Śaivism after the Śaiva Age: Continuities in the Scriptural Corpus of the Vīramāheśvaras

Abstract

This article makes the case that Vīraśaivism emerged in direct textual continuity with the tantric traditions of the Śaiva Age. In academic practice up through the present day, the study of Śaivism, through Sanskrit sources, and bhakti Hinduism, through the vernacular, are generally treated as distinct disciplines and objects of study. As a result, Vīraśaivism has yet to be systematically approached through a philological analysis of its precursors from earlier Śaiva traditions. With this aim in mind, I begin by documenting for the first time that a thirteenth-century Sanskrit work of what I have called the Vīramāheśvara textual corpus, the Somanāthabhāṣya or Vīramāheśvarācārasāroddhārabhāṣya, was most likely authored by Pālkurikĕ Somanātha, best known for his vernacular Telugu Vīraśaiva literature. Second, I outline the indebtedness of the early Sanskrit and Telugu Vīramāheśvara corpus to a popular work of early lay Śaivism, the Śivadharmaśāstra, with particular attention to the concepts of the jaṅgama and the iṣṭaliṅga. That the Vīramāheśvaras borrowed many of their formative concepts and practices directly from the Śivadharmaśāstra and other works of the Śaiva Age, I argue, belies the common assumption that Vīraśaivism originated as a social and religious revolution.

1. Vīraśaivism, Tantra, and the Śaiva Age

By the mid-thirteenth century, Śaivism in the Deccan had already been irrevocably transformed by the decline of the Śaiva Age, as Alexis Sanderson has called it, the golden age of what we colloquially describe as “tantric Śaivism” (Sanderson 2009). Perhaps most remarkably, the Śaiva institutions that had previously dominated the region’s religious ecology were rapidly disappearing, particularly those of the Kālamukhas. Descending from the Lākula traditions, or what Alexis Sanderson has termed Atimārga II, the Kālamukhas left behind precious few of the scriptures that must have originally distinguished their practice from competitors within the Atimārga and Mantramārga, and none in full recensions. Nevertheless, even before Alexis Sanderson and his students had revolutionized our narrative of medieval Śaivism over the past two to three decades, the Kālamukhas were already known to have vanished abruptly, as their landholdings were systematically replaced by another Śaiva tradition rising to prominence in the region, the Vīraśaivas. As a field, we owe our original awareness of this phenomenon to the pathbreaking work of David Lorenzen, who in his monograph, The Kāpālikas and Kālāmukhas: Two Lost Śaivite Sects, compiled a voluminous array of inscriptional evidence to document how Kālamukha maṭhas (monasteries) ceased to be patronized precisely as inscriptions increasingly attested to the presence of Vīraśaiva devotional figures at the same sites.1 Reflecting further on this state of affairs, however, Lorenzen later added an appendix to his work, claiming that Kālamukha maṭhas were not merely displaced, but rather were overthrown by a veritable religious revolution. As Lorenzen writes:

It can even be said the two [Kālamukha and Vīraśaiva] movements represent antipodes of Indian intellectual and religious tradition[:] the Brahmanic and the anti-Brahmanic, the scholastic and the devotional, Sanskrit learning and vernacular poetic inspiration, pan-Indian culture and regional culture, social and spiritual hierarchy and social and spiritual equality…. Vīraśaivism represented not “a reformist schism of the Kālāmukha church” but rather its overthrow.

(Lorenzen [1972] 1991, p. 242)

For Lorenzen, in essence, Vīraśaivism is the quintessential representative of the Bhakti Movement: a fundamentally anti-brahmin, anti-caste “movement”, a radical rupture of social protest, and a purely vernacular religion of the people.2 Lorenzen is not alone, of course, in attributing these features to Vīraśaivism. To the contrary, in the wake of A. K. Ramanujan’s celebrated Speaking of Śiva (Ramanujan 1973), the field of South Asian religions has naturalized his portrayal of Vīraśaivism as a social and religious revolution. Ramanujan, in turn, imported the perspectives of earlier intellectuals writing in Kannada who emplotted Vīraśaivism quite explicitly as an Indian foil for the Protestant Reformation.3 But does this narrative accurately capture the influences that precipitated the emergence of Vīraśaivism? If we depict Vīraśaivism as essentially a devotional (bhakti) revolution, for instance, we might be inclined to delineate the Śaivism after the Śaiva Age as something radically different from its predecessors, those traditions that fall under the category of “Śaiva tantra”. Indeed, most scholarly monographs and articles on Vīraśaivism scarcely mention the word “tantra”, and historicize Vīraśaivism only in relation to other communities traditionally categorized as “bhakti”, as if an unbridgeable chasm separated the two.4 Likewise, even leading scholars of Śaiva philology flag the “movement of the non-brahmin Vīraśaivas” (Sanderson 2012–2013, p. 83) as of interest to what we might call Tantric Studies only for its occasional borrowings from the Śaiva Siddhānta and the Trika of Kashmir.

Yet, if we read this antagonism back into the origins of Vīraśaivism as a moment of rupture, we risk putting forward a thesis that—as I would like to argue as explicitly as possible—is completely in contradiction with our textual evidence. To put matters even more plainly, based on philological evidence, Vīraśaivism did not originate as a revolution or reformation of tantric Śaivism, nor of Kālamukha traditions in particular. Indeed, a large part of the problem facing earlier generations of scholars was that adequate textual evidence had not yet come to our attention. Only a fraction of early Vīraśaiva literature has been studied to date, in part because we have restricted the source languages of our archive to the vernacular, exclusive of Sanskrit, and in part because we lacked sufficient knowledge of what had come before. Of course, print editions of such Vīraśaiva works in Sanskrit did exist, as Vīraśaiva monasteries published a substantial quantity of the tradition’s literary history in the late-nineteenth and twentieth centuries. For institutional reasons, however, scholars trained in early Kannada and Telugu literature have rarely consulted Sanskrit texts, and when they did so, they previously lacked sufficient knowledge of the pre-Vīraśaiva traditions of the Śaiva Age from the region to draw clear connections between the two.5 Likewise, and perhaps more crucially for the present audience, when Śaivism is studied from a philological perspective, vernacular literature is rarely consulted, and in this case, as I will argue, the contemporary Telugu textual context is indispensable for historicizing the early Vīraśaiva works in question.6

In this article, I will make the case that Vīraśaivism emerged in direct textual continuity with the “tantric” traditions of the Śaiva Age, especially the Atimārga II of the Kālamukhas, although in a number of cases early Vīraśaivism was influenced by Mantramārga traditions as well. As an embryonic version of this article was originally presented at the Society for Tantric Studies Conference in 2019, I present evidence that specifically sheds new light on how we define and periodize what we call tantra, but a similar corrective must be taken in our broader narratives of Hindu and South Asian religious traditions as well. In an earlier article in the journal History of Religions (Fisher 2019), I introduced elements of my claim by delineating the canon of what I have called the Vīramāheśvara textual culture of Srisailam. As I demonstrated in that publication, we have access to a rich body of early Vīraśaiva didactic literature that I date to the thirteenth and fourteenth centuries, contemporary with our earliest vernacular evidence for Vīraśaivism and heavily indebted to the textual canons of the Śaiva Age, most notably (but by no means limited to) the Śivadharmaśāstra. Whereas in that context I dealt with issues of historiography facing Religious Studies and South Asian Studies, my project here is primarily philological. Naturally, much work remains to be done in critically editing this textual corpus, and tracing parallels in the citations of many of the otherwise rarely attested early recensions of prior Śaiva scriptures. As a result, the evidence presented here will be extended in subsequent publications on the ritual practice and textual canons of the Vīramāheśvaras.

With such an aim in mind, I will reiterate in greater philological detail the case for dating the Vīramāheśvara corpus to around the thirteenth to early-fourteenth centuries, a significantly earlier date than that of the Sanskrit Vīraśaiva works of Vijayanagara. The principle Sanskrit works in question are the Vīramāheśvarācārasāroddhārabhāṣya, otherwise traditionally known as the Somanāthabhāṣya, the authorship of which I will discuss below; the Śaivaratnākara of Jyotirnātha; and the Vīramāheśvarācārasaṅgraha of Nīlakaṇṭha Nāganātha. Each of these Vīramāheśvara texts, in turn, contains citations from earlier (some likely Kālamukha) Śaiva scriptures, which in many cases match quite closely, barring the usual accretion of textual variants. The contemporary Telugu corpus consists primarily of the Telugu works attributed to Pālkurikĕ Somanātha: namely, the Basavapurāṇamu, Paṇḍitārādhyacaritramu, and Caturvēdasāramu.7 The Śivatattvasāramu attributed to Mallikārjuna Paṇḍitārādhya merits consideration here as well, although most likely dates to a slightly earlier period (a twelfth century dating would be plausible).

First, by bringing these two bodies of textuality into dialogue, I present the evidence that the Sanskrit Somanāthabhāṣya has been correctly attributed to the same Pālkurikĕ Somanātha who is responsible for the three Telugu works mentioned above. As a result, as both text–internal citational evidence and the attribution of authorship to Pālkurikĕ Somanātha are consistent with each other, we can assert with relatively strong confidence that the Somanāthabhāṣya was composed in the thirteenth century at Srisailam. The fact that the Sanskrit and Telugu works in question overlap so pervasively in tone and content, moreover, further allows us to reject the hypothesis, entrenched as it is, that early vernacular Vīraśaivism arose in strict opposition to Sanskritic Śaivism. Second, I will conclude by outlining the principle points of continuity between the Vīramāheśvara corpus and the Śaivism of the Śaiva Age, demonstrating that on textual grounds early Vīraśaivism was directly indebted to its predecessors in the Deccan, and did not constitute an “overthrow” of its legacy, nor a revolution of any kind. While the Vīramāheśvaras drew on a number of distinct textual currents from the Śaiva Age, I focus here on their substantial inheritance from the Śivadharmaśāstra, with particular attention to the concept of the jaṅgama, the human devotee as moving śivaliṅga.

2. Dating the Vīramāheśvara Corpus: Pālkurikĕ Somanātha and the Authorship of the Somanāthabhāṣya

As I have argued at greater length in other venues (Fisher 2019), the tradition we now define as Vīraśaivism, or Liṅgāyatism, was not a new religious movement founded by the poet-saint Basava in the twelfth century.8 Our earliest texts that mention Basava and his exploits—the Telugu (and Sanskrit) works of Pālkurikĕ Somanātha, and the Kannada Ragaḷĕgaḷu of Harihara—can only be dated as early as the thirteenth century, and moreover speak to a wider discursive world that pre-existed Basava himself, in which he merely participated as one historical agent among many.9 Indeed, both Harihara and Somanātha, from opposite sides of the Deccan, speak to a remarkably similar religious worldview, both depicting, for example, the historical śaraṇas or Vīraśaiva saints as incarnations of Śiva’s celestial attendants, the Pramathagaṇas.10 Inscriptional evidence confirms, moreover, that Vīramāheśvara terminology was used prior to and far afield from the city of Kalyāṇa where Basava served as daṇḍanāyaka to Bijjala of the Kalachuris. In other words, we have no plausible historical grounds for situating a singular religious revolution in twelfth-century Kalyāṇa. Rather, the Vīraśaivas in residence there during Basava’s day were already part of a greater trans-Deccan network spanning from southern Maharashtra through coastal Andhra and, if we trust inscriptional evidence, likely penetrating further south into Tamil Nadu as well.11

Nevertheless, although the Vīramāheśvaras may well have traversed an extensive geographical network by the thirteenth century, our surviving Sanskrit textual evidence from the period stems from one single location: the extended domain of the Śaiva pilgrimage site at Srisailam. While we might hypothesize that these texts circulated beyond their locale of composition, whether or not similar texts were composed elsewhere, we can assert with confidence that Srisailam was something of a discursive epicenter, so to speak, in which the thirteenth-century Vīramāheśvaras codified their doctrine and ritual practice. How, then, do we know that the texts I have identified above are Vīramāheśvara works composed at Srisailam at a relatively early date? First of all, as I have discussed in greater length in (Fisher 2019), the texts generally declare their location of composition and religious affiliation fairly explicitly. In the Śaivaratnākara, Jyotirnātha traces his family lineage’s origin to Saurashtra, apparently prior to the demolition of the Somanātha temple by Mahmud of Ghazni. He continues, in the same context, to describe the temple that he and his family had maintained after relocating to Srisailam. In the Vīramāheśvarācāra-saṅgraha, Nīlakaṇṭha Nāganātha pays homage to Mallikārjuna, the form of Śiva at the temple at Srisailam, and proceeds to venerate a number of early Vīraśaiva figures writing in Sanskrit or south Indian vernaculars, none of whom can be dated, based on our evidence, after the thirteenth century. In both of these works, as well as in the Somanāthabhāṣya, the words Vīramāheśvara and Vīraśaiva appear as terms of self-reference to the community in question. All three are structured primarily as nibandhas (compendiums, or anthologies) of Sanskrit scriptural citations, while the Somanāthabhāṣya also elaborates on the verses cited with extended prose commentary. Incidentally, the Telugu works of Pālkurikĕ Somanātha also contain all of these features, incorporating the self-referential term “Vīramāheśvara”, extended descriptions of the Śaiva institutions of thirteenth-century Srisailam, and, as we will see, lengthy anthologized passages of Sanskrit citations.

Who, then, is Pālkurikĕ Somanātha, and why would his dual authorship of works in Telugu and Sanskrit be so significant for our scholarly portrait of Vīraśaivism? Scholars of bhakti traditions of Hinduism will be intimately acquainted with Pālkurikĕ Somanātha for the hagiographies of the early Vīraśaiva saints or śaraṇas he crafts in his vernacular Telugu works. From the perspective of Telugu literary historians, Somanātha’s verse style stands in stark contrast to the school of high Telugu literature that more strictly emulated the idiom of Sanskrit kāvya.12 In short, his writings are marshaled in support of a view that the vernacular in South Asia emerged from the popular religious sentiment of devotion, rather than from the elite courtly world of Sanskrit literature. Based on the portrait of Somanātha’s writings as vernacular hagiography, his works—like those of his near contemporary writing in Kannada, Harihara—have been read almost exclusively in dialogue with the lives of the Nāyaṉārs as recounted in the Tamil Pĕriyapurāṇam. Indeed, such parallels do exist. But, as we will see, by reducing Pālkurikĕ Somanātha’s discursive context exclusively to the Pĕriyapurāṇam, scholars to date have lost sight of the data that allows us to contextualize more precisely the Śaiva worldview from which he wrote.

Among works attributed to him, Somanātha is best known for the Basavapurāṇamu, which narrates not only the life story of Basava, as the name would suggest, but also numerous of his purported contemporaries. The Basavapurāṇamu has been adopted as a principle source for classroom teaching and scholarship on the Vīraśaiva tradition because it can be accessed easily by English speakers through the translation of Velcheru Narayana Rao and Gene Roghair. This Telugu epic in dvipada meter has often been upheld in scholarship as an example of purely vernacular, devotional narrative—disconnected, in other words, from anything remotely Sanskritic and from tantra as a category.13 Pālkurikĕ Somanātha is also generally accepted as having composed the Paṇḍitārādhyacaritramu, a second Telugu prabandha on the life of Mallikārjuna Paṇḍitārādhya, to whom authorship of the Telugu-language Śivatattvasāramu is attributed.14 Indeed, we can be fairly confident that the same author crafted both of these two Telugu works, and in fact, a third as well: Somanātha tells us explicitly at the outset of his Paṇḍitārādhyacaritramu that he had previously completed two Telugu works entitled the Basavapurāṇamu and the Caturvēdasāramu, or “Essence of the Four Vedas”. Speaking about himself in the second person, Pālkurikĕ Somanātha declares the following:

You admirably composed the Basavapurāṇa;

In the Basavapurāṇa narrative, you recounted as history (itihāsa)

The stories of the Gaṇas, those celebrated ancient devotees.

You composed the Caturvēdasāramu with the

   best of heroic devotion (vīrabhakti) in accordance with the Vedas.15

Although Somanātha has professed his own authorship here of the Caturvēdasāramu, its title might give some readers pause: the “Essence of the Four Vedas”, some might suspect, is the polar opposite, at least according to conventional wisdom, of what motivated Vīraśaivism as a religious “movement”. Yet, not only can we infer, pending further examination, that Pālkurikĕ Somanātha did author the Caturvēdasāramu, but we must acknowledge his self-professed motive in doing so: Somanātha authored this vernacular work of Śaiva doctrine, he tells us, to establish the orthodox Vaidika status of what he understands as vīrabhakti. Indeed, this sentiment accords precisely with the view articulated in the Basavapurāṇamu, where we read that devotion to Śiva is inculcated in the Vedas themselves: “O Basava, proclaim the devotion that has been derived from the essence of the Vedas and śāstras”.16 Moreover, we have little reason to suspect that Somanātha’s genuflection to the Vedas was intended disingenuously, or as a means of coopting a textual authority he viewed as foreign to Śaivism. To the contrary, by the thirteenth century in the Andhra country, it would have been quite normative among Śaivas to interpret the Vedas as a quintessentially Śaiva scriptural corpus, in no way contradictory with the Āgamic and Atimārgic literature of the Śaiva Age. For instance, the name of Somanātha’s “Caturvēdasāramu” was by no means unprecedented. Rather, it was likely intended to evoke the earlier Caturvedatātparyasaṅgraha of Haradatta, a garland of Sanskrit verses in the vasantatilaka meter intended to illustrate that Śiva is the essential meaning (tātparya) of the four Vedas, cited frequently in the Somanāthabhāṣya.17 Evidently, we cannot casually presume that the Vīraśaivism of Pālkurikĕ Somanātha intends in any manner to upend the authority of the Vedas as scripture.

Further, we would be remiss in presuming that for Somanātha the vernacular Telugu was in any way divorced from Sanskrit. Contrary to popular perception, his linguistic register is highly Sanskritized, even preserving the sort of lengthy Sanskrit compounds generally taken to be the purview of courtly Telugu literature. For example, to indicate his distaste for interacting with non-Śaivas, Somanātha describes himself in the Basavapurāṇamu with extended Sanskrit compounding as “avoiding contact such as dialogue with and respect for non-Śaivas” (bhavijanasamādaraṇasaṃbhāṣaṇādisaṃsargadūraguṁḍa), and encapsulates his reverence for Vedic canons of textuality in phrases such as “in accordance with all the Vedas and Purāṇas, and the established doctrine of the secret of the stainless liṅga” (akalaṃkaliṃgarahasyasiddhāṃtasakalavēdapurāṇasammataṃbaina) (Basavapurāṇamu p. 7). Moreover, all of Somanātha’s vernacular works are interlaced with direct Sanskrit quotations from Vedic and Śaiva source material. Both the Paṇḍitārādhaycaritramu and Caturvēdasāramu are heavily inflected with long doctrinal digests of Sanskrit source material, as will be discussed below, but Sanskrit citations appear in the Basavapurāṇamu as well. Unfortunately, these quotations are not necessarily apparent to those reading Rao and Roghair’s translation, as the English rendering and footnotes may obscure the shift in language.18

Stylistically, in other words, all of Pālkurikĕ Somanātha’s works give every indication of an authorial imagination well versed in the Sanskrit language. Yet, we can find even more conclusive evidence of shared authorship by directly comparing key passages from the Caturvēdasāramu, Paṇḍitārādhyacaritramu, and the Sanskrit Somanāthabhāṣya that contain direct and unmistakable parallels. In fact, despite linguistic differences, the texts harmonize to a remarkable degree, such that the overlap in content is far too significant to be explained by coincidence. To begin with a particularly striking example, let us examine the maṅgala verse of the Somanāthabhāṣya, which invokes Basava simultaneously as a human incarnation of Śiva’s bull, and leader of the Pramathagaṇas, Vṛṣabha or Nandikeśvara:

May Lord (rājaḥ) Basava surpass all, venerable (pūjaḥ) for his fortitude and

   stainlessness,

The seed (bījaḥ) of shining devotion, keeping the company (samājaḥ) of the

   Pramathagaṇas

Abiding (vartī) within an expansive lineage that removes the affliction (ārti) of the

   humble,

His limitless fame (kīrti) established across the directions, incarnation (mūrti) of the

   Lord of Bulls.19

In both halves of this benedictory verse, Somanātha employs a four-part rhyme scheme of a sort that is rarely encountered in Sanskrit literature but is not at all unexpected in Telugu dvipada verse. In fact, not only do Pālkurikĕ Somanātha’s Telugu works make ample use of this device throughout, but Somanātha is particularly fond of the second rhyming pattern, often making use of the very same rhyming words. To name a single example, the Paṇḍitārādhyacaritramu also opens with an invocation of Basava, incarnation of Vṛṣabha, as the one “who had accumulated fame (kīrti) and merit through the form (mūrti) of the auspicious guru, dwelling (varti) in bliss, pulsating (sphūrti) with the end of scripture”. The content of the verses may differ, but the rhymes are unambiguous parallels. Moreover, the precise same rhyming words appear on multiple occasions in the Basavapurāṇamu as well.20

With the evidence presented thus far, it may remain plausible to suggest that the Somanāthabhāṣya was simply invoking the literary fashions of the day, imitating either Pālkurikĕ Somanātha directly or the broader conventions of early Telugu prosody. Nevertheless, the overlapping content is far more pervasive, including some particularly striking doctrinal passages reproduced in both the Sanskrit Somanāthabhāṣya and the Telugu Caturvēdasāramu. For instance, both texts include an enumeration of a closely matched set of Upaniṣadic scriptures, which both texts refer to as “Śākhā Upaniṣads”, the property of distinct lineages (śākhā) of Vedic transmission.21 It is worth noting that the term śākhopaniṣad itself is not especially common. By employing this term, Somanātha might be taken as revealing that for him, Vedic scripture was not an abstract canon but was embedded within a living sociology of distinct Vedic brahminical communities. Furthermore, both works supply identical proof texts for the incarnation of Vṛṣabha as Basava: “I will become your son, by the name of Nandin, not born from a human womb”.22

Perhaps the most remarkable of these convergences is that the Somanāthabhāṣya and Caturvēdasāramu provide a precisely identical Prakrit etymology of the name of Basava, which makes use of identical grammatical rules and examples from vernacular Telugu usage. Drawing on the Prakrit grammar of Vararuci,23 Somanātha makes the case that the name Basava can be derived systematically from the Sanskrit Vṛṣabha (“bull”), as rules of substitution render the letters b and v interchangeable (vṛkārasya bakārādeśo bhavati, vabayor abheda iti; pavargatṛtīyākṣaramu bakāramu pakāraṃbuvalanaṃ), and the sibilants of various classes are notoriously collapsed into “sa” in Prakrit and several vernaculars (śaṣoḥ sa iti sūtrāt ṣakārasya sakārādeśo bhavati; śaṣoḥ sa yanu vyākaraṇasūtramunaṃ). Hence, “vṛ” can become “ba”, “ṣa” can become “sa”, and “bha” can become “va”, transforming the Sanskrit Vṛṣabha into Basava. Similarly, one may demonstrate Basava’s ontological connection with Śiva, by deriving in a similar manner the name Basava from the first three syllables of Śiva’s name as Paśupati, “lord of beasts” (paśupa). Both texts proceed, then, to illustrate this phonetic transformation with identical examples, such as the Sanskrit word kuṭhāra, meaning an axe, and the Telugu equivalent, guddali, (kuṭhārakuddālatāmarasādipadeṣu…; guddālatāmarasakuṭhāramul’ varusa guddaliyuṁ dāmarayu gŏḍali).24

It is undeniable, at this point, that the Somanāthabhāṣya and Caturvēdasāramu share some direct relation of dependence, but could one text have been written in direct imitation of the other? For multiple reasons, forgery seems implausible. For instance, the Somanāthabhāṣya makes no effort to stake out a reputation for itself through attribution to Pālkurikĕ Somanātha. In fact, the author’s name is mentioned nowhere in the text. Despite their substantial intertextuality, moreover, the two texts are not precise matches: that is, neither the Somanāthabhāṣya nor the Caturvēdasāramu seems intended as a translation of the other. While concerned with several identical themes—for example, both deal with the obligatory Vīramāheśvara topics of sacred ash (vibhūti), rudrākṣa beads, and the bearing of the personal liṅga—the structure of the texts is not identical. Moreover, while a substantial number of the Sanskrit citations in the Caturvēdasāramu also appear in the Somanāthabhāṣya, an equally substantial number do not, and vice versa. As a result, neither text would have been sufficient to provide the source material for the other.

If anything, Somanātha’s Paṇḍitārādhaycaritramu overlaps even more pervasively with the contents of the Somanāthabhāṣya, even if the overlapping content is not so readily memorable. Structured as a garland of narratives of the lives of Mallikārjuna Paṇḍitārādhya and other saints, the Paṇḍitārādhycaritramu has, like the Basavapurāṇamu, been represented as a strictly vernacular and prototypically devotional bhakti literary work. While very little work has been done on the text within the Western academy, it is best known for its occasional polyglossic use of multiple vernaculars (Kannada, Marathi, and Tamil), and secondarily for the chapter in which Paṇḍitārādhya journeys to meet Basava, only to find that the latter has fled Kalyāṇa after the assassination of Bijjala. Anything more than a surface level perusal of the text, however, makes clear that the author aimed to convey Vīraśaiva doctrine as much as narrative, and was as thoroughly acquainted with Sanskrit as with Telugu. For instance, in the first prakaraṇa, we find several extended doctrinal discussions, structured as garlands of Sanskrit citations within a Telugu grammatical medium.

Many of these discussions, moreover, are structurally parallel to sections of the Somanāthabhāṣya, and the verse quotations often run in almost the same sequence in both texts. It is the “almost” here, again, that is key: in most cases, we find just enough variation between the two—a verse missing or an extra citation supplied here or there in either text, or different attributions of sources for the same citation—to be confident that one text could not have been simply copied from the other. The doctrinal digests in the Paṇḍitārādhyacaritramu concern the “greatness” (Telugu: mahima, Sanskrit: māhātmya) of sacred ash (vibhūti), rudrākṣa beads, pādodaka, the worship of the liṅga (liṅgārcana), the bearing of the liṅga, and prasāda, all of which are discussed in the Somanāthabhāṣya as well. The sum total of the evidence is abundant, and only a fraction can be published here for want of space. I have, however, exemplified this citational pattern below in Appendix A, with the original Telugu and Sanskrit of a parallel section from the Paṇḍitārādhyacaritramu and the Somanāthabhāṣya.

What, then, do we make of these pervasive textual parallels in multiple languages? By far the most parsimonious solution—which I believe to be the strongest argument, based on the evidence—is quite simply that all of these works were composed by the same author. In order to confirm the plausibility of dating the Somanāthabhāṣya to the thirteenth century, however, we must further clarify that no textual material contained within the work precludes such a dating. The same, incidentally, must be ascertained for the Śaivaratnākara of Jyotirnātha and the Vīramāheśvarācārasaṅgraha of Nīlakaṇṭha Nāganātha, the other two works I associate with the early Vīramāheśvara corpus. In short, none of the Sanskrit Vīramāheśvara works cite any source texts that would prohibit dating the Somanāthabhāṣya and Śaivaratnākara to the thirteenth century, and the Vīramāheśvarācārasaṅgraha to the early fourteenth century.25 Among readily datable Sanskrit sources in the Somanāthabhāṣya, we find citations from the Somaśambhupaddhati of 1048/9 CE, the ca. eleventh-century Vāyavīyasaṃhitā, and the mid twelfth-century Sūtasaṃhitā.26 In the Śaivaratnākara, we further find an intriguing mention of the fourfold typology of yoga, which Jason Birch has recently historicized to this time period.27 It is worth noting that none of the Vīramāheśvara authors cite the Sanskrit Śrīśailakhaṇḍa, which Reddy (2014) has proposed to date to the thirteenth century on stylistic grounds.28 Also worthy of note is that while these works are intimately familiar with the Śaiva religious landscape at Srisailam, none makes mention of Mallikārjuna’s consort as Bhramarāmbā, who seems to make her debut on the stage of Telugu literature around the turn of the fifteenth century.29

Some confusion may be generated by the fact that the Śaivaratnākara and Vīramāheśvarā-cārasaṅgraha cite a text by the name of Kriyāsāra (“Essence of Rituals”), a title that is most famously associated with a Vijayanagara-period (perhaps fifteenth- or sixteenth-century) ritual compendium hybridized with a Śaktiviśiṣṭādvaita commentary on the Brahmasūtras.30 Not only does the Kriyāsāra as cited by the Vīramāheśvaras, in contrast, contain no discernibly Vīraśaiva or Vedāntic content, but citations attributed to that name fail to match the Vijayanagara text.31 Succinctly, the Kriyāsāra in question is an entirely different work. In fact, none of the Vīramāheśvara texts in question contain Śaktiviśiṣṭādvaita content, and generally invoke the term Vedānta exclusively as a reference to the Upaniṣads. That these three works—the Somanāthabhāṣya, Śaivaratnākara, and Vīramāheśvarācārasaṅgraha—are an interconnected corpus of textuality, moreover, is underscored by the fact that they share a common repertoire of citational texts, a number of which are rarely cited under the same names in other domains of Sanskrit intellectual history, some to my knowledge never otherwise identified in any source to date.

Among the shared scriptural canon of the Vīramāheśvaras, the most foundational and frequently cited source texts include the Śivadharmaśāstra, Vātulatantra, Śivarahasya, Liṅga Purāṇa, and, in the Śaivaratnākara and Vīramāheśvarācārasaṅgraha, the Vīratantra. It must be noted carefully that the recensions of the Vātula and Vīratantra cited are distinct from Āgamic works commonly cited in the Vijayanagara period. Text names that are never mentioned within the Vīramāheśvara corpus, but that are ubiquitous in Vijayanagar period compositions, include the Vātulottara, Vātulaśuddhākhya, and Vīrāgamottara.32 In addition to these theologically significant works, Vīramāheśvara authors share a pattern of citing a number of less widely circulating works, including: the Īśānasaṃhitā; the non-Vijayanagara Kriyāsāra; the Kriyātilaka; the Kālikākhaṇḍa (presumably of the Skanda Purāṇa); the Brahmagītā; the Bhīmāgama; the Mānava Purāṇa; and the Liṅgasāra.33 Outside of the Vīramāheśvara corpus, one of the texts’ closest discursive neighbors seems to be the Śāradātilaka, sharing a number of these sources.34 Although I cannot possibly document all of the voluminous points of textual overlap in this article, including numerous shared citations, suffice it to say that the intertextuality between the Sanskrit Vīramāheśvara works is so strong as to be patently obvious when the works are subjected to a close comparative analysis.

I would caution, however, that there are a number of works attributed to Pālkurikĕ Somanātha that I have not included in this study, and in some cases, I currently harbor significant doubts that Somanātha could have composed them.35 Two of the latter are worth discussing more explicitly, because their content deviates significantly from the discursive norms across languages of the “Vīramāheśvara moment”. Most notable among these is the Anubhavasāramu, a fourth major Telugu work often attributed to Pālkurikĕ Somanātha. The category of anubhava (experience) is already heavily thematized in Vīraśaiva circles by this time in the western Deccan, but is more typically invoked in early Marathi literature than in Telugu. Anubhava does not appear as a technical doctrinal term in the Somanāthabhāṣya, Śaivaratnākara, or Vīramāheśvarācārasaṅgraha. The work currently printed as the Caturvēdasāramu, likewise, requires further explication. While I believe the beginning of this print edition to be the work by that name of Pālkurikĕ Somanātha, as I have argued above, the second half of this publication consists of a Telugu work structured as an elaboration of Māyīdeva’s Anubhavasūtra.36 We find numerous instances of terminology here, as in the Anubhavasūtra, that is highly atypical of Vīramāheśvara thought: for example, caitanya, unmeṣa, terminology from Māyīdeva’s ontology, such as paramātmaliṅga, bhāvaliṅga, and so forth. Both works rely heavily on the ṣaṭsthala system, which only begins to make a brief appearance by the time of the Vīramāheśvarācārasaṅgraha.

This conclusion, then—that Pālkurikĕ Somanātha is the author of the Sanskrit Somanāthabhāṣya—bears significant ramifications for how we as scholars ought to historicize the genres of South Asian religious discourse and practice that we call bhakti and tantra. Indeed, beyond the scope of what can be covered in the present article, the Vīramāheśvaras shared with the Śaiva Age distinctive elements of its ritual culture, which are generally not comprised within our academic definitions of bhakti traditions. Such is the case, for instance, with formal tantric rituals of initiation; while this evidence will be discussed elsewhere, it is worth noting for the moment that in both the Basavapurāṇamu and Caturvēdasāramu, Somanātha refers to Vīramāheśvara initiation as śāmbhavadīkṣā.37 In this light, to accept the Somanāthabhāṣya as composed by the very same Pālkurikĕ Somanātha who authored the Telugu Basavapurāṇamu is to cast fundamental doubt on whether the vernacular of Telugu devotional literature ever existed in isolation from contemporary Sanskrit discourse. In turn, we need to acknowledge that, in the eastern Deccan especially, the tantric Śaivism of the Śaiva Age was not overthrown by Śaiva devotional movements. Rather, it exerted a formative influence on the emergence of the Śaiva communities we classify as bhakti traditions. While these points of continuity are too abundant to enumerate in the present article, I would like to continue by looking closely at one key element that the Vīramāheśvaras had inherited from their predecessors of the Śaiva Age: the role of the liṅga, both the personal iṣṭaliṅga and the jaṅgama, the moving liṅga, as living Śaiva devotee.

3. Before the Vīramāheśvaras: Antecedents from the Śivadharmaśāstra

Centuries before the coalescence of the Vīramāheśvara tradition around the thirteenth century, numerous Śaiva lineages had already carved out an institutional domain at Srisailam. These religious networks spanned not only the central mountain peak, on which the Mallikārjuna Temple is located, but also the wilderness terrain in which it is embedded. Indeed, well before the rise of the Vīramāheśvaras, numerous religious communities, Śaiva and otherwise, had established monasteries throughout the extended sacred geography of the “auspicious mountain”. In the thirteenth century, for instance, Srisailam was home to the regional branch of the Golaki Maṭha of the Śaiva Siddhāntins,38 who held a dominant share in the transregional pilgrimage site, negotiating periodic alliances with the Kalachuri, Cōḻa, and Kākatīya kingdoms (Inden et al. 2000). The eastern Deccan, especially around Srisailam, was also well known for housing Kālamukha lineages of the Simha Pariṣad, who appear often in the inscriptional record. Śākta transmissions of the Kālī Krama and the Paścimāmnāya were also in evidence.39 Beyond the Śaiva and Śākta-Śaiva fold, Srisailam also fostered a shared Buddhist-Śaiva transmission of yogic practices; indeed, some of our richest understudied textual resources for the early development of Haṭha Yoga are in the vernacular languages of the Deccan, especially Marathi and Telugu.40 Given the intense interest it generated across Śaiva communities, it is no surprise that Srisailam was the site at which the surviving Sanskrit Vīramāheśvara canon was first articulated. To the contrary, it is precisely the legacy of the Śaiva Age that made Vīraśaivism as we know it possible.

In the preceding discussion, space has only permitted us to scratch the surface of the textual canons that Pālkurikĕ Somanātha adapted in composing his Sanskrit and Telugu oeuvre. For example, he evidently felt no qualms about supplying material from the Śaiva Siddhānta where convenient.41 Other texts cited by Somanātha, as we have seen, seem to have circulated within a more limited domain, possibly only within the extended coastal region of Andhra and through Orissa. Yet, Somanātha also inherits a far deeper legacy than his more temporally proximate Śaiva sources, such as the Somaśambhupaddhati or Sūtasaṃhitā. Most notably, we find a number of citations in the Somanāthabhāṣya from the Śivadharmaśāstra, perhaps the single most authoritative source for lay Śaiva samaya conduct dating back to the sixth or early seventh century (see for example Bisschop (2018) on the dating of the Śivadharmaśāstra). Indeed, many of those features of Vīraśaivism that scholars have viewed as “revolutionary” and “vernacular”, including caste blindness among initiates, emotional or affective bhakti, reciting the stories of Śaiva saints, and the worship of the jaṅgama, or Vīraśaiva saint, as a moving liṅga, were not at all new to the twelfth or thirteenth centuries, but can be directly traced back to the Śivadharmaśāstra itself.42 Other features of the Śivadharma, although less well known within the academic study of bhakti traditions, were equally foundational to the Vīramāheśvaras, including the belief that Śaiva saints were not at all natural or material (prākṛta) human beings but were rather incarnations of Śiva’s gaṇas on earth.43 The original Śivadharma was couched in the form of a conversation between the sage Sanatkumāra and Nandikeśvara, the latter of whom, equated with the bull gaṇa Vṛṣabha, was later understood to be incarnated as Basava himself. We also find, throughout the Śivadharma, frequent usage of the term śivayogin (Kan. śivayogi) as a religious identity marker, which as Gil Ben-Herut has shown was employed abundantly within early Vīraśaiva literature in Kannada.

Succinctly, the Śivadharma was no minor influence on the Vīramāheśvaras. By now, that Śivadharmaśāstra citations appear within the Vīramāheśvara corpus is somewhat of an established fact rather than a new finding; I have already discussed this myself, for example, in Fisher (2019). Concerning the history of the Śivadharma, research has been well underway for some years aimed at producing a critical edition of the text itself and tracing the outsized influence of the scripture on the history of popular Śaivism. One particularly noteworthy example, in the present context, is the ongoing work of Florinda De Simini on the transmission of the Śivadharma and Śivadharmottara within vernacular currents of south Indian discourse. It may well be the case that the abundant Śivadharma citations preserved in the Somanāthabhāṣya can be of use in reconstructing the earlier history of what has often proved to be an unruly and heterogenous textual transmission. As this work is being conducted elsewhere, my project is not primarily to address the textual history of the Śivadharma itself. My project is, however, both in the present article and within my larger book project, to clarify something that has to date escaped scholarship on the history of Vīraśaivism. Specifically, on historical and philological grounds, we can demonstrate conclusively that early Vīraśaivism—including the Somanāthabhāṣya in particular—was constituted directly from the scriptural and cultural heritage of the Śaiva Age, not least among which is the Śivadharma. To make this case requires that I document, as I have begun to do in this article and in Fisher (2019), the distinctive religious sensibilities that early Vīraśaivas directly inherited from the Śivadharma and other earlier Śaiva sources.

In all likelihood, the Vīramāheśvara exegetes, and their predecessors, possessed far more than a casual acquaintance with the text of the Śivadharmaśāstra. Indeed, the Somanāthabhāṣya incorporates textual extracts from the Śivadharma significantly in excess of the verses attributed by name to that text in our available manuscripts. For instance, in one passage variously described as “Vīramāheśvaramāhātmya” or “Vīraśaivācāra”, a significant portion of the anuṣṭubh passage consists of silent borrowings from the Śivadharma. This was not, then, simply a matter of searching for an authoritative proof-text readily at hand. Moreover, the fact that the Vīramāheśvaras were thinking systematically with the Śivadharma is illustrated by the fact that we can observe what seems to be textual drift, possibly deliberate, in the verses of the Śivadharma themselves. While further manuscript work is needed to confirm this point, we meet with some intriguing Śivadharma citations in both the Somanāthabhāṣya and the Śaivaratnākara that speak either to deliberate redaction of the text or spurious attributions. These verses, moreover, do not appear in the most widely attested recensions of the Śivadharma.44

• One should always bear the nirmālya out of devotion; one should not bear it
     out of greed.
• It is called nirmālya because it is stainless (nirmala). One with an impure body
     should not bear it.
• One should bear the nirmālya on the head, and one should also consume the naivedya.
• Having drunk the prasāda water, one obtains gaṇa-hood.45

Both of these two verses concern the subject of nirmālya, the leftover offerings of food, flower garlands, etc. from the worship of Śiva. By the thirteenth century, nirmālya had become a topic of contention within Śaiva discourse across lineages, with the Śaiva Siddhānta even taking deliberate pains to declare Śiva’s nirmālya as impure, requiring the ritual intervention of shrines to Caṇḍeśvara to purify its contamination.46 Nevertheless, following in the spirit of the earlier precedent set by the Pāśupatas, the Vīramāheśvaras took a strong stance on the matter by not only declaring nirmālya as inherently pure, but requiring that initiates offer all food to their personal śivaliṅgas before consumption such that it would become nirmālya. In contrast, the text we now associate with the most common recension of the Śivadharma does not provide any scriptural support for this practice. It is perhaps no surprise, then, that the redactors of the Vīramāheśvara canon would wish the Śivadharma to speak more forcefully in support of their position on the matter—and this is precisely what we find in the texts. In a similar vein, it is worth noting one additional verse attributed to the Śivadharma by both the Somanāthabhāṣya and the Śaivaratnākara, but this time with one crucial variant. The Somanāthabhāṣya reads: “One must not go to a place in which Śiva is not, where there are none of Śiva’s people (nāsti māheśvaro janaḥ)”.47 The Śaivaratnākara, on the other hand, preserves this variant: “One must not go to a place in which Śiva is not, where there are no Vīramāheśvaras” (vīramāheśvaro janaḥ).48 The fact that the phrase “nāsti māheśvaraḥ” appears to have been replaced in the Śaivaratnākara by “vīramāheśvaraḥ”, a less desirable reading, suggests that the verse was modified either intentionally, or through textual drift within the community, to employ the community’s term of self-reference, Vīramāheśvara. We do not, to clarify, have any evidence that the term Vīramāheśvara was employed in the original Śivadharma.49

It is abundantly clear, then, that the Somanāthabhāṣya and the Vīramāheśvara corpus were substantially indebted to the Śivadharma, and that they invoked—and possibly redacted—the Śivadharma to underpin the authority of their fledgling Śaiva community. What may be less well established, by this point, is the fact that Somanātha was no pioneer in his invocation of the Śivadharmaśāstra within the thirteenth-century Vīramāheśvara community. Rather, the Śivadharmaśāstra was already foundational to the incipient ethos of the Vīraśaivas, or Vīramāheśvaras, even before the community was known by either of those names. Rather, we can illustrate the continuous influence of the Śivadharmaśāstra on the emergent Vīraśaiva community by looking more closely at a predecessor to Pālkurikĕ Somanātha’s works, namely, the Telugu Śivatattvasāramu of Mallikārjuna Paṇḍitārādhya. As of yet remarkably understudied for its contributions to Vīraśaiva thought, the Śivatattvasāramu is, like Somanātha’s Telugu works, internally bilingual, even preserving direct citations from the Śivadharmaśāstra embedded in its Telugu verses. These citations, as well as paraphrased content, allow us to isolate certain elements of the Śivadharma’s worldview that were already prominent in the proto-Vīramāheśvara community before the time of Pālkurikĕ Somanātha.

The Śivatattvasāramu is a Telugu Śaiva verse work of which only 489 verses are currently thought to survive. What do we know, first of all, about Paṇḍitārādhya, purported author of the Śivatattvasāramu? Aside from being the subject about whom Pālkurikĕ Somanātha wrote the Paṇḍitārādhyacaritramu, the name Paṇḍitārādhya appears rather prolifically in the inscriptional record from the twelfth century onward. All things considered, Paṇḍitārādhya can be presumed, to the best of our evidence, to have been a historical personage and perhaps a late contemporary of Basava. Remembered as a native of Draksharama near Guntur in East Godavari District, Paṇḍitārādhya appears based on inscriptions to have been active in the Srisailam region in the late twelfth century.50 We know little for sure about what Paṇḍitārādhya’s doctrinal affiliation may have been, although Somanātha describes him as having studied under a certain Kōṭipalli Ārādhyadēva. As for his authorship of the Śivatattvasāramu, although we have no other substantial works attributed to him to compare, the author of the Śivatattvasāramu names himself as “Mallikārjuna Paṇḍita” within the text itself.51 As scholars of Telugu literature have noted for some time, we also find a few direct citations of the Śivatattvasāramu within the Paṇḍitārādhyacaritramu, making it plausible to believe that the person Somanātha revered in this text was indeed the author of the Śivatattvasāramu (Lalitamba 1975, p. 40, ftn. 25).

A fair amount of ink has been spilled by scholars of Telugu literature questioning whether Paṇḍitārādhya was in fact a “Vīraśaiva”, as the Śivatattvasāramu nowhere mentions the names Vīraśaiva, Vīramāheśvara, or Liṅgāyat. Unfortunately, most of these debates have fixated on the question of whether or not the Śivatattvasāramu prescribes the bearing of the iṣṭaliṅga, as the term itself, and the related prāṇaliṅga, are also nowhere mentioned.52 Although these two later concerns did become integral to Pālkurikĕ Somanātha’s theology, the fixation on these two points within Telugu language scholarship has obscured the substantial doctrinal homologies between the Śivatattvasāramu and Pālkurikĕ Somanātha’s works. The Śivatattvasāramu contains, for instance, a lengthy section in praise of Śiva’s Pramathagaṇas, associating them as later Vīramāheśvara authors do with the narrative of the destruction of Dakṣa’s sacrifice.53 Bhakti as a religious value is celebrated at great length; indeed, we even find references to a number of the Śaiva saints whose stories Somanātha would later narrate in the Basavapurāṇamu and Paṇḍitārādhyacaritramu. The Śivatattvasāramu is equally insistent that caste distinctions must be totally prohibited among Śaiva initiates. Moreover, we even find noticeably proto-Vīraśaiva language, such as an invocation of the term jaṅgama. In short, the substantial points of overlap all have roots in the popular lay theology of the Śivadharma.

In the surviving portion of the Śivatattvasāramu, there are seven verses with direct—although deliberately fragmentary—quotations from the Śivadharmaśāstra, making it the most frequently cited Sanskrit work within the Telugu text. I have reproduced below in Appendix B all seven of these citations. Indeed, in some cases, knowledge of the original Sanskrit from the Śivadharma allows us to emend textual corruptions in the Telugu that the editors appear not to have noted. One Śivadharma verse, for instance, that appears to loom particularly large in Mallikārjuna Paṇḍitārādhya’s imagination is the famous comparison between a dog cooker (śvapacaḥ) and a caturvedī brahmin, which asserts that commensality must be respected between Śaiva devotees, regardless of their caste origin: “Neither a Caturvedī nor a dog cooker who is my devotee is more dear to me. He may be given to, and taken from, and is to be worshipped as I am myself”.54 This Śivadharma verse apparently warrants enough attention that Paṇḍitārādhya weaves portions of this Sanskrit citation through a series of three verses in Telugu. In the process, Paṇḍitārādhya reveals that he is well aware of the lengthy history of anti-caste rhetoric within the Śaiva corpus; the necessity of erasing caste distinction among Śaiva initiates, for him, is clearly no “revolution”, but rather an established point of doctrine.

To the best of our knowledge, then, it appears that Śivadharma vs. 1.36 conveys a fairly unambiguous literal meaning that was greeted favorably, and not undermined, by its interpretive communities. In other cases, what certain terms may have meant to an ideal reader of the Śivadharma in the sixth century is far less clear, and we would be wise to pause before reading back their Vīraśaiva meaning, iconic as it may be today, into the original scripture itself. For instance, Paṇḍitārādhya dwells over an extended series of Telugu verses on the concept of the jaṅgama, or “moving” śivaliṅga, which by the time of the nascent Vīraśaiva traditions unambiguously refers to a human devotee of Śiva, or Śaiva saint. One such Telugu verse in this passage, however, cites directly from the Śivadharma, while simultaneously paraphrasing the textual context of the citation. As Paṇḍitārādhya writes:

• The sentence “liṅgas are said to be twofold”
• States that if one does not worship the jaṅgama liṅga
• As prescribed, having undertaken ritual,
• Pūjās and good deeds become fruitless.55

The Śivadharma verse cited reads as follows:

• Liṅgas are said to be twofold (liṅgadvayaṃ samākhyātam): the moving and
     non-moving.
• The moving is known as “conviction” (pratīti). The non-moving is in the case
     of [liṅgas] made of earth and so forth.56

In this Telugu verse, Paṇḍitārādhya’s point seems to revolve not much around the verse that is actually cited as much as around another slightly subsequent Śivadharma verse, which states that the fixed (sthāvara) liṅga is useless without the moving (jaṅgama) liṅga: “Through disrespect of the jaṅgama, the sthāvara becomes fruitless. Therefore, the wise one should never disrespect the pair of liṅgas”.57 We may assume, then, that Paṇḍitārādhya intends to invoke for his readers not simply the verse cited, but the wider discursive context of the twofold typology of liṅgas as discussed in the Śivadharma. As with all of his partial citations, the meaning of the Śivadharma verses cannot be coherently read without background knowledge simply from the elliptical Sanskrit provided. In this respect, Paṇḍitārādhya’s multilingual idiom appears to have been a foundational influence on Somanātha’s Telugu works, which also weave partial Sanskrit quotations directly into the Telugu grammar of his dvipadas. Thus, succinctly, the Śivatattvasāramu reveals a discursive world in which the Śivadharma was quite well known to his intended audience. He intends, evidently, not to teach something his audience has never encountered before, but to evoke a scriptural canon they can instantly recall even from the mention of a few key words. If the Śivadharma was, then, not new for Pālkurikĕ Somanātha’s audience, it was likely not a novel source of inspiration in Paṇḍitārādhya’s generation either.

Now, the fact that the term jaṅgama predates the advent of Vīraśaivism proper is, in and of itself, not a new finding. In fact, David Lorenzen has already discussed this in his landmark study of the Kālamukhas, noting where the term jaṅgama appears in our inscriptional record as associated with Kālamukha institutions. The Śivadharma verses that mention this term, however, are ambiguous: is a “moving” liṅga a human saint, or a portable miniature śivaliṅga? Within the Vīramāheśvara context, for instance, the related word, caraliṅga (“moving liṅga”) retained the separate meaning of a portable śivaliṅga, since we are provided with detailed measurements of its allowable dimensions (Fisher 2019 , pp. 32–33.). As of yet, we know relatively little about which interpretation of the term jaṅgama or jaṅgamaliṅga would have been most current in distinct pre-Vīraśaiva historical and discursive contexts. Indeed, the original Śivadharma verse itself does not precisely inspire confidence that jaṅgama was originally, in all cases, intended to mean a moving saint, as “pratīti”, the term used in the definition of the “moving liṅga” (caraṃ pratītivikhyātam), does not conventionally have that meaning. Nevertheless, in the spirit of Lorenzen’s inscriptional evidence, the testimony of the Śivadharmavivaraṇa, a rare commentarial voice from the tradition, also suggests that the concept of the jaṅgama was a decidedly pre-Vīraśaiva development: “Intending to articulate that the Māheśvaras also are to be respected like Śiva himself, [the text] points out that they, also, are considered liṅgas”.58

But what, then, does the term jaṅgama mean for Paṇḍitārādhya? While he does not definitively state his position in the Śivatattvasāramu, the nearby context of the Telugu verse cited above suggests that the term jaṅgama did refer to a Śaiva devotee, as the verse appears immediately after a discussion of the pūjā of the Śivabhaktas themselves:

Without having worshipped the Śivabhaktas,

Having performed many crores of pūjās to Śiva

Is useless. To worship the Śivabhaktas

Is to perform crores of pūjās to Śiva, O Rudra.59

Thus, the Śivatattvasāramu provides us with evidence that the jaṅgama had already acquired its conventional meaning as a human saint, and moreover, the text understood this meaning to be associated with the interpretive traditions of the Śivadharma. A further intriguing example occurs in a citation preserved in Jyotirnātha’s Śaivaratnākara, where we meet with a variant reading for this very same Śivadharma verse. As Jyotirnātha cites: “There are said to be two types of liṅgas: the moving and non-moving. The non-moving is made of earth and so forth. The moving is known as the guest (atithi)”. Although we cannot as of yet be certain if the verse was already modified in Jyotirnātha’s text at the time of composition, this seemingly minor variant is doing significant interpretive work: while the original verse may also refer to a portable śivaliṅga, the Śaivaratnākara restricts possible interpretations with the word “guest” (atithi) to provide an impeccable scriptural precedent for the worship of the human jaṅgama.60 Whether further manuscript research locates this shift at the text’s inception or in later textual drift, this Śivadharma verse provides an intriguing snapshot of textual redaction in process. It does suggest, in either case, that early Vīraśaiva exegetes were uncomfortable with the ambiguity in the original Śivadharma, and saw that text as the ideal authenticator of Vīraśaivism’s new approaches to Śaiva praxis.

According to our textual evidence, then, the jaṅgama as moving liṅga was one of many concepts the Vīramāheśvara tradition shares with the Śaiva Age interpretive tradition of the Śivadharma. Should we conclude, then, that the Śivadharma was the sole proximate source for the entrée of these doctrinal elements into Vīraśaivism? As it turns out, textual evidence from further south in the Tamil country complicates matters a bit further. An epigraph preserved from the reign of Kulōttuṅga Cōḻa explicitly mentions the patronage of a group known as Vīramāheśvaras.61 This reference, however, contains little contextual information as to what sort of religious practice these “Vīramāheśvas” may have advocated. We do, however, possess an external source for this evidence, brief as it is, from a Śaiva doxography that seems likely to date back to the Cōḻa period in question. A circa seventeenth-century Tamil work, a commentary on the Ñāṉāvaraṇaviḷakkam by Vĕḷḷiyampalavāṇar, preserves an extensive Sanskrit citation from a work entitled the Sarvasiddhāntaviveka,62 in which we meet with a description of a group of Mahāvratins who espouse a form of practice reminiscent of early Vīraśaivism. According to the Sarvasiddhāntaviveka, these Mahāvratins appear to advocate the bearing of the personal liṅga (liṅgadhāraṇā) as a central religious practice, and insist that the liṅga must be borne on the body only above the navel.63 The Srisailam Vīramāheśvaras attribute just such a restriction to the Vātulatantra, a text that the Tamil commentator Vĕḷḷiyampalavāṇar describes as a “Mahāvratatantra”.64 But moreover, and crucially for the present instance, the Sarvasiddhāntaviveka also links the practice of liṅgadhāraṇā explicitly with devotion to jaṅgamas. As the verses in question pair the term jaṅgama directly with the term guru, clearly a human figure, it appears that the “moving liṅga” quite clearly depicted the human Śaiva devotee for this audience:

He always bears the liṅga on his own head, or his shoulders,

Or on other places above the navel, such as the heart, etc., according to the śāstra.

Liberation [derives] from bearing the liṅga; how much more so from the worship of

  men?

As with devotion to Śiva, so with devotion to the guru and the jaṅgamas.

Even so, devotion to the jaṅgama is called the “particular” (viśeṣa).

Those who are intent on the daily rituals and so forth stated in the śāstra known as

  the Great Vow (mahāvrata)

Set forth for liberation in a single lifetime. Thus, here [on earth], they are “those of

  the Great Vow”.65

In other words, given this contextualizing information, it would appear overly facile simply to conclude that the theology of the jaṅgama was inherited by early Vīraśaivas directly from the raw text of the Śivadharmaśāstra without any textual or institutional intermediaries. As we have seen, the early interpretive context for these Śivadharma verses does attest to the fact that the jaṅgama was previously understood in the Vīraśaiva sense as a human saint. Moreover, while at this time the evidence available to us is fragmentary, the Vīraśaiva understanding of the term jaṅgama can also be traced through at least one intermediary discursive context in circa twelfth-century Tamil region, in which other practices favored by the Srisailam Vīramāheśvaras, such as liṅgadhāraṇā, seem to already be associated with each other. Yet the term liṅgadhāraṇā, central as it had become even to early Vīraśaivism, is not attested in the Śivadharma itself as a component of lay Śaiva practice, nor are the names later attributed to the miniature liṅga borne on the body.

The iṣṭaliṅga, or personal aniconic image of Śiva, is today quite renowned as a definitive marker of Vīraśaiva religiosity: initiates are generally obligated to wear around their necks a miniature śivaliṅga imparted to them upon initiation, and for which they traditionally perform daily pūjā, enshrined in the base that is the palm of the hand (karābjapīṭha). The Srisailam Vīramāheśvaras frequently invoke the concept of the iṣṭaliṅga, most frequently referred to as the prāṇaliṅga or sveṣṭaliṅga. While the Vīramāheśvara terms for such a personal liṅga do not appear in the Śivadharma nor its successors, they do appear in other sources that were directly known to the Somanāthabhāṣya, including the Somaśambhupaddhati. Likewise, preserved within Vīramāheśvara texts, these terms appear in non-Saiddhāntika ritual procedures, such as initiation (dīkṣā), which may originally derive from a Kālamukha, or perhaps a similar Mahāvratin lineage of transmission. This discrepancy underscores the fact that other foundational Vīraśaiva ritual elements cannot be traced to the Śivadharma, and must be excavated elsewhere within the sources cited by Somanātha and his successors.

These and other related issues will be discussed at greater length in other contexts, but suffice it to say for the present moment, an excavation of the Śaiva Age precurrents of Vīraśaivism cannot be limited to the Śivadharma. While I hope to expand upon these findings in future publications, the following points should, I hope, be clear from the present article: (1) early Vīramāheśvara texts such as the Somanāthabhāṣya drew substantially upon the textual resources of the Śaiva Age and their religious systems of value, including, but not limited to, the Śivadharmaśāstra, and (2) the recovery of the history of this inheritance is best approached by bringing both Sanskrit and vernacular textual evidence into dialogue. In the thirteenth century, after all, Śaivism was not exclusively entextualized in Sanskrit, and bhakti was not exclusively expressed in the vernacular.