a number of medieval donative inscriptions that seem to indicate blurred religious boundaries (Orr 2000).
As a final piece of evidence for co-functionality through ritual eclecticism, we have the publications of Gudrun Bühnemann, who has done enormously detailed work on ritual texts of the late medieval period. While it is impossible to go into those details here, Bühnemann’s works of 1999, 2000a, 2000b, and 2001 contain much detail about the appropriation of Buddhist deities and even sādhanas into Hindu ritual manuals.  Table 1, attached, presents clearly some of the evidence of appropriation she has gathered (confirmed also by Sanderson 2009: 240ff). Three very brief and particularly fascinating examples will be given here in the text. The Īśānaśivagurudeva-paddhati (ĪŚP), a late Śaiva manual, imports into its Mantrapāda (MP) the earth goddess Vasudhārā from Buddhist sources and assimilates her to Lakṣmī, and further imports her retinue and mantras: the practitioner is instructed to “bow to the Buddhas and the Bodhisattvas in the beginning of the worship ritual” (Bühnemann 1999: 306)! Furthermore, it also imports Jambhala; what is particularly strange about this is that Jambhala is a Buddhist deity created specifically as a parallel to the popular Kubera deity (they are described with similar appearance and deformities and both called “lord of the yakṣas”), but the ĪŚP MP already includes Kubera himself. (Ibid.) This clearly shows that the text was aiming to be as complete as possible. Evidence that it was influential is seen in the fact that extensive passages from it are adopted by the later Tantrasāra-saṅgraha. Finally, Bühnemann notes the bizarre irony by which the ĪŚP MP gives mantras for the destruction of the enemies of the Vedic dharma (including Buddhists), which include the mantra of Yamāntaka (1999: 314)!
We may sum up this section with a key statement from a Buddhist authority that perfectly captures the sentiment behind this ritual eclecticism or co-functionality. Puṇḍarīka asserts in his Vimalaprabhā (quoting the [[[Bauddha]]] Yoginīsañcāra for support) that all darśanas have equal value (“from the point of view of the relative truth of the world”) for realizing mundane perfections. He cites commonly held theories on the nature of action and calls them “equal” (tulya) and states: “there is no difference [on this level] between the Buddhists and the tīrthikas. The only difference concerns the reality of emptiness.”  (Sferra 2003: 73-74) Isaacson has noted that this verse is found in other Bauddha sources, and a similar sentiment in Śaiva sources as well (Sferra 2003: 74 fn36). Here we have the means to begin to resolve the debate about a “common religious substratum” first put forward by Ruegg in 1964, followed by Beyer and others, criticized by Sanderson starting in 1994, with a rejoinder from Ruegg in 2001 and recently discussed by Sferra (2003) and Gray (2007b). Sanderson rightly rejected the substratum model (SM) insofar as the substratum posited was an undocumented and undocumentable popular folk tradition that somehow intruded its iconography and practices into the rites of the elite. However, the evidence we have seen suggest that there is a common religio-cultural milieu, which might be called a substratum, comprising precisely the observable religious and cultural elements shared by all Indian religions in some measure. No-one would deny that ubiquitous demigods like the nāgas, yakṣas, vidyādharas and various other supernatural spirits are “common cultic stock” to all these traditions, nor that certain practices, like that of the homa, became interreligious and universal. So the actual “substratum,” I would argue, is the one hidden in plain sight, the very sense of commonality that allowed the borrowing that Sanderson has argued for to take place; as Ruegg said in his 2001 rejoinder, the “background of shared categories and concepts” against which borrowing made sense. The debate, then, should focus on what belongs to this common pan-Indian tradition, and what doesn’t, and why or why not. To conclude, a useful insight into this knotty problem is furnished by Lopez (1996: 88), citing Heesterman on the fact that in the Vedic context, there are two opposing but complementary terms: pradhāna, denoting the ’primary’ part of the sacrifice, different for each deity, and tantra, denoting the auxiliary acts that were interchangeable among different sacrifices. It could be that Tantra as a ritual phenomenon diffused across religious boundaries initially precisely because it (at first) applied only to the doctrinally-neutral laukika realm of the pursuit of siddhi, where religious interchange happened freely, as opposed to the transcendent, specific, and soteriological “pradhāna” realm, different for each religion. Once the precedent was established on the tantra level, it became tempting to then surreptitiously borrow within the pradhāna realm too, as we will see.
Ritual eclecticism, as defined by Granoff, specifically refers to the borrowing of others’ mantras and rites, while implicitly or explicitly acknowledging them as other. This, then, is different from ritual appropriation, which attempts to conceal the act of borrowing. This concealment was evidently considered necessary whenever done in the realm of what we called the pradhāna dimension above, but better labelled as lokottara or pāramārthika. For in this soteriological realm, any acknowledgment of borrowing would clearly weaken the status of the religion’s claim to scriptures revealed by omniscient (and supreme) deities unique to that religion. Yet extensive appropriation did take place nonetheless. Detailed research recently published by Sanderson, running to hundreds of pages of evidence, confirms that Śaivism was the dominant religion in the early medieval period, thereby explaining the extensive borrowing from it by other religions (which he also demonstrates) as an attempt to compete for royal patronage. Sanderson writes,
There can be no doubt that for several centuries after the sixth [[[Śaivism]]] was the principal faith of the elites in large parts of the Indian subcontinent and in both mainland and insular Southeast Asia. Only Mahāyāna Buddhism was able to rival it during this period; and when it achieved success in this rivalry, either equalling or excelling Śaivism as the beneﬁciary of patronage, it was in a form led by the Way of Mantras, a system of ritual, meditation and observance in which Buddhism had redesigned itself, if not in essence, then at least in style and range of functions, on the model of its rival. (2004: 231)
Yet this borrowing was not all one way. I will try to show, briefly, that firstly Śaivism probably borrowed from Buddhism in the constitution of its exoteric dimension, and that secondly (as Sanderson describes) Buddhism certainly borrowed from Śaivism in the constitution of its esoteric dimension.
To proceed with the first, the two religions were structurally parallel in terms of their institutions. Though not yet widely recognized by scholars, textual and epigraphical evidence suggests that the elite of the orthodox current of Śaivism (the Śaiva Siddhānta) administered sizable complexes of temples, monasteries, and associated institutions supported by government land grants in many parts of India (Sanderson 2005 and 2009). Both religions possessed a large community of uninitiated householders or “lay” adherents, participating in the religion on an exoteric level and without access to scriptural texts or oral transmission, and nevertheless contributing substantial economic resources. Śaivas borrowed frequently and obviously from the earlier Buddhist model (as well as the Brāhmanical) in the constitution of this exoteric dimension of their religion, centered around temples, āśramas, and associated institutions (Sanderson 2004: 231). This exoteric domain was overseen by the orthodox Saiddhāntika Śaivas, who referred to the laity as upāsakas, the same term used for them in the Bauddha and Jaina spheres. We also see distinctly Buddhist practices prescribed for these Śaiva upāsakas (also known as Rudrabhaktas), such as taking sympathatic joy in the meritorious practices of other devotees (puṇyānumodanā), ritually generated at the three sandhyās of the day.  Furthermore, we are told by the Śivadharma (“The Religion of Śiva”), before the Rudrabhakta does pūjā to Śiva he must generate the four sentiments of maitrī, kāruṇya, muditā, and upekṣā; in other words, the four Immeasurables or brahmavihāras of Buddhism. The Śivadharma and Śivadharmottara also describe religious institutions such as monasteries (maṭha), managed by Pāśupata or Saiddhānta ācāryas, clearly modelled on Buddhist ones, that request and receive substantial donations (śivadāna) from the public and from the king, and also serve the public through dispensaries (ārogyaśālā) and charitable food distribution; inscriptions provide evidence to corroborate the texts (Sanderson 2002, 2004, 2005).
This borrowing, however, was not as thorough and pervasive as the borrowing that took place in the eighth to tenth centuries that produced the last phases of Tantric Buddhism, those of the Yogottara and Yoginī Tantras. Sanderson has presented extensive evidence for this, primarily relying on the science of text-critical philology, in a seminal article of 1994, a more detailed one in 2001, and his Koyasan lecture of 2006,  and of course the monumental 263-page (plus 43 pages of bibliography) 2009 monograph “The Śaiva Age.” Independent corroboration is given by the detailed analysis of the passages borrowed from the Brahmayāmala into the Laghuśaṃvara in Hatley 2007: 176-83. Sanderson’s pre-2009 evidence is summarized in simple tabular form in Table 2. To mention the most salient and indisputable example, Sanderson has discovered that of one the central Yoginī Tantras, the Laghuśaṃvara (aka Cakrasaṃvara, Herukābhidhāna), the core tantra of the Saṃvara cycle, owes at least half its verses to Śaiva sources (primarily the Yoginīsañcāra of the Jayadrathayāmala, the Siddhayogeśvarī, Brahmayāmala, and Tantrasadbhāva)—not in terms of inspiration, but direct copying. He has also discovered evidence of borrowings in a number of other texts. And this borrowing does not center on matters of little consequence. In fact, the entire consecration (abhiṣeka) rite described in Yogottara- and Yoginī-tantras and redacted from them into a standard form in Abhayākaragupta’s Vajrāvalī (1100 ce; the rite is described at Sanderson 1994: 88-92) is modelled in almost every element (ritual actions, mantra forms, role of the maṇḍala) on the standard Śaiva dīkṣā or initiation rite outlined in earlier Śaiva scriptures and described in detail in unpublished paddhatis (paraphrased at length in Sanderson 2002 and 2004U). There is a significant exception in that the Buddhist abhiṣeka as presented in the texts required ritualized sexual congress and consumption of sexual fluids (one’s own and one’s guru’s); this did not occur in the Śaiva rite, except in some forms of Kaula initiation. I propose that in adopting rites from the Śaiva Tantras, the Buddhists, conscious of the need to prove their own versions superior, increased the transgressive element by including it in the initiation rite, thus creating a greater empowerment—for in this ritual world greater transgression, successfully negotiated, yielded greater power.
The Śaiva rite of initiation was said to guarantee liberation (nirvāṇa-dīkṣā) at the moment of death or earlier, while the parallel Bauddha rite was said to guarantee the generation of empowered bodhicitta that inevitably leads to liberation. We know the Śaiva rite was earlier, for Dharmakīrti (prior to 650 ce) attacked the idea that it could grant liberation (SANDERSON 2001, 2006), in addition to the evidence of the textual criticism already alluded to.
Another crucial piece of evidence for the direction of borrowing is provided by Sanderson in his 1994 article that showed clearly that the Saṃvarodaya (of the Saṃvara cycle) borrowed its listing of the 24 sacred pīṭhas from the Śaiva Tantrasadbhāva (1994: 95), a listing which contains no known Buddhist sites but several known Śaiva ones. Scholars such as Davidson (2002: 206-211) have resisted this conclusion, which is odd as the Buddhist tradition itself is aware that this was an attempt to “take over” Śaiva sites. This is seen in the well-known myth of the subjugation of Śiva, studied in its Sarvatathāgata-tattva-saṅgraha version by both Snellgrove and Davidson, and translated in full in an even longer version by Gray (2007a: 30-34) from Indrabhūti’s commentary on the Cakrasaṃvara. Here we are explicitly (and of course polemically) told that Bhairava and Umā dwelt in these 24 sites, along with their hordes of yakṣas, gandharvas, rākṣasas, nāgas and so on: “Having taken control of the three realms, they overpowered the world. In these twenty-four places they [behaved like] worldly ones, acting out their lust and hatred. Wherever they were, Lord [[[Īśvara]]] Bhairava and Umā engaged in love-play, along with their retinue.” The Buddhist forces of Heruka and Vajrayoginī are pictured as conquering them, and then taking on their names and appearances: “The names of those disciplined were attached to the heroes [[[vīras]], a tantric term] and yoginīs who disciplined them.” (Gray 2007a: 33-34) A clearer self-awareness of appropriation we could hardly ask for in this type of literature. But we have yet another piece of evidence of this kind: that of the Śaiva response. This is found in a story narrated by Jayadratha in the 13th cen. Haracaritacintāmaṇi, discussed by both Sanderson (1994) and Granoff (2000). In this story, Bṛhaspati declares his intention to delude the demons by converting them to Buddhism; as they are devout Śaivas, this must be done by creating Buddhist texts and practices that look like Śaiva ones but claim even greater power. The key passage is: “…Once I have established these [[[Buddhist]]] icons [depicted trampling on Śaiva deities] and seen to it that they are widely recognized, I shall add certain mantras culled from the Śaiva Tantras; and lifting various passages out of these scriptures of Śiva I shall fabricate and propagate a system of [[[Buddhist]]] Tantric ritual.” (Sanderson trans., 1994: 93) Thus we have explicit acknowledgment of appropriation from both sides.
Sanderson’s arguments have been questioned and disparaged by Ronald Davidson in a chapter of his 2002 book that betrays significant confusion about the difference between tantric and non-tantric Śaiva sects, a lack of necessary chronological perspective, and ignorance of the contents of the Śaiva canon, among other problems. He agrees that there must have been “sustained [[[Śaiva]]] influence” but questions whether Śaiva texts were the primary sources for the Yoginī Tantras, without arguing for any other sources, other than the uncorroboratable idea that the nebulous figures of the Buddhist siddhas might have been a source. He speculates that the more likely scenario is one of mutual influence, but provides no evidence whatever for the notion. He criticizes Sanderson’s work for a) “chronological difficulties,” a bizarre statement that suggests a lack of careful reading on Davidson’s part, for Sanderson’s work is notable for an almost obsessive attention to valid means of establishing reliable chronology; and b) a “lack of examination of the sources of Śaiva formulations” (203) despite the fact that no such (textual) sources can be shown to exist for the scriptural register of Śaivism, other than a non-specific indebtedness to classical Sāṅkhya and Yoga and certain Upaniṣads. Certainly there is no evidence for the borrowings from Bauddha texts that he imagines “must have” taken place; the instances he cites from Bühnemann’s work are from a much later period than the one in question. Davidson’s “philological” footnote (386: fn105), where he analyzes what he takes to be an error in Sanderson’s textual criticism, is even more absurd, for it indicates either a failure to read the cited source carefully or a surprising lack of awareness of the basic methods of textual criticism. In the same footnote, he accuses Sanderson of a “curious theology of scripture” because of the latter’s putative “privileging” of the Śaiva sources, which is ironic as it is precisely what one could accuse Davidson of, but much more accurately in his case, since he is clearly resisting the conclusions that the evidence obviously suggests when the principle of Occam’s razor is applied. A more complete refutation of Davidson’s claims can be found at Sanderson 2009: 189-92 (with notes 455, 457, and 461), corroborated fully by the analysis in Hatley 2007: 184-87 and partially in Gray 2007b. Note that the latter article and the detailed introduction to Gray’s recent annotated translation of the Cakrasaṃvara Tantra (2007a) demonstrate clearly that later redactions of the text (such as that of commentator Bhavabhaṭṭa) removed some Śaiva elements and added more and more Bauddha ones, taking the text further from its Śaiva/Śākta origins, which were glaringly obvious in the earlier redaction of the text commented on by Jayabhadra. (Gray 2007a: 6-7)
We may incidentally note that the fact of Śaiva influence on the Yoginī Tantras was acknowledged as early as 1828 by H.H. Wilson, who wrote upon examing some manuscripts of these tantras for the first time that “the Buddha creed has been modified by Tāntrika admixture;” we see that for him this means Śaiva/Śākta traditions when he elaborates that “the Śākta form of Hinduism is . . . the chief source of the notions and divinities foreign to Buddhism” (Wedemeyer 2001: 239). This perceptive assessment, corroborated soon after by Burnouf (Wedemeyer 2001: 240), was apparently dismissed in time along with much other work of early “Orientalists” whose major advantage, as Davidson rightly notes (though with infelicitous phrasing, 2002: 13), was a lack of interest in producing a predetermined conclusion from their work, which is hardly the case with some modern Buddhologists.
We should also note that the very title of this corpus of works, the Yoginī Tantras, reflects their origin in Śaiva and Śākta modes of practice, for the yoginīs were central to the latter, as seen in the early Yāmala Tantras, especially the extant Brahmayāmala (aka Picumata; Hatley 2007) and the Jayadrathayāmala (partially critically edited by Sanderson but as yet unpublished). Indeed, the earliest external reference to Śaiva texts that we have, that of Dharmakīrti, refers to them as Ḍākinī- and Bhaginī-tantras (Sanderson 2006: 149). There is a truly a surfeit of evidence for the redaction of the Buddhist Yoginī Tantras from those of the Śaivas; but as we have space for only one more piece of evidence, let it be a powerful one, from within the tradition. David Gray reports that the Chinese biographer Hui Li tells us that when King Harṣa Śīlāditya was travelling in Orissa, he was reproached for his support of Nālandā by “Hīnayāna” monks, who argued that “The Monastery of Nālandā and its ‘sky-flower’ [i.e. absurd] doctrine is not different from the Kāpālika sect.” (Beal 1911: 159, cited in Gray 2007)
Yet one final point remains to be made. The fact of borrowing and appropriation having been established, what is its significance? What does it tell us? Davidson has presented a convincing picture of the political and economic instability of India in the early medieval period, in which competition for royal patronage was absolutely necessary for the survival of a religious institution (Davidson 2002, chs. 2-4). Part of the strategy, then, was the “sacralization of the sociopolitical environment” (114), for example, through the rite of abhiṣeka, which paralleled in certain ways the coronation rite of the king, and through symbolism of the maṇḍala, which is remarkably similar to the image of the king at the center of circles of vassals, exerting his power over his realm. Another strategy, then, was the wholesale appropriation of Śaiva rites and scriptures; this is explained by the enormous success Śaivism exhibited in the winning of royal patronage and its resultant dominance in this period (Sanderson 2005 and 2009). This was the vital information that Davidson was missing in his picture of things in the early medieval period. It is also demonstrated by the similarly pervasive appropriation of Śaiva texts by the Vaiṣṇava Pāñcarātrikas. Of course Davidson is correct in that all these groups were striving to respond to the volatile political environment, and the interreligious borrowing might be viewed in this light as a case of desperate times calling for desperate measures. Yet it strikes me as very problematic that in such cases modern academics tend to attribute to religious actors motives that are solely economic or political, despite the obvious fact that our experience with religious agents in our own period show us that many of them are, indeed, religious—that is, pious, in their own terms. Therefore the failure of academics to attribute religious motives to the agents they study is an intellectual failure (as also noted by Wedemeyer in his review of Davidson, 2006: 374), an overcompensation perhaps to distance our field from theology. In this light I propose that we must consider the other obvious but unstated reason that this interreligious borrowing took place: it must have been the case that some of the borrowers believed there was ritual, and sometimes even spiritual, value to the texts of the other tradition, and believing further that “All that which is well spoken is spoken by the Buddha,” they saw their appropriation as entirely valid and indeed compassionate. The common academic presumption that all such appropriation is part of a political agenda of jockeying for position, and that religious motives cannot be dealt with effectively in our intellectual discourse, shows how far we still have to go to accurately understand the religious mind.
It is my hope that this article goes some small distance towards proving what I firmly believe will inevitably become the consensus view: that in the early medieval period, the notion of a “Hinduism” over and against “Buddhism” is misleading and false.
 Cf. Sanderson, Alexis. "Tolerance, Exclusivity, Inclusivity, and Persecution in Indian Religion During the Early Mediaeval Period." In Honoris Causa: Essays in Honour of Aveek Sarkar , edited with a foreword by John Makinson (Allen Lane, 2015), pp. 155–224.
 Sārdhatriśatikālottara-vṛtti, p. 5.
 See Smith, Imagining Religion, pp. 4-5, citing Robert Sokol and Peter Sneath’s Principles of Numerical Taxonomy. Rodney Needham seems to be another source for the concept. Note that Indologists (e.g. Brooks, Tribe, Lopez) seem to conflate the concept of polythetic definition with Wittgenstein’s “family resemblence“ theory, which Smith does not do.
 With the important caveat that transgressive acts characterize only certain sects.
 The Mañjuśrīmūlakalpa is sometimes cited as a prototypical text of this class, though it is much later than most of the texts labelled Kriyātantras.
 Smith (loc.cit.), following the usage in the taxonomial endeavors of the biological sciences, argues that a thing need have a “large number” of the requisite criteria. I would prefer a majority of them for inclusion in the class.
 Note that here I am dating the beginning of each class of text, not implying that they succeeded one another; in fact, we have Kriyā Tantras from the 12th century as well.
 He sees the earliest evidence for this element in the Sarva-tathāgata-tattva-saṅgraha, which was translated into Chinese in 723 ce; he calls it (loc.cit.) “if not the first certainly one of the earliest and most influential texts in which a tantric way to liberation is taught.“
 As Xuan-zang was a ”keen observer,” Hodge concludes “It is likely that any specifically tantric texts and practices that were already in existence at that time had not yet gained general acceptance in the main centers of Buddhism he visited, such as Nālandā.“ (Op.cit.) Davidson notes that Xuan-zang dismissed “ignorant users of spells [[[dhāraṇīs]]]“ (2002: 118).
 Hodge 2003: 9-10. Yi-jing adds that there is “no way of comprehending them without an oral transmission,“ presumably because the commentaries had not yet been written. He also implies that a permament maṇḍala [[[Wikipedia:Chinese|Chinese]]: tán] had been installed at Nālandā where he stayed (673 ce). [NB: this is also how Gray reads the passage; 2007: 54 fn 212.] If so, the only conclusion possible is that these teachings reached Nālandā between 645 and 673, indicating perhaps a very rapid spread.
 Though there was and is debate on whether the goals of the different yānas are in fact the same. Gellner argues that the three yānas have three different soteriological ideals: arhat, bodhisattva, and siddha. Snellgrove asserts that the Vajrayāna is “as distinguishable from the Mahāyāna as this is distinguishable from the so-called Hīnayāna“ (1987: 129) while Tribe indicates that the Vajrayāna was always connected to, if not actually dependent on, the Mahāyāna (2000: 220-22).
 Utilizing David White’s useful term ’mesocosm’ (White 2000: 9), referring to any of the several ritual items or maps that mediate between the forces of the universe and the practitioner’s (subtle) body: most usually the mesocosm is a maṇḍala or yantra, but may also be an icon, image or (more rarely) even a rosary, sword, or mirror. (Törszök 2003: 191 and Sanderson CHECK)
 Some features that some scholars would like to include in such a definition, such as nonduality and emphasis on the feminine in worship or practice, are disallowed from an all-encompassing definition because they do not characterize a significant pan-Indian school of Tantric Śaivism, that of the early Śaiva Siddhānta.
 The Niśvāsa is connected to the earlier Śaiva textual stratum of the Ātimārga (i.e. the non-tantric Śaivism of the Pāśupatas, Lākulas, etc.), and of course Śaiva texts owe something to earlier Indian literature in general, but not often in terms of specific traceable textual dependence. See Sanderson 2001: 16-17.
 It is of course possible that a large number of Buddhist Tantras were lost, if they were centralized in monastic libraries destroyed by the Muslims; but this is not strongly likely in my view, due to the decentralized and grass-roots nature of much of Tantra.
 But note that the Mantrapīṭha, unlike the Yogottara Tantras, does not require sexual congress within the initiation ceremony or at any other time. I will argue that this is because the Bauddha Tantras, each phase of which followed closely on the heels of their Śaiva counterparts, tried to “up the ante“ and out-do their rivals in transgression and thus in ritual power.
 These parallelisms outlined at Sanderson 1988: 678-9. A detailed exploration of the textual indebtedness of the Yoginītantras to the Śaiva Vidyāpīṭha is to be found at Sanderson 2009: Note that the three-phase periodization of Śaivism I have given here is much simplified and should not be taken as a strict chronology; for example, Śākta and transgressive elements were present from the beginning, in certain groups only, and in various unsystematized ways. But the chronology is valid in terms of prominence if not in terms of advent.
 Where he cites a verse on the guru found in several Bauddha texts, gurur buddho gurur dharmo guruḥ saṅghas tathaiva ca, which in my view seems to be formed on the basis of the famous Hindu verse: gurur brahmā gurur viṣṇur gurur devo maheśvaraḥ.
 I find two early Saiddhāntika Śaiva sources without any effort: the Svāyambhuva-sūtra-saṅgraha Yogapāda, 20.43 teaches: vyomarūpaśivadhyānī vyomamanāḥ sudhīḥ | sarvasaṅgavinirmuktaḥ parāṃ nirvṛtiṃ ṛcchati || “The wise one who meditates on Śiva as the void, his mind absorbed in the void, attains the highest release, completely free of all attachment.“ (The Vijñāna-bhairava, a Śākta source, also associates the highest reality with śūnya in verse 127. Perhaps there is Buddhist influence at work here, but if so, it only reinforces my argument, as I show below.) Another important early source, the Kālajñāna recension of the Kālottara, has (v. 97c): cinmātrapuruṣo jñeyaḥ śivo “Śiva must be understood as a being consisting only of consciousness.“ (Quoting from the critical edition of Dominic Goodall.)
 Torella suggest that Dharmakīrti was more appropriate for this role than Dignāga because the former accepted the legacy of Bhartṛhari, another formative influence on Śaiva discourse. (1992: 338, fn 7)
 In the Nareśvaraparīkṣā-prakāśa ch. 1; see Watson 2006: 87, fn 129.
 Which is not to say that borrowing and appropriation did not happen on the soteriological level (the best-documented example is the creation of the Buddhist Laghuśaṃvara-tantra by “cannibalizing” text-passages from Śaiva scriptures; see Sanderson 2009: CHECK), but when it did, the appropriating redactors sought (sometimes clumsily) to conceal the source of their material.
 See Abhayākaragupta’s Niṣpannayogāvalī, which includes Gaṇeśa, Kārttikeya, and the Dikpālas in the periphery of its deity circuit (Bühnemann 1999: 303), and the Dharmadhātu-vāgīśvara-maṇḍala, a Yogatantra which includes in the fourth circuit of its maṇḍala the Vaidika dikpālas, the Mothers, the planets, and sectarian deities such as Viṣṇu and Śiva (Vira & Chandra 1995: 70-75).
 See Sanderson 2004: 231; Sanderson, 14th Gonda Lecture, 24 November 2006 (summary available at alexissanderson.com); and most importantly the monumental “The Śaiva Age: an Explanation of the Rise and Dominance of Śaivism During the Early Medieval Period,” in The Genesis and Development of Early Tantrism.
 Confusingly, unlike Buddhism Śaivism distinguishes laukika from Vaidika, but both terms refer to the Vedic domain: the first refers to the common-or-garden Vaidika brāhmin of the type who does pañcāyātana-pūjā, and second refers to Vaidikas who are primarily devoted to Śiva and are pursuing liberation.
 Since these texts generally date from the 15th century and later, the term ‘Hindu’ is appropriate.
 This may be seen as an example of what Agehānanda Bharati called the ‘Pizza Effect’: the exportation of a cultural element, its alteration by the borrower, followed by its re-importation in the altered form by the original lender. More research would be needed to identify what has been altered in this case other than the name.
 This word Sferra (and others) translates as ‘heretics’. But its literal meaning is ‘pilgrims,’ as well as ‘rule-followers,’ and refers simply to non-Buddhist laukikas, though certainly with a derogatory connotation.
 iha lokasaṃvṛtyā vicāryamāṇaḥ sarvadarśanasiddhāntaḥ samāno laukikasiddhaye , Yoginīsañcāra 11.2ab
 See the Kiraṇavṛtti ad 6.11-12 and the Mataṅgavṛtti VP p. 98.
 See Śivadharma chapter 1.
 A handout from the lecture, given at the International Conference on Esoteric Buddhist Studies in September 2006, showing parallel passages indicating redaction from the Śaiva to the Bauddha, is available on his website.
 But note that careful hermeneutical exegesis of the scriptural sources soon resulted in the sexual ritual being performed mentally or not at all. (Tribe 234-5) Certainly these same initiations when performed today (e.g. by the Newars) do not have a sexual component, despite drawing on very old paddhatis.
 Though he uses the word Kāpālika (2002: 202), not realizing that that word as a sectarian term designates the non-tantric sect also known as the Somasiddhānta. The kāpālika aesthetic was deeply influential on certain branches of Tantric Śaivism (e.g. the Yāmala and the Kālīkula), but there was no Tantric sect with that name. This error is not surprising, however, for these relationships were only clarified by Sanderson in the last six or seven years. Indeed, Davidson exhibits some confusion in his endnote (386: 105) when he observes that Sanderson does not use the term “Kāpālika“ in his 2001 article.
 Sanderson has shown (2001) that the Bhaginī-tantras refers to the Vāma-tantras of Tumburu and his four sisters (bhaginī), and Hatley (2007) and Gray (2005) have shown that at this period, the only Buddhist usage of the word ḍākinīs is entirely opprobrious, grouping them with piśācas and the like.
 Indeed, the term abhiṣeka was first used in that context. The Tantric abhiṣeka modelled an esoteric religious rite on the royal one, and then extended those rites to the king himself, as part of a patronage-winning strategy (Sanderson 2005).
 Addressing the reasons for Buddhist appropriation, Sanderson writes: “In their view they were not succumbing passively to an alien influence. Fully conscious that they were assimilating the dominant Śākta Śaiva idiom of the region, they justified their doing so as a means of converting non-Buddhists, taking their practices and encoding them with Buddhist meaning so that outsiders could rise effortlessly through what was familiar to them to what would save them, a view exactly reflected in Jayadratha’s myth of the compilation of anti-Śaiva iconography, Śākta Śaiva liturgy, Mantras, and Buddhist doctrine as a means of luring devout Śaivas away from their faith.” (2009: 233)
 Sanderson (2001, fn 50): “The ritual systems taught in the Śaiva and Pāñcarātrika Saṃhitās resemble each other so closely in morphology and syntax that they have the appearance of two dialects of a single ’Tantric‘ language.“ As he has shown, this is due mostly to borrowing; see Table 3.
 Adhyāśayasañcodana-sūtra: yat kiñcin maitreya subhāṣitaṃ sarvaṃ tadbuddhabhāṣitam, where of course subhāṣitam is carefully defined as inspired, meaningful, principled, bringing about extinction of afflictions, and so on. Cited at Lopez 1996: 30.
Goodall, Dominic and Harunaga Isaacson. “Workshop on the Niśvāsatattvasaṃhitā: the Earliest Surviving Śaiva Tantra?” in Newsletter of the NGMCP 3 (Jan-Feb 2007): 4-6.
Hodge, Stephen. 2003. The Mahā-vairocana-abhisaṃbodhi-tantra. London: Routledge Curzon.
Ruegg, David Seyfort. 2001. “A Note on the Relationship between Buddhist and ‘Hindu‘ Divinities in Buddhist Literature and Iconology: The Laukika/Lokottara Contrast and the Notion of an Indian ‘Religious Substratum‘“. In: Le Parole e i Marmi: Studi in onore di Raneiro Gnoli, ed . Raffaele Torella. Rome: Istituto Italiano per l’Africa e l‘Oriente.
Sanderson, Alexis. 1988. “Śaivism and the Tantric Traditions.” In The World's Religions, edited by S. Sutherland, L. Houlden, P. Clarke and F. Hardy. London: Routledge and Kegan Paul, pp. 660-704. Reprinted in The World's Religions: The Religions of Asia, edited by F. Hardy. London: Routledge and Kegan Paul (1990), pp. 128-72.
. 2001. “History through Textual Criticism in the study of Śaivism, the Pañcarātra and the Buddhist Yoginītantras.” In: Les Sources et le temps. Sources and Time: A Colloquium, Pondicherry, 11-13 January 1997, edited by François Grimal. Publications du département d'Indologie 91. Pondicherry: Institut Français de Pondichéry/École Française d'Extrême-Orient.
. 2005. “ A Commentary on the Opening Verses of the Tantrasāra of Abhinavagupta. ” In Sāmarasya: Studies in Indian Arts, Philosophy, and Interreligious Dialogue in Honour of Bettina Bäumer, ed. Sadananda Das and Ernst Fürlinger. New Delhi: D.K. Printworld.
. 2007. “The Śaiva Exegesis of Kashmir.” In: Mélanges tantriques à la mémoire d’Hélène Brunner / Tantric Studies in Memory of Hélène Brunner, edited by Dominic Goodall and André Padoux, Pondicherry: Institut français d'Indologie / École française d’Extrême-Orient, 2007. Collection Indologie 106, pp. 231–442 and (bibliography) pp. 551–582.
Sferra, Francisco. 2003.“Some Considerations on the Relationship Between Hindu and Buddhist Tantras” in Buddhist Asia 1: Papers from the First Conference of Buddhist Studies Held in Naples in May 2001. Giovanni Verardi and Silvio Vita, eds. Kyoto: Italian School of East Asian Studies.
. 1994. The Īśvarapratyabhijñā-kārikā of Utpaladeva with the Author ’s Vṛtti. Critical edition and annotated translation (Serie Orientale Roma, LXXI). Roma: Istituto Italiano per il Medio ed Estremo Oriente.
Tőrzsők, Judith. 2003.“Icons of Inclusivism.” In Mandalas and Yantras in the Hindu Traditions by G. Bühnemann, with contributions by H. Brunner, M.W. Meister, A. Padoux, M. Rastelli and J. Tőrzsők. Leiden: E.J. Brill.