Sutra Journal Logo

A curated journal on art, culture and dharma

The Śaiva Religion and its Philosophy in Context
Part one

by Christopher Wallis May, 2016

Shiva Kailash Temple

A Shiva panel from the Kailash temple (Cave 16), Ellora. QuartierLatin1968


Modern scholarship now frequently acknowledges that the concept “Hinduism” was constructed by Western ‘Orientalist’ [1] scholarship, and that there was no such indigenous term until the nineteenth century.[2] Yet such acknowledgment is often merely lip-service, for the term continues to be used to describe the entire complex and diverse Indian religious milieu, other than Buddhism and Jainism, throughout the common era. Though there is utility to an etic (non-indigenous) term that scholars define clearly and use specifically, this is not such a term, and it remains true that we have not yet unpacked the manifold implications of the fact that as a premodern religious (as opposed to cultural) designator, “Hinduism” is fundamentally a fiction, there being no single religious practice or belief that unites all putative “Hindus” in the premodern period. It is a fiction that has obscured our ability to discern clearly the complex interrelationships of the competing religions of the volatile and productive medieval period. An examination of the primary sources for this period shows us five primary religious traditions, each seeing themselves as a distinct and complete path to liberation as well as a vital cultural institution competing for patronage: those of the Vaidikas, Śaivas (including the Śāktas), Vaiṣṇavas, Bauddhas, and Jainas. [3]

The purpose of this essay is twofold: in Part One, I seek to identify Shaivism as a distinct religious institution in the early medieval period (600-1100 ce) that certainly did not see itself as part of a “Hinduism,” since that concept didn’t exist in this period; but neither did it see itself as part of Brahmanism, the tradition that is the antecedent to Hinduism. In Parts Two and Three, I seek to delineate the doctrines that made Shaivism distinct from other medieval Indian philosophies (with which it was of course interrelated, just as Judaism, Christianity, and Islam are interrelated), that we readers may have a clearer picture of the Śaiva religious-cum-philosophical textual landscape of the period. In this article I draw heavily on the work of Alexis Sanderson (and his students), work that culminated in a book-length 2009 study demonstrating the dominance of Shaivism as a distinct religious institution throughout South and Southeast Asia in this period.[4]

Shiva as Lord of Dance and Music

Shiva as Lord of Dance and Music, 11th century India:
Orissa, 1000-1099 Sandstone 45 x 17 x 10 in.
Carved in deep relief, the figure of Shiva is
depicted dancing on a pyramid of five heads with
sunken eyes placed upon a lotus base.
Five heads or skulls play an important
role in esoteric tantric practices.

Part I: Evidence for Shaivism as a single religion in the early mediaeval period

Before addressing what distinguished Śaiva thought within the broader sphere of Indian intellectual activity, we must define Shaivism as an entity. We may begin by briefly examining evidence that Shaivism was a single self-contained religion and thought of itself as such. By “religion” I mean “an institution consisting of culturally patterned interactions with culturally postulated superhuman beings” (Melford Spiro’s famous 1966 definition), to which I would add the clause that religion is also an institution that offers its adherents either salvation or heavenly rewards or both. In the Indian milieu, where religious traditions often are not clearly demarcated, we must also ask what circumscribes and differentiates one religious tradition from another. I see these five criteria as operative (in order of importance): 1) a body of texts that belong to that system and no other; 2) authoritative teachers consecrated in that system and no other; 3) the fact that the system makes an effort to distinguish itself from others; 4) competition with other religious systems, such as the claim to offer definitive salvation above and beyond them; and 5) the belief in a founder, usually conceived as historical, and unique to that system. Shaivism possessed all of these in the period in question.

To take the last first, Shaivism as a distinct initiatory religion (as opposed to a non-initiatory devotional tradition) did have its putative historical founders, though their historicity is heavily obscured by hagiography and myths. The earliest instantiation of sectarian monotheistic Shaivism for which we have certain evidence is that which became known as the Atimārga (‘Higher Path’), more commonly designated by one of its three primary branches: the Pāśupatas, Kālamukhas, and Kāpālikas. All three are explicitly part of one tradition, though, putatively founded by one Lakulīśa, thought to be an avatāra of Śiva, who is said to have descended in a cremation ground and animated the body of a deceased brāhmin to reveal his teaching to the world.[5] That Lakulīśa was possibly a historical figure (though the name was invented later) can been seen in the accounts of his religious instruction to three or four named primary disciples, at least one of which (Kuśika) began a lineage that is documented for ten generations.[6] The second and substantially larger instantiation of initiatory Śaiva religion (śiva-dharma[7]) called itself the Mantramārga (known to scholars as Tantric or Āgamic Shaivism), putatively founded by one Śrīkaṇṭha Nātha, also identified as an incarnation of Śiva. His teachings were the five “streams” (srotas) of the Śaiva scriptural corpus.[8] As has now been well documented, these Mantramārgic scriptures offered detailed cosmologies and rituals unique to Shaivism. Foremost amongst the latter was the unprecedented (apūrva) claim to offer a salvific initiation (the nirvāṇa-dīkā)—which could be conferred only by a consecrated Śaiva guru temporarily or permanently identified with Śiva—that guaranteed liberation at or before the time of one’s bodily death.[9] Further, these scriptures taught unique theological doctrines (e.g., the three malas, five Acts of God [pañca-kṛtya], five kañcukas, 36 tattvas, and so on), and argued that only liberation achieved through Śiva’s scriptural teachings is true and final liberation. More specifically, Śaiva theologians had an inclusivist view which taught that the other Indian religions granted some salvation from bondage and suffering, but argued that their soteriological goals do not reach to the highest levels of reality,[10] i.e. the Pure Universe (śuddhādhvan) where only God exists,[11] which is attained only through Shaivism, they argued. [12] Thus Shaivism fulfills the first and fourth criteria listed above.

To address the third criterion: Buddhism and Jainism are generally regarded as separate religions from Brahmanism/Hinduism specifically because they explicitly reject the spiritual authority of the Veda and are therefore labelled as heterodox (nāstika). [13] But precisely such heterodoxy belongs to Shaivism in varying degrees (especially in its Śākta dimension). Here is the evidence for that assertion. First, Śaiva sources of the early mediaeval period argue that whenever Śaiva and Vedic injunctions conflict, the Śaiva one supersedes the Vedic. Śākta-Śaiva exegete Abhinavagupta (c. 1000) writes in the Tantrasāra (ch. 4), “The capacity for annulment belongs to the Śaiva scriptural injunctions alone, as established by reason and by countless scriptures.” [14] Even the more conservative Śaiva scriptures argue that Vedic injunctions are valid only in the sphere of civic religion, and that the Veda has no soteriological value to the initiated Śaiva. [15] Though most texts do recommend that the initiate maintain his Vedic religious duties, these must be understood by him as being done purely for the sake of appearances, to uphold the established social order. In fact, we are told, if one makes the mistake of believing that it is the Vedic observances in combination with the Śaiva that have religious value, then neither will bestow their respective benefits, for a hybrid practice (śabala-karma) is said to be fruitless. [16] Abhinavagupta goes further in asserting that the Vedas are not only soteriologically irrelevant, but in fact “drag down [into a lower birth] those whose minds are deluded“ (TĀ 37), i.e., those who believe the Vedas will liberate them. Indeed, it is said in the scriptures that when Śiva’s blessing descends on a ordinary Vaidika, it takes the form of the realization that his Vedic practice is inadequate and thus leads him to seek a Śaiva Guru.[17]


Shiva Mahadeva in 8th-century Kashmiri sculpture.
Cleveland Museum of Art. 1989.369

The objection might be raised here that I am engaged in constructing a category just as artificial as that of Hinduism, and that instead of a single Śaiva religion I should see merely a collection of interrelated sects and cults. Yet the Śaivas in the early mediaeval period themselves believed in the existence of a coherent Śaiva religion, for we have evidence that they viewed the members of the other Śaiva sects as co-religionists, no matter how far removed they were in doctrine or practice, and did not view Vaiṣṇavas and others as co-religionists. We see evidence of this in both scriptural and exegetical writings; for example, in his Tantrasāra, Abhinavagupta quotes the Pārameśvara-tantra, which says, “All these Vaiṣṇavas and others are stained by limited knowledge and attachment. They do not discover the highest reality, being deprived of the wisdom of the Omniscient One,” to which he adds the comment, “Thus they do not look to the higher View (darśana), and therefore, they are simply averse to correct discernment, true scriptures (sad-āgama), and the instruction of true Gurus. It is only those who are pierced by an intense and stable Descent of Power ( śaktipāta) from Śiva, [and who subsequently] refine and purify their conceptual understanding through true scriptures and [a true Guru], who [finally] enter their own ultimate nature.”

An important piece of evidence that Śaivas saw their religion as coherent as well as distinct has been already pointed out by Sanderson: that of the rule of supplementation outlined by Jayaratha in his Tantrāloka-viveka.[18] This is the rule that whatever detail is missing from one’s primary source-text (mūla-tantra) is to be supplemented from other texts within the Śaiva canon, even if the proper performance of a single ritual act thus requires a combination of information from multiple texts (four, in the example), the collation of which crosses several sectarian and doctrinal boundaries but remains within the Śaiva canon. Sanderson argues that the entire Śaiva canon of many hundreds of scriptural texts “is seen as a single complex utterance” (2005: 23). Thus our first criterion is yet further strengthened.

For further evidence that Śaivas thought of themselves as constituting a distinct religion, we may briefly mention the rite of li goddhāra. This is a special ritual designed for those who wish to convert to Shaivism from another dharma; it is designed to remove one’s former sectarian marks (the visible and invisible defining features of one’s religious identity) so that one may be initiated as a Śaiva. Anyone performing a soteriologically-oriented practice in another tradition prior to coming to Shaivism had to undergo this rite in the period we are discussing; note that this was just as true for a Vedic sannyāsin or vānaprastha as for a Buddhist or Vaiṣṇava. [19]

Finally, clinching evidence is found in the fact that the Vaidikas also confirmed Shaivism’s status as a separate religion: an inferior one to be repudiated. Texts such as Medhātithi’s Manusmti-bhāya condemned Śaivas as well as Jainas and others as “outside the Veda” (veda-bāhya) and therefore having no religious value.[20] The Vaidika Purāṇas labelled Śaiva and Vaiṣṇava Tantric texts as ’scriptures of delusion‘ (moha-śāstra). [21] The seventh-century exponent of Vedic orthodoxy, Kumārila, even argued that some Śaivas were further from the norms of the Veda than Buddhists and Jains, due to their impure and foreign/barbaric (mleccha) practices such as eating from a skull-bowl.[22] The famous upholder of Brāhminicaldharma Aparāditya cautioned the true Vaidika against adopting Shaivism, and presents a detailed argument against Śaiva doctrines. [23]

Next month: Part Two: Shaivism’s Sources and Influences


[1] In this context, the term references the famous 1978 book Orientalism by Edward Said.

[2] See, e.g., Gauri Viswanathan, ‘Colonialism and the Construction of Hinduism’, in Gavin A. Flood, ed., The Blackwell Companion to Hinduism, Oxford, 2003; Will Sweetman’s short essay on the subject available from the University of London’s School of Oriental and African Studies website southasianstudies/keywords/; and the entry s.v. “Hinduism” in the Harper Collins Dictionary of Religion, Jonathan Z. Smith, ed. While no one disputes that the term Hinduism was coined in the nineteenth century, still under dispute is the question of which period the term ‘Hindu’ may be appropriately used. On this point we have David Lorenzen’s important 1999 article “Who Invented Hinduism?“ in Comparative Studies in Society and History, 41.4. This piece shows that “Hindu“ as a term of self-designation came into use around 1400, after the period I consider here. I have no argument with those who use the term in reference to the last six centuries. Lorenzen’s citations further show that it was a term originally used exclusively in the context of contrasting indigenous Indian religion over and against Islam. He does not show, for example, that the concept “Hindu“ in this context and period did not include Jainas (and of course Buddhism had disappeared from India by this time). Nor is it true that the term sanātana-dharma was used in the classical period in the same sense that ‘Hinduism’ is used today; for example, in the Bhagavad-gītā, Kṛṣṇa uses the term to refer to his teaching specifically, and none of the Sanskrit commentators take it as referring to Śaiva and Śākta dharmas as well as Vaiṣṇava and/or Vedāntin ones.

[3] These are the indigenous terms for the followers of each of these religions. As Sanderson has pointed out, the derivation of these Sanskrit words is in the sense of “one who follows the teachings of deity X,” e.g. śivaproktaśāsanam > śaivam > śaiva . Note that the adherents of these religions did not choose to create an abstract term by suffixing ‘–ism‘ (-tva, -tā), though such suffixes were common. We do, however, see the term śivadharma in the sense of “Śiva’s religion,” i.e. Shaivism.

[4] "The Śaiva Age: The Rise and Dominance of Shaivism during the Early Medieval Period," by Alexis Sanderson, in: Genesis and Development of Tantrism, edited by Shingo Einoo, University of Tokyo, 2009. This piece also treats the relation of Shaivism to the other four major Indian religions mentioned above. Though I have not been able to incorporate the findings of Sanderson’s massive study into the present article, its arguments only strengthen the conclusions I have reached here.

[5] See the account in the original (6th cen.) recension of the Skandapurāṇa (critically edited by Isaacson and Bakker as well as by Bhaṭṭarāī), discussed in Bisschop 2006.

[6] I refer to the 356 ce inscription recording a land grant of a Pāśupata priest for the support of a shrine found in Ramesh and Tewari 1990: 4-6, as well as other Pāśupata inscriptions in Valkhā with dates corresponding to 369, 374, and 375 ce (ibid: 6-7, 10, 20; cf. Sanderson 2009: 50). This is the earliest epigraphical evidence for Atimārgic Shaivism.

[7] Some reading this article will wish to argue that the word dharma does not mean ‘religion’, to which I respond that this extremely multivalent term derives its specific meaning from context, and that there is ample evidence to suggest that it can indeed mean religion as I have defined it here. We have space to detail only a small portion of this evidence in the present article.

[8] Including many hundreds of texts, far too much material for one person to compose; but the attribution Śrīkaṇṭha is nonetheless given, e.g., at Tantrāloka 37.13-17. Note the Saiddhāntika, Trika, and Krama sampradāyas all were ascribed historical founders (see Sanderson 1995) and the Kaula phase of Shaivism (or perhaps it should be called Śāktism) had its own founder, possibly historical: Macchanda, the Fisherman, later Sanskritized as Matsyendra Nātha (see Sanderson 2007: 264, n91).

[9] There are countless references in the primary sources to this; for the simplest and most direct see Mālinīvijayottara-tantra 4.8: muktiś ca śivadīkayā, ‘Liberation is bestowed by Śiva’s initiation.’

[10] Abhinavagupta’s Tantrasāra, introduction: “Only those scriptures uttered by Parameśvara [Śiva] are (completely) authoritative. [They are authoritative] due to their judicious acceptance of [some] doctrines taught in other scriptures, and due to their teaching [additional realities] established by superior reasoning to that of those doctrines. Therefore, the knowledge taught in other sacred texts (āgamas) liberates one from bondage, [but] only to a certain extent, not from all of it. However, the scriptures of Parameśvara do liberate one from all of it.“ (śāstraca parameśvara-bhāitam eva pramāam. apara-śāstroktānām arthānātatra vaiviktyena abhyupagamāt tadarthātirikta-yukti-siddha-nirūpaāc ca tena aparāgamoktajñāna tāvata eva bandhāt vimocakam na sarvasmāt sarvasmāt tu vimocakaparameśvara-śāstra).

[11] I argue that the translation ‘God’ for Parameśvara or even Śiva is appropriate in monotheistic contexts; Shaivism was always either monotheistic or monistic, in the sense that if other gods were acknowledged, they were regarded as emanations or servants of Maheśvara, who is One ( eka eva), deputized by him to carry out their specific functions (see, e.g., Abhinavagupta’s Tantrasāra, ch. 11).

[12] See Sanderson 2009: 301, n715 for a short list of key passages on this topic. There are also rare exceptions in the canon, such that as the theologian Kṣemarāja, who in his Pratyabhijñā-hṛdaya argues that some Vedāntins, and those who follow the spiritual philosophy of the Grammarian Bhartṛhari, attain to the Pure Universe.

[13] This entailing of course the claim that all sects labeled “Hindu” may be so called because they share a common denominator, that of deriving their teachings, deities, or at least authority from the Veda (hence, astika). With reference to the medieval period, this is hardly the case, as we will see further on. It would be more accurate to say that nearly all of the groups now labeled Hindu at least payed lip service to the authority of the Veda. It was this that eventually enabled their assimilation into the Hinduism of the modern period; the exceptions were entirely weeded out. See the article by Sanderson, serialized in Sutra Journal March through May of this year.

[14] śivacodanāyā eva bādhakatvaṃ yuktisiddha m …anantāgamasiddhaṃ .

[15] See, e.g., the Matagapārameśvara-tantra CP 2.2-8b.

[16] See the Mataga, loc. cit. and the Sarvajñānottara pp. 97-98 (IFP MS T334, available on the Muktabodha Digital Library website); see also Tantrāloka 4.249-51 and –viveka thereon. These references were provided to me by Prof. Sanderson. Note here that in Shaivism, the way in which one thinks of the meaning and purpose of ritual action influences the outcome (possibly due to Buddhist influence), as opposed to the mechanistic ritual view of the (Mīmāṃsaka) Vaidikas. Note also that it is just such a Vaidika/Śaiva hybridization condemned in the early texts that one does in fact see in the late medieval period (1200-1500), presaging the constitution of modern Hinduism.

[17] Mata ga-pārameśvara, Vidyāpāda pp. 56 and 98.

[18] TĀV ad 4.251cd, cited and discussed at Sanderson 2005: 23-24.

[19] See Tantrāloka 22.42-48. Someone in the Vaidika sphere of the brahmācārya or ghastha āśramas did not undergo the ligoddhāra, but this was because such people were not involved in the practice of Brāhmanism for soteriological reasons, but belonged only to the sphere of civic religion.

[20] See Manusmti-bhāya vol. 1, p. 57 and Āgama-prāmāya p. 26; these references (and the following) were furnished to me by Professor Sanderson.

[21] E.g., Kūrma Purāa (1.16.119-20).

[22] See Tantravārttika, vol. 1, p. 94 (note that this was in the 7th century, before Tantric Buddhism took up these transgressive practices as well) and Sanderson’s article in the current issue of Sutra Journal.

[23] Yājñavalkyasm ti- īkā ad 1.7. Note that this is in the early 12th century, the height of widespread Śaiva influence in the political and religious spheres. Should we consider Aparāditya’s arguments a desperate plea from the (temporarily) losing side?

by Christopher Wallis

May, 2016

Get the One and Only Sutra Journal Newsletter

Sutra Journal Logo

© 2016 All Rights Reserved by Sutra Journal and Respective Authors.

Sutra Journal has readers from 170 countries.