1 Introduction

It is well known that religious agents in premodern South Asia appealed to the purported timelessness and transcendence of their scriptural sources to secure not only their legitimacy but also that of the ideas found in them. Equally well known—in no small part due to the work of Alexis Sanderson on the social history of Śaiva and other religious traditions in early-medieval Kashmir and elsewhere—is the fact that premodern South Asian religions do not appear as unchanging and immobile traditions in social stasis. Quite the opposite: the various religious traditions of medieval South Asia were nothing if not innovative in idea and practice, most notably in their literary productions. These myriad religious traditions, moreover, had a measurable and not insignificant influence on contemporaneous social life.

In beginning to address the question of religious change in premodern South Asia I would like not merely to point out that the religious practitioners of the day were surely able to distinguish new religious, and other, ideas and practices from received tradition—just as we are today—a fact that itself calls into question the reification of the sort of social stasis and lack of historical awareness posited in previous Indological scholarship.¹ I also will argue that a self-conscious, emic theory or explanation of scriptural authority and social change may be found in the history of religions in premodern Kashmir. I wish to examine the significance of novelty, of innovation, as it was conceived by the authors I propose to place under study, and to identify its role in establishing, or challenging, religious authority in the period in question, all of which I will do by exploring an exemplar that illustrates what I suggest should be taken as a maxim in the study of South Asian religions (and religion more generally), namely, that change is not inimical to religion, even if particular religious agents are not infrequently inimical to change.²

In particular, I propose to examine selected writings of some among the most prominent of the Śaiva tantric “post-scriptural” authors of the Kashmir Valley, who thrived there in the period reserved by the title of the present essay. They include Somānanda (circa 900–950), Utpaladeva (circa 925–975), and Abhinavagupta (circa 975–1025), as well as Jayaratha (early 13th C.E.), the author of the Tantrālokaviveka commentary on the Tantrāloka (hereafter TĀV and TĀ, respectively). Their textual contributions are properly associated with what continues, sometimes, to be labeled “Kashmiri Shaivism,”³ and they offer an important opportunity to reevaluate the nature of various sources of religious authority in premodern South Asia, as well as the manner of establishing and deploying such authority in practice in India’s premodernity.

This is so, firstly and primarily, because the authors in question explicitly wedded themselves to the notion that novelty, the production of something new, was not only acceptable, but was an asset to their religion. This they maintained over and against the ethos of Vedic religious authority, which was defined by its claim to offer a kind of knowledge that was beyond question, its infallibility guaranteed by the purported timelessness and authorlessness of the Vedic scriptures themselves. Secondly, and related to this, they lived and wrote in the Valley at a time when Kashmir was economically dynamic and culturally and religiously pluralistic, and was perhaps the most active center of Sanskrit literary production in the Indian sub-continent not only in its day, but in Indian history (after the time of the Guptas).⁴ Indeed, the bibliography of Kashmiri contributions, one feels, is practically asymptotic to that of premodern Sanskrit learning tout court.⁵ Quite a number of innovative literary works were produced there in this period, moreover; and some among the Śaiva authors in question actively participated in this cultural and literary tradition by composing texts that contributed to selected non-tantric literary genres. A study of these authors thus opens a window into the ways in which religion functioned in larger intellectual and public spheres in a period of tremendous intellectual and cultural flourishing, even if the present essay will examine the internal logic of the selected tradition’s philosophical and religious writings rather than explain the relationship between such writings and those of other, more “secular” genres.

The wider cultural context of the Kashmir Valley is of note, because scholars have sometimes linked the purportedly transcendent nature of religious authority—and of textual production more generally—to a concomitant, forced stasis in the social order. Rigid and purportedly timeless ideas about the nature of reality and the place of various peoples situated in it, the thinking goes, produced and supported rigid rules concerning the innate rights and capacities of particular individuals and groups, with such rights and capacities inevitably being hierarchically and inequitably defined, all in a manner as resistant to change as the transcendent sources of religiously sanctioned authority were held to be. The precise social implications of a religious worldview that welcomes innovation, in a context in which innovation in literary production was not uncommon, is therefore of primary concern in defining the role of religion in public life in premodern South Asia.⁶

In what follows I will proceed in three stages. First, I will offer a précis of the problematic associated with the issues to hand, this by examining Pollock’s theorization of the relationship of theory (śāstra) to practice (or prayoga) in premodern South Asia. Pollock is selected for further examination because his writings on premodern South Asian cultural change are, in my view, the most sophisticated and engaging of any scholarly treatment of the subject to date, even if I will ultimately adopt his view only in part, challenging it in part as well.

Following this, I will examine key textual passages from the writings of the tradition here placed under study, this to illustrate the ways in which transcendentally authorized religious ideas can be conceived simultaneously to be historically situated and, indeed, new—and therefore demanding of a new social consciousness. Here I wish to support the claim that social change in premodern South Asia tends almost invariably to be incremental, not revolutionary, in nature. It regularly—not to say exclusively—involves modifications of social norms and strictures, not wholesale changes thereof. The relevant textual passages will be examined in two sections, the first dealing with the tradition’s emic theory of textual authority, the second with its religious practices.

Finally, I will conclude by charting the implications of this exemplar in a concentric series of contexts, culminating with the study of religion, broadly conceived. If I might be allowed to argue by way of example—ab uno disce omnes—I will suggest in conclusion that the present illustration offers insight into the nature and functioning not only of religion in premodern India, but also of religion tout court.

2 Theory and Practice: Sheldon Pollock on śāstra and prayoga

The issues associated with endeavors to negotiate cultural and social change may be set in the brightest relief if we read them in light of what is in my view the most sophisticated treatment to date of the relation of śāstra—i.e., (textual) theory—on the one hand, and cultural and intellectual practice or prayoga, on the other, that of Sheldon Pollock (as articulated in a pair of articles: “Mīmāṃsā and the Problem of History in Traditional India” [1989] and, perhaps more importantly, “The Theory of Practice and the Practice of Theory in Indian Intellectual History” [1985]). Pollock has argued, in what is the most significant theoretical contribution of this pair of articles, that classical Sanskrit theoretical works in premodern South Asia articulated an epistemological formulation according to which there was no dialectical relationship between theory (śāstra) and practice (prayoga). New practices were not seen as able to challenge or shape theoretical models, Pollock has argued, due to the claimed a priori status of the theoretical works themselves, which came to be understood to represent (in, he suggests, the language of Ryle) a form of “knowing that” a given social or cultural form was in its very nature just as it appeared in practice, as opposed merely to offering a “knowing how” the social or cultural form in question functioned or could function (based on empirical observation).⁷

This is to say that the purportedly timeless and authorless works defined the world in terms of what it is and should—and must—be, rather than articulating a contingent theory, one subject to revision on the basis of on-the-ground practices, or in other words in a dialectical relationship with events. “For here,” Pollock argues, “on a scale probably unparalleled in the pre-modern world, we find a thorough transformation—adopting now Geertz’s well-known dichotomy—of ‘models of’ human activity into ‘models for,’ whereby texts that initially had shaped themselves to reality so as to make it ‘graspable,’ end by asserting the authority to shape reality to themselves.”⁸ Pollock ultimately identifies three consequences to this theory of theory and its practical application: (1) the “ ‘creation’ of knowledge” was understood to be exclusively a divine activity; (2) knowledge was “by and large viewed as permanently fixed in its dimensions”; and (3) third and finally, there could be “no conception of progress, of the forward ‘movement from worse to better,’ on the basis of innovations in practice.”⁹

One product of such “ahistoricality”¹⁰ was an apparent stasis in Brahminical views of the normative social order, at least as such views were codified in the classical Sanskrit śāstras par excellence—the Vedas, which, to reiterate, were reified as authorless and timeless works and said therefore to convey transcendent and unchanging truths.¹¹ The śāstras were self-consciously understood to articulate a range of social and other normative views that were consciously held to be unassailable, because the works in which they were recorded were said to stand entirely outside the realm of mundane and contingent existence:

History, one might thus conclude, is not simply absent from or unknown to Sanskritic India; rather it is denied in favor of a model of ‘truth’ that accorded history no epistemological value or social significance. The denial of history, for its part, raises an entirely new set of questions. To answer these we would want to explore the complex ideological formation of traditional Indian society that privileges system over process—the structure of the social order over the creative role of man in history—and that, by denying the historical transformations of the past, deny them for the future and thus serve to naturalize the present and its asymmetrical relations of power.¹²

3 The Kashmiri Śaiva Example in Its Theoretical Form

If I question Pollock’s model in part, it is not because it is nowhere operative in premodern South Asia or, mutatis mutandis, in contemporary Indian religions. For, decidedly it was, and sometimes is; and Pollock’s theory adroitly explains a strategy for dealing with cultural and social change that in its primary features is also operative in other religious traditions of other times and places. Indeed, it signals what is a fundamental concern for religion and does much to theorize and explain what is a common response to it.¹³ Where I differ with Pollock is with his understanding of this model as monolithic, as functioning equally across the gamut of premodern śāstric modes of discourse and across textual genres and religious traditions.¹⁴ For the facts as we can discern them “on the ground” in the period in question challenge the universal applicability of his model, and the construction and use of religious authority in premodern South Asia was regularly if not uniformly more complex than Pollock’s model allows.

Two dimensions of textual production in premodern South Asia complicate his view of the śāstras—the normative, technical, and for our purposes religious writings—of classical India. These are, firstly, (1) the fact that the canon—or, more precisely, canons—of Hindu scripture remained (and in certain instances continue to remain) open, allowing for a proliferation of scriptural sources regularly to emerge “on the ground”; and, secondly, (2) the fact that “ahistoricality”—the very basis of the authorless and timeless transcendence that, Pollock has shown, furnished (religious) authority—was conceived in multiple ways, including, in the tradition placed under consideration in this essay, in a manner that allowed purportedly divine and timeless teachings simultaneously to be associated with the particular biographies of historically-located religious figures (“perfected ones” or Siddhas, in the present example). This pair of dimensions of scriptural production allowed for the recognition of and advocacy for change in Hinduism in ways that have heretofore passed without mention: religious authority could be conceived of as simultaneously new and timeless; and in a context in which religions could readily proliferate textual sources of a staggering variety, religious authority was complex, both because it was layered and because it could and did change over time. It is precisely for these reasons that the example offered by the Śaiva tantrics of Kashmir both challenges and confirms elements of Pollock’s theory.

I begin the case-study with the examination of a theoretical work, a śāstra, one inspired by Śaiva tantric scriptures and views, but one voided of explicit references to such texts: Utpaladeva’s Īśvarapratyabhijñākārikās (hereafter ĪPK), the preeminent expression of Pratyabhijñā philosophy, for which the author wrote both a short and a long autocommentary, respectively the Īśvarapratyabhijñākārikāvṛtti (ĪPKVṛ) and the Īśvarapratyabhijñākārikā-vivṛti or -ṭīkā (the latter of which is lost excepting that a number of fragments of the work have been recovered in recent years; see chapter 5 of this volume). The ĪPK and ĪPKVṛ offer an exoteric explanation for the nature of the divine and the place of the practitioner in the world. They were written by a Brahmin who in his public life would have adhered to caste norms and the associated rules of purity and pollution. Yet, as is well known, Utpaladeva and the other authors of the Pratyabhijñā simultaneously would have allowed, in theory, at least, for the contravention of these rules in the context of the family of tantric initiates, a complex practice summarized by the famous dictum, “internally a Kaula, externally a Śaiva, and in social practice a follower of the Veda.”¹⁵ (I will return to this maxim in what follows in order to examine the wider textual context in which it is recorded.)

My contention is that the ĪPK and its commentarial tradition exemplify the possibility of self-consciously introducing religiously—indeed, what are held to be transcendently—authorized innovations, and that they also serve to illustrate and explain the possibilities and limits of such innovations on contemporaneous social life. Consider, to begin, the by now well-known invocatory verse (maṅgala) of Utpaladeva’s masterwork:

Having somehow become a servant of Maheśvara, wishing to offer assistance to all people (janasyāpi), I will facilitate the recognition of [the Lord], which brings about the achievement of all successes.¹⁶

Utpaladeva’s stated wish to assist “all people” (janasyāpi) requires further examination, and I will return to this statement in what follows. More immediately, one may note two key and perhaps obvious qualities of the maṅgala as it is offered to the reader. First, Utpaladeva claims access to the transcendent, for he says he has “somehow become a servant of Maheśvara,” i.e., of Śiva. Second, he promises to bring others to a similar level of accomplishment by accounting for his insight in the very text that is to hand, paving a path for others to the same ennobled state.

The first quality is anticipated in the maṅgala with which Somānanda opens his Śivadṛṣṭi (hereafter ŚD), a verse that is also by now well known:

May Śiva, who has penetrated/possessed my form by warding himself off by means of his own self, pay homage to his (all-)extensive self by means of his own power.¹⁷

Here, as before, the author’s association with Śiva is noted, though it is rather more prominently asseverated than is the association with Śiva described by Utpala in the ĪPK—for the ŚD uses the Kaula scriptural idiom of possession. We thus clearly are not met with those who count themselves merely to be informed intellectuals, but with those who claim to be both scholars and practitioners, who understand themselves to be close to—or in fact the very embodiment of—God in the form of Śiva.

So much is confirmed in the history of the transmission of the teachings of the Pratyabhijñā that is found in a textual passage appended to the Śivadṛṣṭi, said to be an autobiographical account of Utpaladeva’s teacher, Somānanda.¹⁸ While this quasi-mythological passage claims a divine origin for the Pratyabhijñā, origins that were shrouded with secrecy, it emphasizes the emergent quality of the teachings—their very entry from obscurity into circulation among men, or, in a word, their novelty. For the teachings, we are told therein,¹⁹ were kept by “great-souled sages” (mahātmanām ṛṣīṇām) until they retreated from the world at the dawn of the Kali Yuga, at which point it was incumbent on God (deva) himself to cross down to earth “in order to grace [humanity]” (devaḥ … anugrahāyāvatīrṇaḥ … bhūtale).²⁰ This he did by instructing the irascible and imprecating sage Durvāsas to perpetuate the lineage of initiates. Durvāsas is said to do so by producing a mind-born son named Tryambakāditya, who stands at the head of a fifteen-generation lineage of such mind-born Siddhas (all named Tryambaka).²¹ The last falls in love with a Brahmin woman who bears him a son named Saṃgamāditya. He in turn is said to have moved to Kashmir,²² and over the course of the generations the lineage and its esoteric teachings were passed down to Somānanda. Thus, we are told, the teachings were not lost (nocchidyate).

I would further suggest that the type of autobiographical expression offered by Somānanda is also found in Abhinavagupta’s famed autograph verse, with which he opens virtually all his Trika commentarial works and which Sanderson has examined, already, in detail. Sanderson (2005, 89) has rendered this bivalent (śleṣa) maṅgala verse as follows:

May my heart shine forth, embodying the bliss of the ultimate, [for it is] {one with the state of absolute potential made manifest in the fusion of these two, the ‘Mother’ grounded in pure representation, radiant in ever new genesis, and the ‘Father’, all-enfolding [Bhairava], who maintains the light [of consciousness] through his five faces} / {formed from the emissions produced through the fusion of these two, my mother Vimalā, whose greatest joy was in my birth, and my father [Nara]siṃhagupta, [when both were] all-embracing [in their union]}.²³

These expressions, in turn, reflect the tradition of the poet’s autograph verse, which, it must be added, is something of a late phenomenon in the history of Indian literature. I offer for comparative purposes one example thereof, that of Ratnākara (as rendered by David Smith), which opens his famed Haravijaya:

Ratnākara, son of Amṛtabhānu who was the son of him who lived at Gaṅgāhrada on Himādri’s peak and was of Śrīdurgadatta’s lineage, wrote this poem, a lovely composition, which is beautiful because it is based on the deeds of Śiva, whose crest is the digit of the moon.²⁴

The similarities lie precisely in the fact that by identifying his parentage and/or family lineage each author stakes a claim to his unique biography and thereby his historicality, even if each does so without explicitly situating himself historically with any traceable reference to a calendrical date or reign of any particular royal sovereign. The autograph, then, furnishes precisely what this term suggests, a marker of the unique identity of the author.²⁵

It bears reiterating that the historical accuracy of such accounts is not here pertinent; what is of concern in these and other similar expressions is the fact that their authors conceive of themselves not only in terms of their historicality, the mere fact that they appear in time, but also in terms of their historicity:²⁶ the authors in question claim for themselves—and whatever (religious) authority they claim—the possibility of being mapped historically, located in particular social, political, religious, and other contexts due to their own distinguishable personages and individual biographies.

Most significant is the fact that Utpaladeva is explicit in the ĪPK about the nature and significance of the spiritual path he describes: it is not only efficient and effective, but it is explicitly new.²⁷ This Utpaladeva tells us in the concluding passage of the fourth and final section (adhikāra) of the work, where, reflecting on his completed text, he says:

Thus I have set forth this new (nava), easy path as it was explained by my distinguished teacher [Somānanda] in the śāstra, [entitled] the Śivadṛṣṭi.²⁸

We see this sentiment reiterated, moreover, in the ĪPKVṛ, where Utpaladeva glosses the term “new” (nava) with the synonym abhinava;²⁹ and Abhinavagupta reproduces this latter term in the shorter of his two sub-commentaries, the Īśvarapratyabhijñāvimarśinī (ĪPV), where he adds that the path in question is both “included in” all the various secret śāstras (sarvarahasyaśāstrāntargata) and, because it has been concealed from public view, is not well known (sannigūḍhatvād aprasiddha).³⁰

Isabelle Ratié, the first scholar explicitly to note the significance of this passage, has suggested that Utpaladeva refers to the novelty of the path he described in the ĪPK because he wishes to indicate that it does not authenticate itself by any overt appeal to scripture.³¹ Now, it is undoubtedly the case that the Kārikās are unmistakably philosophical in tone and content, and that they require no special or extensive knowledge of esoteric scriptural sources for their comprehension; and one may note that Utpaladeva’s teacher, Somānanda, similarly obscured the scriptural influences on his articulation of the Pratyabhijñā.³²

I would propose, however, that one must understand a rather different justification for Utpaladeva’s description of the path he describes as “new.” It is not—or, rather, it is not merely—the fact that one need not appeal to Śaiva scriptures to follow it; rather, I suggest that the primary sense of Utpaladeva’s description of the path as “new” is that it is new to humanity. As is evinced in Somānanda’s “autobiography,” the Pratyabhijñā teachings were said to have been concealed by great sages at the beginning of the Kali Age, preserved as they were by means of the secret lineage described therein, and were only revealed subsequently for the benefit of those who would come to be initiated into its ways. Some justification for linking these mythological origins to the attested novelty of the path may be found, moreover, in the fact that Abhinavagupta explicitly invokes the narrative found in Somānanda’s “autobiography” in a passage of the ĪPVV that comments on ĪPK 4.16, precisely the verse in which the novelty of the Pratyabhijñā’s path is proclaimed. And in doing so Abhinavagupta may be understood himself to associate the lineage with Utpaladeva’s claim to novelty.³³

The Pratyabhijñā thus presents the reader with a self-conscious (theoretical) understanding of their śāstra as divinely sanctioned, guaranteed by the identity of the authors in the lineage with Śiva himself; and, while thus divinely sanctioned, the śāstras are intimately, inextricably, tied to the biographies of individual, monadic agents—Utpaladeva, Somānanda, and those preceding and following them in their lineage. Thus, and perhaps not unlike the theology associated with, e.g., the avatāra theory, we are here offered a view of Siddhas or “Perfected Ones” who have emerged in human form in the world in their particular historical contexts, contexts to which they could—and did—actively and explicitly respond. Finally, this concurrence of historicity and divine authority allows the ĪPK in particular to claim to offer something that is plainly and explicitly “new” (nava), yet nevertheless “transcendent,” even “timeless,” as it were, insofar as authorized by Śiva himself. The Pratyabhijñā thus achieves in its explicit and implicit theoretical formulations the apparently paradoxical aims of offering something both historically located and new, on the one hand, and transcendent, on the other, the teachings being ultimately authored and authorized by Śiva himself and thus held to be uncontained by any historical bounds.

4 The Kashmiri, Śaiva Example in Practice

This much, then, represents the Kashmiri non-dual Śaivas’ theory, their self-conscious view of the place of their śāstra and those who carry it in history. But, what of their practices? What did their theory suggest they should do, and what, so far as can be known, did they do?

As is well known, the Śaivas of the Pratyabhijñā understood the various practices related to varṇa or birth to involve mere social custom, their view being that caste identity was not based in any real difference in the natures of the individuals so marked by it. We see this view exemplified perhaps nowhere more clearly than in a passage of Somānanda’s ŚD (at 3.42cd–47), where the question of the need to purify a fire installed in an outcaste’s, a caṇḍāla’s,³⁴ house is addressed:

svecchāto bhāvarūpatve parādhīnā kutaḥ sthitiḥ 3.42

kṣīravad yadi vocyeta parādhīnaṃ jaḍaṃ bhavet

etayaiva diśā śodhyaṃ śuddhanyūnādidūṣaṇam 3.43

abhagne ’sya svarūpatve śuddhanyūnādikaṃ kutaḥ

patadgrahādike hemni hematvaṃ mukuṭādike 3.44

sthitam eva na hemno ’sya kācid asti vibheditā

caṇḍālasadmago vahnir na vahnir yadi kathyate 3.45

tad evaṃ syād athocyeta vahneḥ saṃskāracodanā

śāstreṣu varṇitā kasmāt kāryārthaṃ kāryam eva tat 3.46

na svarūpavibhāgo ’tra tathā tatra vyavasthiteḥ

saṃjñākaraṇamātraṃ tad vyavahārāya kalpitam 3.47

Given that he [i.e., Śiva] exists of his own volition in the form of (all) the entities (that make up the universe), how is existence dependent on another than himself? If, for example, you say it [i.e., the purported dependence] is one similar to (the example of curds, whose genesis depends on the) milk (of which they are comprised), it [i.e., the universe] would be insentient, dependent on another. The fault (attributed to our system) that must be corrected—being pure, being diminished, or the like—is precisely the result of this (wrong) point of view. How can there be something pure, something diminished, etc., when his nature is undivided? The fact of being gold simply exists in gold, (be it) in (the form of) a golden spittoon, etc., or in (the form of) a tiara, etc. The (fact of being) gold is in no way divided [i.e., differentiated] whatsoever. If you argue that a fire installed in an outcaste’s house is not (properly) called a fire [i.e., it is not a proper, ritually-purified fire], we reply: that may be so [i.e., this does not contradict our notion of the uniformity of the nature of fire as such]. But, if you say (in reply) that injunctions involving the rites of passage for fire are explained in the (Śaiva) teachings, (we reply:) what is the goal of the (ritual) action (in question)? It is the action itself.³⁵ There is no division of its [i.e., the fire’s] nature, here. It is the same for his [i.e., Śiva’s] abiding there [in the world]: that [i.e., the distinguishing of “pure” from “impure” elements in the world, or the distinction of that which is said to be Śiva and that which is said not to be] is conceived of merely as the assignation of names for the purpose of everyday speech/everyday activity [vyavahāra].³⁶

That caste distinctions should not disqualify anyone from hearing the teachings of the ĪPK has also been stated plainly enough, if somewhat obliquely, in the maṅgala verse of the ĪPK, where, as noted above, the “new path” (nava mārga) is explicitly said to be offered to “all people” (janasyāpi). While the maṅgala is merely suggestive of any implications regarding caste, however, the ĪPKVṛ and Abhinavagupta’s ĪPV and ĪPVV further clarify what is intended by the expression; and what Utpaladeva means by “all people” is explained with increasing clarity as one delves into the succeeding levels of commentary.

The ĪPKVṛ indicates that one should understand the term to refer to all people without exception: Utpaladeva glosses janasyāpi with imam akhilam.³⁷ Abhinavagupta, in turn, addresses the caste implications of this pledge—to help all people—in a more explicit fashion. In the ĪPV subcommentary, he indicates that the term jana refers to any person regardless of the circumstances of his birth (janasyeti yaḥ kaścij jāyamānaḥ tasya); and he further specifies that no distinction by the “office” held by a candidate for initiation (adhikāriviṣaya) may serve to restrict access to the teachings of the Pratyabhijñā.³⁸ He later adds that the term in question should be understood to refer to anyone afflicted by the suffering of incessant births and deaths.³⁹

Finally, Abhinava also is explicit, in both the ĪPV and ĪPVV, in stating that jāti should not be considered an appropriate measure of a candidate’s eligibility for entry into the soteriological path of the Pratyabhijñā. This he does in the passages of his two sub-commentaries that address the concluding verse of the Kārikās, where Utpaladeva has again indicated that the text is meant for “[all] people.” Referring to himself in the third person, he says (at ĪPK 4.18):⁴⁰

In order that [all] people (jana) can attain perfection effortlessly, Utpala, the son of Udayākara, has presented this Īśvarapratyabhijñā [treatise].

Abhinava’s ĪPV suggests that with “people” (jana) Utpaladeva refers to any person, for neither jāti nor any other criterion is innately disqualifying.⁴¹ He reiterates the same in the ĪPVV as well, where, additionally, he excludes the individual’s “conduct” (ācāra) from consideration in determining whom the teachings may help.⁴² Note also that the ĪPVV explicitly links the present passage with the one found at the beginning of the ĪPK, saying: etac ca prārambha evāsmābhir vitatam.⁴³ Simply, distinctions marked by birth—including caste distinctions—are to be ignored.

The professed catholic reach of the work is significant, not only due to its inclusivism, a quality well attested in the context of tantric rules of initiation, even if this cannot be taken to imply the existence in practice “on the ground” of a concomitant equality of participation, or status, within the context of the initiatory “family” (kula); but it is further significant that it is declared, albeit obliquely, in the ĪPK (and ĪPKVṛ), a text (and commentary) that has all the hallmarks of works intended for audiences of non-initiates.⁴⁴ I submit that there must have been some appreciation for a certain novelty in the apparent inclusivism professed in the work. This, then, may constitute another reason for Utpaladeva’s description of the Pratyabhijñā path as “new”: it offers a certain novelty in being (more or less explicitly) indifferent to caste distinctions.

While the use of the term jana in the ĪPK offers a certain ambiguity of expression, Abhinava’s subcommentaries are explicit in stating the inclusivism of the śāstra. One can thus see operative here a certain stratification of expression, a certain obliqueness with regard to any reference to caste in the ĪPK and ĪPKVṛ, but a more explicit asseveration of the sociological parameters of the tradition in Abhinavagupta’s sub-commentaries. So much of course reflects the complex of practice of the tradition, exemplified as it is by the dictum quoted above, “internally a Kaula, externally a Śaiva, and in social practice a follower of the Veda.”

This dictum and the complex practice it invokes—one of engaging different rules and norms in differing contexts—is explicitly tied to the question of social practices, those associated with caste and life-stages, as the context provided by Jayaratha for this quotation clearly indicates. For he considers the significance of preserving the social order—varṇāśramācāra—precisely where he quotes the complex form of practice that is here of concern.⁴⁵ Moreover, the entire matter is raised in the context of coming to understand precisely whether the Śaiva practitioner should respect the Veda-based rules of purity and impurity. For the relevant section of the fourth āhnika of the TĀ is governed precisely by the question of how the jīvanmukta should behave in the world, and Abhinavagupta explicitly cites the eighteenth chapter of the Mālinīvijayottaratantra there as an authoritative voice in the matter.⁴⁶ Note, in addition, that it is the matter of śuddhi and aśuddhi, as in was in ŚD 3.42cd–47, which is here (at TĀ 4.213cd) explicitly placed in question. The TĀ and TĀV go on to suggest that yogis know for themselves what is pure, what impure; and they further state that it is their very awareness that should determine how they act, for the normative rules of purity do not apply to them.⁴⁷

Following this, the key objection is raised: are there not (Vedic) injunctions that require one to follow the rules of purity?⁴⁸ The opponent asks, that is, whether tantric Brahmins must respect Vedic authority as well as Śaiva truths. How is it that, given the apparent validity or “truth” of both sets of śāstric injunction, the Vedic and the Śaiva, one can be invalidated, the other affirmed? How can one be internally a Kaula, externally a Śaiva, and a Vaidika in social practices when to be one sometimes requires one to contravene the strictures of the other?

The answer given not only suggests that it is not always inappropriate in practice to contravene the rules of purity and impurity, of śuddhi and aśuddhi; it further offers a rationale—a theory—for this complex of practice. This it does by invoking a general, and common, śāstric sensibility, namely, that the śāstras operate by first issuing blanket rules, which may later be superseded in certain specified contexts. This is a fundamental organizational structure of, among other śāstric works,⁴⁹ the system of Pāṇini’s Aṣṭādhyāyī—which (as is well known) sets out exhaustively to account for correct speech not by listing every possible form of speech, but by structuring classes of rules that first define the utsarga or generally applicable rule, only then to itemize the apavādas or exceptional rules that can override the former in the appropriate contexts. So much is explained in Patañjali’s Mahābhāṣya as follows:⁵⁰

Now, one may object by asking whether it isn’t the case that, when this (system of) instruction of (correct) speech is operative, one must recite each word in the course of acquiring the (correct forms of) words. (That is:) Every word must be recited, as in ‘cow, horse, man, elephant, bird, deer, Brahmin.’ Reply: Certainly not! It is not expedient to recite each word in the course of acquiring the (correct forms of) words. For, so much has been communicated as follows: ‘Bṛhaspati imparted to Indra the study of words by uttering each one individually over the course of a thousand divine years, and he did not reach the end of it.’ Bṛhaspati was the teacher [pravaktṛ], and Indra was the student [adhyetṛ]; and the time of instruction did not reach its terminus over the course of a thousand divine years. What of (the use of this method) today? One who plainly lives a long time (today) lives for (only) a hundred years … How, then, should these (correct) words be acquired? A certain characteristic associated with the generic [sāmānya] and the particular [viśeṣa] should be activated, by means of which [people] can—with little effort—become acquainted with the extremely massive flood of words. What, pray tell, is that? Utsarga and apavāda. A general rule [utsarga] is (first) to be employed; (then,) an exception [apavāda](, which overturns the general rule). But under what circumstances is the general rule to be used, under what circumstances the exception? The general rule is to be used generically. E.g.: karmaṇy aṇ (A 3.2.1). The exception to it (should be used) in a particular circumstance. E.g.: āto ’nupasarge kaḥ (A 3.2.3).

Here, we see that it is for efficiency’s sake that this model is adopted. And we are given a sample application in Pāṇini’s grammar, the details of which are explained in Appendix A. Essential to recognize is the fact that the prescription delineated by the general rule, the utsarga, always stands; yet, on some occasions its application is suspended by a special prescription, in the defined context of the apavāda or exceptional rule.⁵¹ Whenever no apavāda applies, however, the utsarga automatically prevails, for whenever or wherever the special rule does not apply (or must itself be suspended), one reverts by default—invariably and automatically—to the general, standing rule.

It is this very śāstric mode that is explicitly invoked by Jayaratha in reference to the very phenomenon we have here, that of a complex of religious practice, Vedic and Śaiva.⁵²

TĀ 4.230ab: ‘If you argue that this [Śaiva rule] is invalidated by the Vedic one, (we reply:) why shouldn’t it be the other way around?’⁵³

…

Considering that one may object by asking how it is that both [śāstras] are equally real, given that in certain contexts the injunction(s) [vidhi] associated with purity and the like are invalidated, even though they apply universally, for all people, [Abhinavagupta] says:

TĀ 4.230cd–231ab: ‘If you think about it clearly, a rule that is an exception [apavāda] invalidates a rule generally applied [sāmānyavihita vidhi], because it applies in a particular domain.’

So, if you properly consider the procedure of invalidation, then (you will realize that) no injunction [codanā] whatever loses reality [sattvahāni]. To explain: the rule that is the exception [apavādavidhi]—by nature specific because it is (generally) void of any occasion (for application)—supersedes the general rule [utsargavidhi], which, being one that always has met with its occasion (for application), is by nature generally applicable [sāmānyātmaka]. This is what those who know language say …

TĀ 4.231cd–232ab: ‘Moreover, purity and impurity, which are generally enjoined, are simply superseded when a man knows reality. This is how it has been explained here [in the Mālinīvijayottara].’

In the face of the objections of a more orthodox opponent who demands to know how the less permissive rules of the Veda cannot prevail over Śaiva scriptures, Abhinavagupta brilliantly inverts the hierarchy by appealing to a mainstream sāstric interpretive rule: it is the exception that preempts the normative, overarching rule and not the other way around. The rule that applies most generally, the Vedic one, is the utsarga. It is invariably prescribed—Jayaratha here described the utsarga as sarvatra labdhāvakāśa and sāmānyātmaka—but it is occasionally superseded by the exceptional rule or apavāda (described by Jayaratha as niravakāśatvād viśeṣātmā), where in this instance the apavāda is the Śaiva rule, occasioned as it is by the special knowledge of the accomplished yogin who sees reality for what it is, as transcending the conventions of social position as determined by birth.

Nevertheless, whenever this special rule does not prevail—viz., outside the confines of tantric knowledge and practice—one must revert by default to conforming to the general rule, the utsarga (i.e., the Veda-based strictures), which is never nullified but only occasionally and temporarily superseded. So much is the position Abhinavagupta articulates in the TĀ, and Jayaratha indicates that in doing so he responds to the potential criticism that tantric rules authorize practices that are proscribed by Vedic norms, implying thereby that the tantric practices should be abandoned. This complex of practice, offered by Śaivas who engaged in rites that transgressed social and other contemporaneous cultural norms, was thus said to be the necessary product of the mainstream theoretical model that they understood to prescribe it.

The implications of this exemplar, then, are evident. Divinely sanctioned (in no small part by the authors’ identification with Śiva), the Pratyabhijñā offers a tradition that prescribes exceptional practices that were introduced in historical time but are nevertheless rooted in an “ahistorical” authority. The knowledge it offers is fixed in its dimensions: there is no theology of “progress” in this śāstra. Each śāstra, moreover, Vedic and Tantric, speaks in a voice that in some important sense offers not merely a “knowing how,” but rather a “knowing that”; for each individually denies the possibility of a dialectical relationship between theory (śāstra) and practice (prayoga) in Pollock’s sense of the problematic. (New political developments could not alter its vision of the nature of Śiva as a ubiquitous consciousness, for example.) To this extent, then, Pollock’s model is applicable to these śāstras.

Yet, the tantric śāstra here placed under study nevertheless allows for a theology of novelty, because its “ahistorical” authority was said to have been introduced by particular agents in particular historical moments. The system thus allowed for, indeed anticipated, the emergence of new exceptional rules, and because the canon remained open, it was ever possible that another text, a new apavāda as it were, could emerge in a context in which the general parameters of social action—of religious practice or prayoga—otherwise and inevitably would remain operative.

A dialectic, then, between theory and practice in premodern South Asia, while it might not be found in the self-representation of any single given work, can be found in the proliferation of multiple śāstras, which in supplementing as well as competing with one another could accommodate innovation, explicitly or implicitly, by introducing novel theoretical models as circumstances warranted. The dialectic may only be found in the scriptural gestalt, however; for, it is only by reading multiple scriptural sources with an awareness of their mutually shared contexts, pedigrees, and, most importantly, structured mutual relations that one might begin to identify patterns of religious (and other forms of) change in Indian premodernity.⁵⁴

5 Conclusion

I note, in closing, a concentric series of five implications of this phenomenon, each of an increasing degree of generalization, and in doing so I intend to point to possible avenues of further inquiry that lie beyond the scope of the present (and admittedly particular) case-study. The first involves a relatively narrow methodological concern in the study of Śaiva post-scriptural sources: the present exemplar stands as a reminder that knowledge of the wider context of “orthodox” śāstric texts and concerns should inform scholarly interpretations of the post-scriptural literature; for even works that are primarily rooted in and explicitly concerned with esoteric scriptures, as is the TĀV, frequently record a substantial interaction with the mainstream systems of learning (including both various śāstric systems such as Pāṇinian grammar and the classical philosophical darśanas).

Second, there is the matter of the social context, the Kashmir Valley of the period in question, in which the authors of the Pratyabhijñā lived and wrote. While it is worth noting that the tantric initiatory “family” (kula) would have frequently offered an uncomfortable home for low-caste and outcaste initiates (as many or few as there might have been in practice), entry into the tantric kula could also have afforded the initiate a certain place of privilege. Initiation was, from the perspective of the tradition, a significant institution, one that the tantric practitioner would have coveted and shared sparingly; and membership would have conferred not only a certain prestige but also a special form of knowledge that those within the tradition would have held to be paramount: a certain access to the divine.

Thus, while this form of privilege does relatively little globally to address the social inequities enshrined in normative, caste-based strictures, the initiation of low-caste or outcaste individuals into the secret śāstras nevertheless constitutes a social innovation over and against preexisting Vedic (or, perhaps more accurately, dharmaśāstric) norms. The change it offers, however, comes not in the form of social revolution but in the form of a particular and circumscribed social inclusion. Generalizing from the present exemplar, one suspects that social change in premodern South Asia will most often be found in its incrementalism, as opposed to efforts in total revolution such as the toppling of the institution of royal sovereignty or the wholesale displacement of Brahminical authority. Social change, simply, will likely be found, where it is to be found, in the often-subtle exceptions that contradict overarching and often oppressive rules that were more globally applied, in Kashmir or elsewhere, in premodern South Asia.

Third, the model of social conduct witnessed in the Pratyabhijñā exemplar finds correspondences in the narrative tropes of (non-tantric) Śaiva mythology. Śiva is not only paradoxically erotic and ascetical in the mythological literature, as is well known, but he also appears there as both asocial or anti-social and integrated into society (and often simultaneously so), as is also well known. Take, for example, the insult of Dakṣa’s sacrifice, where Śiva is depicted as a pariah while simultaneously standing as the ideal husband in the eyes of the eminently acceptable Satī. One may therefore query whether the sort of complex of social practice examined in the present essay constitutes a phenomenon more widely attested across various Śaiva traditions.⁵⁵

Fourth, the exceptionalism offered by the utsarga-apavāda model is commonly attested across a range of premodern Indian cultural systems, most notably, for the purposes of the present essay, those that address social rules and life. The Kṛṣṇa of the Mahābhārata plainly contravenes a series of generally applicable social strictures, for example, rules that he otherwise would wish vigorously to maintain; and the rules of conduct for crisis-situations (āpaddharma) more generally are premised on the idea that the prescribed actions are exceptional both because they contravene otherwise normative behaviors and because they are deemed inactionable outside the only context in which they may apply: that of a social emergency. And one cannot overlook, in this context, the frequency with which the Veda is acknowledged as authoritative in Hindu traditions while the contents of Vedic works are nearly entirely ignored. It is common, in sum, to find a layering of practice in South Asian religion and social life; and such complexes of practice, allowing as they do for the exception, in its place, to supersede normative strictures, should probably be counted as the normative model of social organization in premodern South Asia.⁵⁶

Fifth and finally, I address an issue relevant to the study of religion, broadly conceived. It is instructive that Pollock appealed to Geertz in formulating the theoretic that defines his understanding of the relationship of śāstra to prayoga. Indeed, he finds Geertz’s famed distinction of “models of” from “models for” precisely in the article in which he famously defines religion as “(1) a system of symbols which acts to (2) establish powerful, pervasive, and long-lasting moods and motivations in men by (3) formulating conceptions of a general order of existence and (4) clothing these conceptions with such an aura of factuality that (5) the moods and motivations seem uniquely realistic.”⁵⁷

Geertz apparently did not anticipate the possibility of a complex of (sometimes mutually-contradictory) beliefs or a complex of (sometimes mutually-contradictory) practices. His model instead elaborates a more singular structure or phenomenon, referring as it does to “a system of symbols” that establishes powerful moods and motivations in men by formulating “a general order of existence,” which is clothed in “an aura of factuality,” all of which implies a model of a singular source of religious authority. There is in this, as in other similarly situated definitions,⁵⁸ an understanding of religion as involving an appeal to an ultimate that does not seem naturally to allow for variegated and equally legitimate sources of authority, that fails to account for the type of “And” model found in South Asia and, one imagines, elsewhere, as well;⁵⁹ and in defining religion in this manner, we see foreclosed any possibility that a religious tradition might innovate or simultaneously access disparate sources of authority in order to recommend diverse—or even new—doctrines and/or practices, as we see in our premodern Indian exemplar.

The example here studied, then, begs for novel theoretical models of religion, ones that could attend more fundamentally or at least more explicitly to the capacity of religion to accommodate, indeed self-consciously to cultivate, change. To proffer as much would refocus our possible ways of reading premodern religious traditions, allowing us to identify patterns of social and religious change that are not defined by modern understandings of the same (as exemplified in Pollock’s dependence on the concept of “progress” to measure any conceptual openness to social change in India’s premodernity). It would also require one to cultivate a view of religion that emphasizes not its interest in certainty, in its capacity to memorialize particular ways of acting or models of knowing—be they models of “knowing that” or “knowing how”—but its interest in possibility. Cultivating such a new model (and thereby a new interpretation, even a new scholarly definition) of religion could, that is, nurture a view of religion that emphasizes its more active and dynamic, and malleable and practical, dimensions, the capacity of religious practitioners and authors themselves self-consciously to formulate systems of understanding and to stake claims where any number of claims could have possibly been made.

This would, in short, require one to understand religion not primarily as the maps of systems and symbols and practices that are charted by religious agents, but instead in terms of the strategies they use to sketch their maps. Those who have crafted religious ideas and ideals, I think, were—and are—well aware of the fact that religions offer changing models of action and modes of reflection in the context of a changing world that is perennially met with novel problems, begging resolution. Those of us who read and theorize their practices should, in turn, seek to shape our reflections on religion in a manner that allows more naturally for the innate flexibility, the freedom of possibility—of change—in idea and practice, with which religious agents developed and develop their models for action.

Abbreviations

Acknowledgements

I would like to thank Srilata Raman and Shaman Hatley for inviting me to the conference in honor of Alexis Sanderson, titled “Śaivism and the Tantric Traditions,” at which I presented a shortened rendering of this essay. I also thank Eric Steinschneider for all he did to facilitate the running of this well-organized event. Prior to this, I delivered an earlier version of this essay in the Hindu Studies Colloquium at Harvard Divinity School on February 20, 2014. I thank Anne Monius for the invitation to speak there and for her thoughtful and instructive comments on my lecture. I also thank Ben Williams and James Reich for their comments on my lecture and for their hospitality during my time in Cambridge, MA. Finally, I thank Shaman Hatley for having read a draft copy of this essay and for the insightful comments and suggestions he offered in doing so.

Appendix A: utsarga and apavāda as Exemplified in Aṣṭādhyāyī (A) 3.2.1 and 3.2.3

3.2.1. karmaṇy aṇ.

3.2.3. āto ’nupasarge kaḥ.

1. The Pāṇinian example here offered is of a general rule, or utsarga, that is preempted by a special rule, or apavāda, in a particular circumstance. The general rule is offered at Aṣṭādhyāyī (A) 3.2.1; the special rule is subsequently given at A 3.2.3. The rules here are used to generate what are referred to as upapadasamāsas.

2. A principal purpose of the grammar, of course, is to allow one to generate the correct form of speech by the application of a series of rules that build, as it were, the correct form of speech in a series of operational steps. The point of the present example is that the general rule, A 3.2.1, regularly—indeed, always—applies,⁶⁰ unless the parameters of the special rule, A 3.2.3, apply, in which case the general rule is suspended. If the special conditions that trigger the special rule are not met, the general rule is reverted to as authoritative and is applied.

3. A 3.2.1. states that the affix aṇ is added after a verbal root (dhātoḥ is to be read into A 3.2.1 by anuvṛtti, or ellipsis, from a preceding rule) when the object of the verb is in composition with it as an upapada. This yields, e.g., kumbhakāra (“potter”) and nagarakāra (“city-builder”); kāṇḍalāva (“branch-cutter”) and śaralāva (“reed-cutter”); and vedādhyāya (“the learning of the Veda”) and carcāpāṭha (“the reading of Carcā”). In each of these cases, the final “a” of each word is present because of the affixation of aṇ. (The a of aṇ is precisely this letter or akṣara, while the ṇ of aṇ is a “marker,” an anubandha or it, which is dropped out in the derivational process and does not appear in the final form of the given word or words being derived, though it does serve to trigger certain rules and operations in the grammar, and so has a meta-function in the derivational process.)

4. On the other hand, the affix aṇ is not allowed to be added in cases where the special rule found in A 3.2.3 does apply. This rule states that the affix ka will come after a verbal root that ends in “ā,” if and only if the verbal root in question is not preceded by a preverb (upasarga), and only when the object of the verb is in composition with it (karmaṇi here is extended from A 3.2.1 to the present rule by way of anuvṛtti or ellipsis). Examples of the application of this special rule are: goda (“giver of cows”) and kambalada (“giver of a blanket”). In both of these examples, the final “a” of each word is present because of the affixation of ka. (The a of ka is precisely this akṣara, while the k of ka is again a “marker” or anubandha/it that is dropped out in the derivational process and does not appear in the final form of the word being derived.)

5. If we were to have been able to use the affix aṇ instead of the affix ka in the present pair of examples, we would have been required to preserve the long ā in the verbal root dā, which is present in both examples (goda and kambalada), generating thereby the wrong forms. Using rule 3.2.1 instead of 3.2.3 (and thus using aṇ instead of ka) would give us the wrong forms *godā and *kambaladā in place of goda and kambalada. This would be bad!

6. Similarly, in an instance where the verb employed does take an upasarga or preverb, we have an exception to the special rule, which prohibits the use of ka in place of aṇ precisely when there is a preverb affixed to the verb. Thus, even when the verbal root in question ends in long ā and is in composition with its object as an upapada, rule 3.2.3 cannot apply. A pair of examples is given with the verbal root dā + sam. The apavāda rule (A 3.2.3) is not allowed to take effect, and therefore ka is not appended to the verb. Thus, we derive gosaṃdāya (“who ceremoniously gives a cow”) and vaḍabasaṃdāya (“who ceremoniously gives a vaḍaba horse”), and here the final vowel (a) of each term is the product of aṇ and not ka.

7. What is at stake in this last pair of examples (viz., those of gosaṃdāya and vaḍabasaṃdāya) is the status of the long vowel (ā) in the verbal root dā + sam. If the affix ka were used in the present examples, then the “marker” (anubandha, it) k would have triggered another rule of the grammar (A 6.4.64, see paragraph 10, below), one that deletes this long vowel. If this marker (k) is not in play, on the other hand, as it is not, for example, when the affix marked with the anubandha ṇ is used (as in aṇ of A 3.2.1.), then the rule in question that deletes the long ā does not apply.

8. So, to review before concluding: we here have a general rule, A 3.2.1, which applies generally. It is the utsarga. We have a special rule or apavāda (A 3.2.3) that applies within what otherwise would have been the domain of this general rule, but which is commanded by the special rule under special circumstances, with the application of the special rule blocking the application of the general rule or utsarga.

9. How does this figure in the present example? Well, A 3.2.3 should apply wherever we have the requisite special circumstances, namely whenever a verbal root ending in “ā” is in composition with its object as an upapada, as it did in the derivation of goda and kambalada. But the apavāda rule (A 3.2.3) does not apply in the derivations of gosaṃdāya and vaḍabasaṃdāya, even though in both cases we have a verbal root that ends in ā and that is in composition with its object as an upapada. This is so, because in both cases the verbal root appears with a preverb (upasarga), and A 3.2.3 specifies that it can apply only in instances when there is no preverb (it reads: anupasarge).

10. Because the examples gosaṃdāya and vaḍabasaṃdāya include a verbal root that appears with an upasarga (the verb is dā, the upasarga or preverb is sam, of course), they are no longer to be marked by ka. This is significant because of what k does: according to A 6.4.64 (āto lopa iṭi ca), the final ā would have been elided, the rule effecting as much either because it appears before an ārdhadhātuka suffix (defined at A 3.4.114) augmented with iṭ, or when the affix begins (as would have been so in the present instance) with a vowel and has an indicatory marker (aunbandha or it) of k or ṅ (the affix a in ka begins with a vowel—a of course being a vowel and the first member of a single-lettered affix—this after taking into account that the k is a marker and not a part of the affix itself).⁶¹

11. Thus, in these instances A 3.2.3 does not apply. And here is the key to understanding the present example of utsarga and apavāda: the utsarga or general rule (i.e., A 3.2.1) is thus again deemed operative and automatically so, by default. This is so because it is understood always to apply, only excepting when an apavāda overrules or suspends it. What this means is that there is no need to write yet another rule to account for the derivation in instances where the rule of exception or apavāda is itself rendered inoperative by an exception; for upon the suspension of the apavāda rule, the utsarga is immediately and automatically understood once again to apply. As such—returning now to the pair of examples here reviewed—the final a of gosaṃdāya and vaḍabasaṃdāya results from aṇ being applied, not ka. One returns to the original rule by default, thus providing for the final letter a of both gosaṃdāya and vaḍabasaṃdāya. And the anubandha ṇ, unlike k, does not trigger A 6.4.64 and thus does not lead to the elision of ā in the verbal root dā appearing in both compounded words. In a word, apavādas can only be written in contexts where an utsarga is always otherwise operative.