1.4 Source Criticism

[Full title: History and Structure of Svacchanda Tantra; (4) Source Criticism]

53 1.1.4 Source Criticism In this way, a wide range of traditions or strata have influenced and moulded the composition of Svacchandatantram both in the early period of its compilation and in the ongoing course of its transmission. Following this tradition criticism, which has identified the text's respective strata, internal analysis should pursue literary criticism, which would attempt to identify the actual documentary sources used by compilers in constructing the text. Ostensibly, the degree of influence or impact of each stratum should directly correspond to its proximity to the text compilers, and therefore to the probability that they employed oral and written sources drawn from that stratum or tradition. Extremely detailed similarity or pervasive knowledge would strongly suggest that the text compilers and redactors had direct knowledge of the documentary sources of a specific tradition. Unfortunately, in the case of Svacchandatantram, since many of the documentary sources of these traditions are lost, not yet published, or known only in altered form, literary criticism can only preliminarily identify or hypothetically reconstruct probable sources. This lack of positive, external evidence demands that literary criticism construct hypotheses from internal, negative, and analogous evidence. First, internal evidence such as consistent differences in content and style may point to different sources; second, negative evidence such as textual hiatuses and discontinuites may delimit sections drawn from these sources; and third, analogous evidence based on the compositional history of similar anonymous texts in Sanskrit and other branches of literarture may suggest the procedures followed in compiling these sources. These methods of determining individual sources, their number, and their sequence also depend upon and affect redaction criticism, which attempts to identify those who actually compiled, transmitted, and altered the text. This retrospective and internal

54 criticism, naturally, can not afford to ignore the slightest irregularity that might hint at the hand of a redactor, and thus runs the constant risk of over-interpreting and multiplying the number of sources, recensions, and compilers. 1 As a corrective to this tendency, a schematic and simplying presention of the more secure results of this analysis will precede a detailed discussion of methods and illustration of their application. 2 The main deity, title, and dialogue participants of the present recension of Svacchandatantram strongly support the hypothesis that at its basis lies a written document from the "Bhairava" tradition. This document, the "Bhairava" source, as it were, likely recorded practices associated with the worship of 1 For an example of a methodic attempt to read a text as a unity, in reaction to this type of criticism, v. Georg Feuerstein, The Yoga-Sutra of Patanjali, An Exercise in the Methodology of Textual Analysis (New Delhi: Arnold Heinemann, 1979). Feuerstein programmatically argues (pp. 39-40): "Past scholars, on the whole, proceeded from the assumption that the Yoga-Sutra cannot possibly be a single homogeneous textual entity owing to its apparent disorganisation. In the following I shall commence from the opposite end as it were by presupposing the perfect homogeneity of the text. In other words, I shall look for points which seem to contradict, or at least seriously challenge, this basic working hypothesis. In this way I hope to avoid the fallacy common to all attempts of textual criticism so far, namely to cut more and more slices from the cake until it simply vanishes out of sight and nothing but disconnected fragments quite meaningless in themselves - are left behind." He also attempts to isolate this critical tendency as a cultural basis: *... before one contrives distortions, interpolations, etc., one must ask onself seriously whether what seems to be 'corrupt', 'confused' or 'patched together' is not merely the result of an unwarranted demand for absolute logical consistency characteristic of our specific thought pattern." As an exemplar in Indology of the dissecting type of text criticism. Feuerstein appropriately cites Friedrich Weller's critical study of the Kathopanisad: Friedrich Weller, Versuch einer Kritik der Kathopanisad, Deutsche Akademie der Wissenschaften zu Berlin, Institut fur Orientforschung. Veroffentlichung 12 (Berlin: Akademie-Verlag, 1953). - 2 V. infra in the summaries of individual books.

55 Bhairavah, or specifically, Svacchanda-bhairava. Alongside information about formula, gestures, and other aspects of ritual, panegryic interludes probably mixed with inconographic cum meditational data. This Bhairava source might have served to codify the essential elements of the cult of Svacchanda-bhairava, and might have been used as a guide to rituals described not only in the text of Svacchandatantram, but also known through separate written and oral tradition. Thus, though not exhaustive, this source likely contained more detailed instructions for specific types of ritual, particulary, the malevolent rites conducted in the cremation grounds. Perhaps a desire to exalt this Svacchandah form of Bhairavah formed the primary motive behind the compiling of this text. In that case, the Svacchanda stamp of nomenclature, attributes, and the like, could have been secondarily imprinted on an "Ur-Bhairava" source. Tantric texts of this period in many traditions, however, attest to the creation and multiplication of deities as the representation and personification of doctrinal and meditational concepts. 1 Accordingly, the "svacchanda-" epithet, and the details of his form may simply reflect a theological characterization of Bhairavah when viewed as the supreme Shiva, rather than the proper name of a particular minor deity who came to be identified with Bhairavah. 2 Therefore, without 1 So Marie-Therese de Mallmann asserts at the beginning of her Introduction a l'iconographie du tantrisme bouddhique, Bibliotheque du centre de recherches sur l'asie centrale et la haute asie 1 (Paris: Librairie Adrien-maisonneuve, 1975), p.1: "L'on se rappellera tout d'abord que cette iconographie est TOUJOURS le reflet ou l'illustration de notions religieuses ou de concepts philosophiques. En consequence, aucun detail n'est gratuit: mensurations, couleurs, gestes, attitutdes, expressions de physionomie, attributs, etc., ont tous une signification precise." 2 V. supra section 1.1.3 for a discussion of the significance of the attribute Svacchandah.

56 additional evidence, it can not be assumed that Svacchandabhairavah represented the exclusive deity of a region or group, or that organized cults existed devoted to aspects of Bhairavah, among the larger groups of devotees who worshipped Shiva as Bhairavah More probably, a fluid interchange of traditions related to Bhairavah persisted that enabled and encouraged the elaboration of different Bhairava texts and the addition of material resulting in enlarged recensions of Svacchandatantram. 2 Given the lack of information about the early history of the cult of Bhairavah, therefore, discriminating between an "Ur-Bhairava" and a "Svacchanda-bhairava" source at the base of this text constitutes a speculative over-refinement. Throughout the present text of Svacchandatantram, there occur various short rites, meditations, and teachings, which show no intrinsic connection to Bhairavah, and which sometimes reoccur embedded in larger rituals. The meditation on the Pranavah offers 1 In later periods better documentation exists of the historical process in which local and tribal deities became absorbed into the Hindu pantheon by their identification with Bhairavah through the meditation of ascetic worshippers of Bhairavah who circulated through these regions on the periphery of regions of Hindu cultural dominance. V. Gunter-Dietz Sontheimer, Biroba, Mhaskoba, und Khandoba, Schriftenreihe des Sudasien-Instituts der Universitat Heidelberg 21 (Weisbaden: Franz Steiner Verlag, 1976), esp., pp. 95 ff. (For an early period, illustrating the same process, cf. Heinrich von Stietencron, "The Saiva Component in the Early Evolution of Jagannatha," in The Cult of Jagannath and the Regional Tradition of Orissa, ed. Anncharlott Eschmann, et al., South Asia Interdisciplinary Regional Research Programme, Orissa Research Project. South Asia Institute, New Delhi Branch, Heidelberg University. South Asian Studies .8 (New Delhi: Manohar, 1978), pp. 119-123.) 2 This appears likely from the references to a theology of partial manifestations of Bhairavah referred to by the text and by Kshemaraja. (Cf. the preceding section on the Bhairava stratum, and Ksemaraja's reference to other Bhairava aspects and texts, in bk.1, p.10.).

57 the best example. It is described as part of the grand initiation liturgy of the fourth book, and as an independent practice of the adept in the sixth book. 1. As noted previously, practices like the Pranava meditation occur in many of the strata that have influenced the development of the text. But as their immediate source for these short practices, the compilers of Svacchandatantram likely drew upon an early Saiva tradition of ascetics who stood close to the antecedent Pasupata tradition and who had collected their practices for usage by their own nascent sect. The combination of this "early Saiva" source and the Bhairava source marks the first stage in the compositional history of Svacchandatantram. In their work, these compilers effected or rather reflected the association of Svacchanda-bhairava and his specific practices with more generic practices like the Pranava meditation and Bahurupa formula of Aghorah. The structure of the initial anukramanika, and the dialogue frame may date to this time. The assumption of a single early Saiva source likely telescopes a process that stretched over a long period, perhaps even after secondary redaction with different documents. The older, less elaborate content and disparate if not disorganized structure of this early Saiva material, however, tends to corroborate the hypothesis that the long prehistory of its assemblage and evolution had reached some written closure, before the qualitatively distinct next stage in the compositional history of Svacchandatantram. This second stage occurred when compilers incorporated extensive material from the established Saiva tradition into the text of Svacchandatantram. The clearcut and numerous parallels with other extant agamah strongly warrant assuming a redaction 1 Cf. the summaries of bk.4, pp. 159 ff, and bk.6, pp.104 ff. Recognizing these similarities, Kshemaraja provides cross-references to the fourth book in his commentary on the sixth. (V., for example, bk.6, p. 107, p. 109, et al.)

58 " using written documents, a "scriptural Saiva source. Where possible, the redactors welded and incorporated the earlier Saiva material into the later ritual procedures and teaching, or set them side by side, leaving evidence of redaction in redundancy and repetition. From this second source might derive, for example, the Pranava meditation employed as part of the larger initiation ritual. This meditation, in turn, might derive from the same earlier Saiva material preserved in an earlier form in other parts of the text and reintroduced in the initiation ritual in a more developed form. These later compilers apparently made only minor adjustments to the frame dialogue and did not completely update the introductory table of contents, which does not mention, for example, all the classes of Saiva initiates or the types of Saiva initiation. 1 Their primary motivation appears to have been the desire to place Svacchanda-bhairava at the center of an already established Shiva cult of the agamah.2 As evidence of this process. in the institutionalizing of Bhairavah, they substituted the name Bhairavah for Shiva only sporadically and left Shiva in many of the books that describe the standard agamic rituals. In contrast, the name Bhairavah occurs exclusively as the name of the deity in, for example, the cremation ground rites. 1 Thus the introductory table of contents (Bk.1, p. 12, vs.10 b), which refers to the path (adhva), may refer not to the grand initiation liturgy of the fourth book via six paths, but rather to the simpler initiation using the diagram of Svacchanda-bhairava in the fifth book (pp. 11 ff), or to the apparent initiation of the adept into the use of the Bhairava formula summarily indicated in bk.13, pp. 90-92. 2 Cf. Anncharlott Eschmann, "Hinduization of Tribal Deities in Orissa: The sakta and Saiva Typology," in The Cult of Jagannath and the Regional Tradition of Orissa, pp.79-117, for a discussion of the process by which once minor deities become assimilated to mainstream tradition and become the center of elaborate cultic ritual.

59 Svacchandatantram does not contain other important and standard Saiva scriptural sections, e.g. that on the installation of temple images such as the lingam, an omission which may mirror the motives and historical circumstances of its compilers and the receptive limits of the already established text of Svacchandatantram. The other agamah, as well, undoubtedly underwent further revision and expansion, and thus may not be presumed to reflect a state of Saiva scriptural activity prior to that evidenced by Svacchandatantram, which may preserve earlier features of this tradition. 1 Or, the omission or incomplete treatment of many topics by the compiler of Svacchandatantram, may be interpreted as indicating that an oral tradition accompanied the text, or similarly that other Saiva scriptures complemented the text in these areas, and obviated the need for absolute exhaustiveness. The tenth through twelfth books of Svacchandatantram, which contain several overlapping treatments of the worlds (bhuvanani) and planes (tattvani) that compose the universe, provide a notable example of the way in which the text preserves early stages in the development of Saivism. Here the process of combining material from Samkhya, Puranic, and Pasupata sources with properly Saiva categories and meditational schemes, remains visible in several unpolished and only partially standardized versions. Thus these sections illustrate the process through which compilers and redactors elaborated the Saiva cosmology, an historical process that elsewhere in this and other texts often appears collapsed into two chronologically discontinous stages: that of the early Saiva and that of the scriptural Saiva. 1 For example, the purificatory initation contained in the middle of the account of the cosmology may be a survival of an earlier and simplier initiation rite corresponding to a simpler worldview. (V. the summary of bk.10, pp. 141 ff.)

60 Thus the addition of scriptural Saiva material, which placed Bhairava at the center of the agamic cult, made Svacchandatantram into a Saiva scripture, and promoted later commentaries both from the dualistic Siddhantin school and the monistic school represented by Kshemaraja. Nevertheless, the later redaction of the text apparently did not parallel that of the other agamah. Instead of only revising and adding material related to agamic ritual, redactors inserted material of a more "tantric" character at similar and thus appropriately receptive points in the text. 1 Although successive redactors, drawing on different documents and traditions, likely engaged in this supplementation, given the absence of information identifying individual hands, this material may be collectively attributed to a "tantric source. * The accretion of Svacchandatantram through this phase of supplementation from the tantric source probably continued interwoven around the next discernible major redaction. Here, in this phase, redactors affixed additional sections of a specific Bhairava character to the text. The finished quality and self-contained structure of these sections, added to internal indicators of conflation, differentiate them from the primary Bhairava source, as a "later Bhairava" source, as it were. 2 This later Bhairava source actually probably comprised a number of distinct documents, but, once again, given the impossibility of ascertaining the number of redactors or redactional steps, for convenience, a single label may suffice. More importantly, these additions demonstrate the continued life of the text of Svacchandatantram within circles of Bhairava devotees, and 1 V., for example, the summary of bk 7, which examines the elaborate meditation on time that has been apparently inserted in the middle of simpler practices. 2 V., for example, the summary of bk.13, pp. 95, ff, for an apparently supplemental collection of Bhairava practices.

61 hypothetically then, in continuity with the milieu of the primary document. Within this same line, and probably close after this redaction using the later Bhairava source, occurs the last major stage in the compositional history of Svacchandatantram that brought the text close to the form of its final recension. At this time, redactors from the milieu of the last stratum discussed, the Kaula, appended well formed pieces of their own tradition to the text. 1 Apparently, by this time, the structure of Svacchandatantram had hardened to a degree that precluded a rifacimento to allow better integration of the segments from this "Kaula" source. Instead the redactors fastened them as appendices with only minimal editorial links. The editing into books likely followed the secondary swelling by conflation of the text of Svacchandatantram.² Perhaps this divsion even preceded the later redactors who appended the material from the Kaula source, and led them to append their material in the form of another book. This hardening in the structure of Svacchandatantram coincides with the formation of its now extant recension and with its entrance into the final stage of its redaction. Here the act of editorial retouche, which as reflected by devices such as cross-references also ran concurrently with the conflation and assimilation of the major sources, becomes predominant. Consequently, circumscribed interpolation can be assumed to have replaced expansive supplementation. These redactors not only added material but also in the course of rectifying the language of the text, likely deleted 1 V. the summary of bk.15. 2 As an example of the secondary editing into books, the unitary topic announced in the dialogue at the beginning of bk.7, (p. 167) time (kala-), and the portion (-amsakam), has been broken into two parts, with the second topic discussed in the eighth book. This separation may have occurred due to the increase by interpolation of the seventh book.

62 material. Perhaps these deletions went beyond the substitution of standard forms and even phrases to include substantive excisions and abridgements. 1 Although, naturally, no direct evidence exists for such large scale cutting, Ksemaraja's commentary provides ample witness that the manuscripts which had reached him carried a text not only variant but also somewhat corrupted, fragmented, damaged and needing reconstitution. 2 Evidently, a chronological as well as a socially qualitative gap still separated the last anonymous editors from the sectarian commentators. These pre-commentatorial redactors or editors, thus, partially revised the language, harmonized the various books, and interpolated verses and smaller sections reflecting their own doctrinal position. 3 The label "Saiva redactorial" source might best cover their collective imprint where discernible as affecting the extant recension of Svacchandatantram. As with preceding sources, the documents that may have served as the paradigm for their operations on Svacchandatantram, remain obscure and only hypothetically 1 Cf., for evidence of deletion in the Puranic literature, Rocher, The Puranas, pp. 37-38. For example, many authors on dharmah, both commentators and nibandhakaras, quote verses from the Puranas which can not be found in any extant version. 2 V. bk. 13, p. 120, where Kshemaraja, in a rare account of his commentatorial activity states: "And thus, as a rule, the interpolation of other books, the reversal (of the order] of books, and the reversal of readings of this book, [that have] a hundred branches, [which] are seen, are fabricated by the ignorant. That is removed by us through the investigation of ancient books, as regards the sense . (evam ca prayaso granthantarapraksepo granthaviparyasah pathaviparyasasca asya granthasy durmedhobhih parikalpitah sataksakho drsyate/sa 'smabhih puratanapustakanvesanato yavadgati apasarita ........)" " 3 V., for example, bk.4 p.13, where an editor has apparently added a verse marking the end of the daily ritual ("nityakarma") and the beginning of the occasional ritual ("naimittikakarma"), using these terms that do not appear elsewhere in the text except in the commentary of Kshemaraja.

63 identifiable, since the text never refers to any other scriptures by name or even scriptural traditions.