You are on page 1of 470


Shaman Hatley

a dissertation in Religious Studies

Presented to the Faculties of the University of Pennsylvania in Partial Fulllment of the Requirements for the Degree of Doctor of Philosophy 2007

Dissertation Supervisor

Graduate Group Chair

The present study would have been impossible without the support of numerous individuals and institutions. Several years of my graduate studies at the University of Pennsylvania were funded by the university itself, while the Muktabodha Indological Research Institute and the American Institute of Indian Studies provided the fellowships necessary for conducting dissertation research. The departments of Religious Studies and South Asia Studies at Penn both provided me travel grants at various times, and sustained my intellectual journey in numerous other ways. For access to manuscripts and other research materials, I am grateful to the curators and staff of the National Archives of Kathmandu, Asiatic Society of Bengal, Department of Sanskrit (Manuscript Division) of Vi vabh rat University, Bhandarkar Oriental s a Research Institute, Van Pelt Library, Nepal Research Center, and AIIS Center for Art and Archeology. I am grateful to the staff of the AIIS ofce in Pune for their logistical support as well, and particularly to Prof. S. S. Bahulkar of the Tilak Maharashtra Vidyapeeth, who arranged for my afliation with the t.m.v. and graciously served as my academic advisor in India. My studies at the University of Pennsylvania have indebted me to numerous teachers; beyond the dissertation committee, I owe special thanks to Professors Aditya Behl, George Cardona, Stephen Dunning, the late (and much missed) Wilhelm Halbfass, Christian Novetzke, and Ludo Rocher. I am particularly indebted to the latter, who rst taught me Sanskrit and supervised my m.a. studies. I am also much obliged to Prof. Alexis Sanderson of the University of Oxford, with whom I had the privilege of reading part of the Brahmay mala at an early stage in my research, and a whose guidance has been invaluable. Professors Michael Meister and Guy Welbon, members of the dissertation committee, have been continual sources of knowledge and encouragement over nearly a decade, and I am grateful for their feedback on the dissertation at various stages in its formation. Most of all, I wish to thank Prof. Harunaga Isaacson, from whom I have learned much of what I know; I am lost for ii

iii words to express my gratitude for his years of painstaking mentorship. Without his detailed feedback on drafts of this thesis, it would have been a much inferior work. Several others have read and made valuable remarks on sections of the dissertation, especially Alberta Ferrario, Ayesha Irani, Prof. Kathleen Kesson, Prof. John Nemec, and Dr. Judit Trzsk. I have also beneted from the opportunity to present and discuss work drawn from or connected with the dissertation at the AIIS Junior Fellows Conference in Gurgaon (January 2006), the University of Virginia at Charlottesville (November 2006), the Workshop on Early Saivism at Pondicherry (January 2007), and Concordia University (January 2007). The present study would have been much poorer but for the increasing availability of searchable electronic texts. In particular, I am indebted to Dr. Somadeva Vasudeva and Prof. Harunaga Isaacson for sharing extensive material with me, including transcriptions of manuscripts of unpublished sources; to Dr. Judit Trzsk, who provided me an electronic copy of her edition of the Siddhayogevarmata; and s to Prof. Alexis Sanderson, for providing me draft editions of portions of the Jayadrathay mala, including its entire third book (the Yoginsac raprakarana). As a student a a . of Saivism, I am also indebted to Dr. Mark Dyczkowski, who has made available numerous transcriptions of unpublished tantric texts through the Digital Library of the Muktabodha Indological Research Institute. His collation of manuscripts of the Tantrasadbh vatantra has been of signicant value in this study. I also thank Prof. a Vidya Dehejia, who gave permission to reproduce one of her photographs (figure
A 2.6). The typesetting of the critical edition of the Brahmay mala using L TEX was in no a

small measure enabled by technical help from Dr. Somadeva Vasudeva, from whom I have learned much concerning Saivism as well. Finally, I owe an incalculable debt to my family, who have tolerated a great deal so that I could complete this thesis, and given me much joy in the process. This work is dedicated to them.


THE BRAHMAYAMALATANTRA AND EARLY SAIVA CULT OF YOGINIS Shaman Hatley Dissertation supervisor: Harunaga Isaacson

The present thesis comprises a study of the Brahmay malatantra, a scripture of early mea dieval tantric or esoteric Saivism, and its cult of yogins: ying, shapeshifting female deities whose occult powers practitioners sought in visionary, transactional encounters. Composed prior to the ninth century, and perhaps considerably earlier, this unpublished work of onehundred and one chapters and more than 12,500 verses constitutes one of the most signi cant sources for the study of early Saiva ritual and goddess cults. After introducing the text, the tradition, and the gure of the yogin, chapters 2 and 3 review the extant literary, art historical, and epigraphic sources concerned with yogins, with a focus on the background and early development of their Saiva cult. It is within this context that the Brahmay mala a is situated. Particular problems addressed include the relationship between the yogin cult of the Brahmay mala and the Brahmanical Mother goddesses (m tr), the post ninth-century a a. temples of yogins, early tantric Saiva literature, and the Buddhist yogintantras. Chapter 4 investigates the Brahmay malas form, textual strata, provenance, and social and geographic a horizons, while chapter 5 examines the position the text articulates for itself within the Saiva tradition. Part II of the dissertation consists of critical editions and translations of several chapters of the BraY , which appear in print for the rst time. a



List of Tables vii List of Figures viii

I History and Sources

1 Introduction 1.1 1.2 1.3 2 2

The Brahmay malatantra and Cult of Yogins . . . . . . . . . . . . . . . . . a

The Brahmay mala in Scholarship . . . . . . . . . . . . . . . . . . . . . . . 23 a A Note on Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

Early Evidence for the Cult of Yogins: the Literary, Sculptural, and Epi graphic Sources 31 2.1 2.2 2.3 2.4 2.5 Introduction: Yogins and Mother Goddesses . . . . . . . . . . . . . . . . 31 Early Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 The Later Literary Evidence . . . . . . . . . . . . . . . . . . . . . . . . . . 73 Temples of the Yogins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

The Cult of Yogins and its Background in Early Tantric Literature 131 3.1 3.2 3.3 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 The Background of the Yogin Cult in Tantric Saiva Literature . . . . . . 133 Scriptures of the Saiva Yogin Cult: the Vidy ptha and Kaula . . . . . . 153 a . v

vi 3.4 3.5 Yogins in Early Buddhist Tantric Literature . . . . . . . . . . . . . . . . 162 Buddhist and Saiva Yogintantras: the Case of the Brahmay mala and a Laghucakraamvaratantra . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 s . 3.6 4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187

The Content, Structure, and Provenance of the Brahmay mala 190 a 4.1 4.2 4.3 Introduction: Select Topics in the Brahmay mala . . . . . . . . . . . . . . 190 a Structure and Textual Strata . . . . . . . . . . . . . . . . . . . . . . . . . . 200 On the Provenance of the Brahmay mala . . . . . . . . . . . . . . . . . . . 211 a

5 To Name a Tantra: Identity, History, and the Saiva Canon in the Epithets of the Brahmay mala 237 a 5.1 5.2 5.3 5.4 5.5 5.6 Brahmay mala . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238 a Picumata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243 Nav ksaravidh na: the Method of Nine Syllables . . . . . . . . . . . . . 251 a . a Mulatantra: the Root Scripture . . . . . . . . . . . . . . . . . . . . . . . 258 Dv daas hasraka: the Tantra of Twelve-Thousand Verses . . . . . . . . 273 a s a Vimala and the Ucchusmatantra . . . . . . . . . . . . . . . . . . . . . . . . 275 .

6 Conclusion 282

II The Brahmay mala: A Critical Edition and Translation of a Chapters i, ii, lv, lxxiii, and xcix
Appendices 424 A Chapter Colophons of the Brahmay mala in its Oldest Nepalese Codex (nak a 3-370) 425 Works Cited 434


List of Tables

3.1 4.1 4.2

A parallel passage in Brahmay mala lxxxiv and Laghuamvara 26 . . . . . . . 178 a s . Forms of Bhairava in the Brahmay mala . . . . . . . . . . . . . . . . . . . . . 192 a Siddh ntatantras of the Middle Stream (madhyamasrotas) according to BraY a a xxxviii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217

4.3 4.4 4.5 4.6

Tantras of the Left Stream (v masrotas) and Lower Stream (adhahsrotas) . . . 218 a . Bhairavatantras of the Right Stream (daksinasrotas) . . . . . . . . . . . . . . . 218 . . A parallel passage in the Brahmay mala and Niv satattvasamhit . . . . . . . 219 a s a . a The Nav tman Mantra in the Niv sottara and the Vidy r ja of the Svaca s a aa chandatantra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220

4.7 4.8 4.9

The Nine Tattvas in the BraY . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 a Individuals mentioned in BraY i . . . . . . . . . . . . . . . . . . . . . . . . . 230 a The Eight Cremation Grounds in BraY iii . . . . . . . . . . . . . . . . . . . 233 a

4.10 The Eight Sacred Mounds (p. ha) in BraY lxxxiv . . . . . . . . . . . . . . . 234 t a 5.1 5.2 Deities of the Nav ksar Vidy . . . . . . . . . . . . . . . . . . . . . . . . . . . 252 a . a a The Ten Streams of Revelation according to BraY xxxviii . . . . . . . . . . 267 a


List of Figures

1.1 1.2 1.3

National Archives of Kathmandu ms 3-370, folios 3v4r. . . . . . . . . . . . nak 3-370, detail of folios 3v and 4r, left third. . . . . . . . . . . . . . . . . . Siva enshrined within a circle of sixty-four yogins. Yogin temple, R npur a

4 4

Jharial, Orissa. AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . 12 2.1 Mother goddess bearing lotus and infant, making the abhayamudr . Kusaa . na-era Mathur . AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . 38 a . 2.2 Skanda (left) and ve m trs. Kusana-era Mathur . State museum, Mathur . a. a a . . AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 2.3 2.4 Kubera and two m trs. Kusana-era Mathur . AIIS Photo Archive. . . . . . . 40 a. a . . Seven m trs, with Skanda (left). Kusana-era Mathur . State museum, a. a . . Mathur . AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 a 2.5 Bird-headed Mother goddess carrying an infant in her likeness. Kusana. . era Mathur . State museum, Mathur . AIIS Photo Archive. . . . . . . . . . . 43 a a 2.6 Horse-headed yogin from Lokhari, U.P., with like infant. Photograph by Vidya Dehejia, published in Dehejia, Yogin Cult and Temples, 159. . . . . . . 43 2.7 C munda and Vin yaka/Gane a, R me vara Cave (no. 21) m tr shrine, a a a s a. . . . s Ellora. Sixth century. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 2.8 Siva and the Seven Mother goddesses (detail). Saptam tr cliff shrine bea. tween Badoh and Pathari, Madhya Pradesh. AIIS Photo Archive. . . . . . . . 59 2.9 Saptam tr Panel, with Vrabhadra (left) and Gane a (right). State Mua. . s seum, Ashapuri, Madhya Pradesh. AIIS Photo Archive. . . . . . . . . . . . . 59 viii

ix 2.10 Yaksa at pillar base, Nagarjunakonda. National Museum, New Delhi. AIIS . Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 2.11 Theriomorphic gana or dv rap la. Deorani temple (circa 6th century); Tala, a a . near Bilaspur, M.P. AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . 62 2.12 Head of a fanged, large-eared yaksa. Pawaya, Gwalior, M.P. 5th6th cen. tury. AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 2.13 Mah k la and K l? Ellora, R me vara cave temple, west wall of Mothera a a a s goddess shrine. AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . 71 2.14 Siva Gajasamh ramurti, Vait l Deul temple, Bhuvanesvar. AIIS Photo a . a Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 2.15 Frontal view of the Hr pur yogin temple. AIIS Photo Archive. . . . . . . . . 112 a 2.16 Hr pur yogin with damaru drum, standing upon wheel and rat (?). AIIS a . Photo Archive (detail). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

2.17 Hr pur yogin on lotus and elephant, drinking from a skull-bowl. Photoa graph by Don Stadtner, from the AIIS Photo Archive. . . . . . . . . . . . . . 114

2.18 One of nine k ty yans (?); Hr pur yogin temple, to the right of the a a a entrance, exterior facade, standing upon severed head. AIIS Photo Archive. 115 2.19 Lion-faced yogin, Hr pur temple. Photograph by Don Stadtner; included a in the AIIS Photo Archive. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 2.20 Ekap dabhairava, Hr pur yogin temple. AIIS Photo Archive. . . . . . . . . 116 a a 2.21 C munda, Hr pur yogin temple. AIIS Photo Archive (detail). . . . . . . . . 116 a . . a 2.22 Veneration of sandals representing a yogin, depicted below a yogin im ages pedestal. Bher gh t yogin temple. AIIS Photo Archive. . . . . . . . . . 119 a a 5.1 The Origin of Scripture (Sivatantrotpatti) according Niv satantra, Uttarasutra s a 1.2325 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260 5.2 5.3 The Descent of Scripture (astr vat ra) according to Svacchandatantra 8 . . 261 s a a The Descent of Scripture (astr vat ra) according to Brahmay mala i and s a a a xxxviii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263


BraY a nak ngmpp ksts tss aiis ms, mss cod. corr. em. conj. (. . . )err.
ac pc

Brahmay mala a National Archives of Kathmandu Nepal-German Manuscript Preservation Project Kashmir Series of Texts and Studies Trivandrum Sanskrit Series American Institute of Indian Studies manuscript(s) reading of the codex in question correction emendation conjecture syllables marked by the scribe as erroneous ante correctionem post correctionem folio(s); r = recto, v= verso (e.g. f. 4r) identies text for which the reading adopted seems insecure text passage deemed corrupt, for which no conjecture is offered conjectural translation a marginal or interlinear insertion syllables only faintly legible

f., ff.

. . . () . . . (?) +. . . + (. . . )

xi Bya A Byb B Byc C Byd D Bye E Kjncod Kjned Stksts Stcod Tsk,kh,g Brahmay mala codex nak 3-370 (in part i) a Brahmay mala codex nak 3-370 (in the edition and translation) a BraY codex nak 5-1929 (in part i) a BraY codex nak 5-1929 (in the edition and translation) a BraY codex nak 1-143 (in part i) a BraY codex nak 1-143 (in the edition and translation) a BraY codex nak 1-286 (in part i) a BraY codex nak 1-286 (in the edition and translation) a BraY codex nak 6-2608 (in part i) a BraY codex nak 6-2608 (in the edition and translation) a Kaulaj nanirnaya codex nak 3-362 (ngmpp reel a48/13) a . P. C. Bagchis edition of the Kaulaj nanirnaya (see bibliography) a . Svacchandatantra, ksts edition Svacchandatantra, ms nak 1-224 Tantrasabh va mss ka, kha, and ga, as reported by Dyczkowski a

Part I History and Sources

Chapter 1

1.1 the brahmayamalatantra and cult of yogins i

On the eighth day of the waxing moon of the month of M gha in the year 172, a Nep l-samvatSunday, 12 January, 1052 c.e.a certain Jay karajva, who resided in a a the vicinity of Kathmandus Pa upatin tha temple, completed copying a Sanskrit text s a called, among other names, the Brahmay mala. This endeavor had undoubtedly oca cupied him a long while, for the text lls three hundred and fty-eight long, doublesided palm-leaf sheets (figures 3.12). Fortunately, the product of his labors found its way into the manuscript collections of Nepalese royalty, and is today housed in the National Archives in Kathmandu.1 But for this, precious little knowledge of a vast and important pre ninth-century work would have been possible,2 a fate which has befallen all too many works of tantric literature. The other extant codices of the Brahmay mala (hereafter BraY ), also of Nepalese provenance, all appear to descend a a from this eleventh-century manuscript.3 Although the scripture itself was not composed in Nepal,4 no complete manuscript appears today to survive outside of this
nak accession no. 3-370. See the mss description in the introduction to the critical edition. Luciano Petech records the colophon and date of the manuscript as follows: Ms. Brahmay mala, National a Archives, iii.370. Colophon: samvat 100-50-2 M gha-ukl s. amy m adityadine R.-P.-r-Baladeva-r jye sra s a. t a. s a Paupativ stavya sr-Jay karajvena Brahmay malam n ma sastram likhitam. Written at Pa upati N th. The s a a . a a s a . . . date is veried for Sunday, January 12th, 1052. Petech, Mediaeval History of Nepal (circa 7501482), 2nd ed., 44. (Petechs samvat 100-50-2 is a typographical error for samvat 100-70-2; the ms reads samvat a cu 2.) 2 The dating of the Brahmay mala is addressed in chapter 4, section 3. a 3 Manuscripts of the BraY are described and discussed in the introduction to the critical edition. a 4 The question of the provenance of the BraY is taken up in chapter 4. a

3 region. The BraY designates itself a tantra: a treatise, ostensibly of divine origin, cona sidered authoritative scripture within the cult of Siva that is based on the tantras (t ntrika)i.e. Tantric Saivism. The most signicant emic term for this tradition a is the Mantram rga or Way of Mantras.5 It should not be inferred from the a paucity of extant manuscripts that the BraY was an unimportant or little-known a work of tantric literature. A glimpse of the authority it once commanded may be had in the writing of the Kashmiri polymath Abhinavagupta (. circa 1000 c.e.), who cited the scripture often despite its degree of cultic and doctrinal remoteness.6 Somewhat like the Rudray mala, the actual text would largely be forgotten;7 yet its aura a of authority would persist, on the basis of which there would emerge several new Brahmay malas. I am aware of ve: a South Indian text connected with the cult of a Bhadrak l, in which some traces of the older BraY are discernable;8 another South a a
5 On the term Mantram rga and its scope, see Alexis Sanderson, Religion and the State: Saiva Ofa ciants in the Territory of the Kings Brahmanical Chaplain, 229 (n. 1). On the term tantra, see Sander son, Saivism and the Tantric Traditions, 66162. It is noteworthy that in addition to Mantram rga, a the BraY in one instance uses the term Tantram rga in what appears to be a comparable sense. BraY a a a xcv.29cd30ab:

tantram rgaprasiddhena vidhin varttayan sad a a a pr pnuy ta mah devi siddhayo manasepsit h | a a a a.
psit h a a a . ] em.; psit By


Ever repeating [the mantra] with the method established in the Way of Tantras, O Great Goddess, one would obtain the siddhis the mind desires. (pr pnuy ta appears to represent the correct optative, pr pnuy t, metri causa, while siddhayah is noma a a a . inative in form but accusative in sensea common phenomenon in the BraY .) Harunaga Isaacson a (personal communication, September 2007) draws my attention to the fact that the Puracary rnava s a . refers to Tantric Saivism as the Way of Tantras as well, in 1.149ab: vedam rgam parityajya tantram rgaikaa a . tatpar h, [those who,] abandoning the Way of the Vedas, are devoted exclusively to the Way of Tantras a. . . . . 6 For a discussion of some of Abhinavaguptas references to the BraY , see chapter 4, section 3 (n. a 57). 7 On the Rudray mala, of which no old version survives but which became the foremost locus of a a ascription in Hindu Tantric literature, see Goudriaan, Hindu Tantric and S kta Literature, 4748. It is possible that the old BraY continued to have a place in Nepalese tradition into the early twentieth a century; Harapras d S str, who more than a century ago cataloged the oldest codex of the BraY , a a a reports on views of the text then current in Nepal: in the opinion of the Pandits of Nepal the full .. texts of Brahma Ymala is a lakh and a quarter of slokas, and that it belongs to all the six schools of Tantra. The present work, extending over 1200 [sic; 12,000?] slokas, belongs, however, to the western school. A Catalogue of Palm-leaf and Selected Paper Manuscripts Belonging to the Durbar Library Nepal, vol. a ii, reprinted in Reinhold Grnendahl, A Concordance of H. P. S stris Catalogue of the Durbar Library and the Microlms of the Nepal-German Manuscript Preservation Project, 61. 8 This text is preserved, incomplete, in a Devan gar transcription in the collection of the Institut a

Figure 1.2: nak 3-370, detail of folios 3v and 4r, left third.

Figure 1.1: National Archives of Kathmandu ms 3-370, folios 3v4r.

5 Indian Brahmay mala related to this of which only a few chapters survive;9 a short a text preserved in a Bengali manuscript expounding a series of ritual diagrams (cakras or yantras), with no discernable relation to the older BraY ;10 a text of the cult of T r a aa by this name transmitted in an untraced Bengali manuscript, a section of which has been published;11 and a Brahmay mala preserved in a single, fragmentary Nepalese a ms, which though eclectic, draws directly from the older BraY .12 Additionally, the a BraY has been the locus of ascription for several hymns of praise (stotra),13 and a
Franais de Pondichry (manuscript T. 522), copied from a manuscript from Tirukkalukkunram in Tamil Nadu. Very recently, this text has been discussed by Alexis Sanderson, who provides evidence for its South Indian provenance and shows several respects in which it demonstrates continuity with the older Brahmay mala. Sanderson, Arthavavedins in Tantric Territory: the Angirasakalpa Texts of the a Oriya Paippal dins and their Connection with the Trika and the K lkula. With critical editions of the a a Par japavidhi, the Par mantravidhi, and the *Bhadrak lmantravidhiprakarana, 277 78. a a a . 9 This text, transmitted in Trivandrum University Library ms no. 982, is mentioned by Sanderson, ibid.; he reports that it breaks off in the fth chapter. I have been unable to consult the text thus far. That it concerns the cult of Bhadrak l is suggested by passages Sanderson cites in ibid., 278 (n. 143). a 10 Asiatic Society of Bengal manuscript g6392. 11 Two chapters of this text were published as an appendix in Girsacandra Ved ntatrtha (ed.), a T r tantram. Srgirscandraved ntatrthasankalitam. With an Introduction by A. K. Maitra. By all appearaa a ances from a much later period, this Brahmay mala seems unrelated to the older text of the same a name. The ms on which the edition is based apparently comes from the collection of the Varendra Research Museum Library. Maitra, introduction to ibid., 22. However, this might no longer be available, for the archives relevant catalog of manuscripts makes no mention of it. Several Brahmay mala manuscripts are described, but none appear likely to be the basis for the published text: a a a number 279 transmits the Srr dh kavaca in the Brahmay mala (the colophon reads iti srbrahmaa y male haragaursamv de sr r dh kavacam sampurnam). Number 250, of sixteen folios, consists of a a . a a a . . . . stotra of the goddess T r (the colophon reads iti srbrahmay male t r devy h s mr jye samj m [?] aa a aa a. a a . a. tak r disahasran mastotram sam ptam). Number 310 (six folios) is the Caitanya-kalpa of the BraY aa a a . a . (iti brahmay male sivadurg samv de caitanyakalpalm. Number 303 (ve folios) consists of the Gt s ra a a . a aa . attributed to the BraY (no colophon is provided). Sachindra Nath Siddhanta, A Descriptive Catalogue a of Sanskrit Manuscripts in the Varendra Research Museum Library, vol. I, 22829. The material published from the Varendra BraY codex narrates the journey of the Brahmanical sage Va istha to Mah cna, a s .. a where the inebriated Buddha instructs him in the Chinese method (cn c ra) for the worship of T r . a a aa As Joel Bordeaux brought to my attention (personal communication, May, 2007), this has close parallels in chapters 12 of the Mah cn c ratantra, a text published by Marion Meisig. Die China-Lehre Des a a a Saktismus: Mah cn c ra-Tantra, kritisch ediert nebst bersetzung und Glossar. a a a 12 nak 1-1557 (ngmpp reel a165/16). This undated paper manuscript, in Newari script, consists of twenty-one folios; the scribe appears not to have completed his task, or else to have copied from an incomplete exemplar. The text calls itself by several of the titular epithets of the old BraY (on which see a chapter 5): Picumata, Nav ksaravidh na, and, of course, Brahmay mala. The text breaks off in the twelfth a . a a chapter, unfortunately. It directly draws upon or echoes the old BraY pervasively, in effect comprising a a digest of the latter. Its seventh chapter, for instance, incorporates material from BraY liv and lv on a the subject of Mah k la and the Mothers, and secret signs (chomm ). The old BraY does not appear to a a a a be its only source, however, for this chapter contains material related to the Tantrasadbh va as well: a a passage from ff. 14r14v has close parallels in Tantrasadbh va 18.1825. a 13 Theodor Aufrecht makes reference to at least fteen distinct works having the BraY as locus of a ascription, mostly apotropaic hymns (the kavaca genre) and hymns of the thousand names genre of panegyric (sahasran mastotra). Aufrecht, Catalogus Catalogorum. An Alphabetical Register of Sanskrit Works a and Authors, vol. i, 382; vol. ii, 87; and vol. iii, 81. See also n. 11 above. I have had no opportunity to

6 possibly more literature that has not come to my attention. Though surviving in just a handful of manuscripts, the old BraY has in fact fared a better than most early Saiva scriptures. As Dominic Goodall shows, a relatively small number of the early tantras of the Saivasiddh nta survives; furthermore, many of the a texts listed in ancient descriptions of the tantric Saiva canon exist only in versions of post eleventh-century, South Indian provenance.14 And little at all survives of the scriptures of the proto-tantric cults of the Saiva Atim rga, the Path Beyond, which a apparently formed the immediate background of the Saiva Mantram rga.15 a Tantric traditions have been characterized in popular culture and no small amount of scholarship as quintessentially concerned with goddesses and/or ritualized forms of sexual activity.16 But that there is no intrinsic connection with either, and that the two are not concomitant, should in fact be obvious; indeed, goddesses and ritual coitus have little or no place in several of the major, early tantric traditions. Tantric Buddhist practice systems based on the yogatantras and earlier literature infrequently accord high cultic status to female deities, while sexual ritual appears to have been a comparatively late development.17 Extant sources of the Vaisnava tradition of the .. P car tra appear to give no place to ritualized sexuality, while the Laksmtantraa a a . work of the early second millennium shown to have been inuenced by nondual consult the manuscripts of these works; but with titles such as Suryakavaca, R makavaca, and Caitanyaa kalpa, it seems most improbable that any of them is connected to the old BraY . a 14 Goodall, introduction to Bhat. a R makan. has Commentary on the Kiranatantra. Vol. 1: Chapters 16, .t a .t . xxxvili. 15 On the division of Saivism into Atim rga and Mantram rga traditions, see Sanderson, Saivism a a and the Tantric Traditions, 66469; and Sanderson, The L kulas: New Evidence of a System Intermea diate Between P c rthika P supatism and Agamic Saivism, passim. Sanderson discusses what can a a a be gleaned concerning lost scriptures of the Atim rga in ibid., and History through Textual Criticism a in the Study of Saivism, the Pacar tra and the Buddhist Yogintantras, 2931. a 16 To catalog the varied guises in which such visions of the tantric traditions have and continue to surface is no small task; Hugh Urbans recent monograph tracing the genealogy of modern understandings of Tantra provides a useful starting point: Tantra: Sex, Secrecy, Politics and Power in the Study of Religion. 17 Coitus and the ritual engagement with other varieties of impurity appear to have their earliest attestation in the Sarvatath gatatattvasamgraha. See Steven Weinberger, The Signicance of Yoga Tantra a . and the Compendium of Principles (Tattvasamgraha Tantra) within Tantric Buddhism in India and Tibet, . 197200. These elements take on added importance in the Guhyasam jatantra and related works, literaa ture in some respects transitional between the yogatantras and yogintantras, and classied accordingly as mah yogatantras. See Anthony Tribe, Mantranaya/Vajray na: tantric Buddhism in India, 21213. In a a the present thesis, chapter 3, section 4, discusses the emergence of goddess cults in Tantric Buddhism.

7 ist Saivismseems exceptional in according theological preeminence to the divine feminine.18 And in the early pan-Indian and living South Indian traditions of the Saivasiddh nta, female deities have subsidiary roles, while sexual ritual is little ata tested.19 Arguably, these represent the predominant tantric traditions ourishing in South Asia and the lands of its cultural inuence through much of the early medieval period.20 Yet tantric traditions did develop that centered upon goddesses, including ones which harnessed sexuality as an important element in a wide range of practices focused upon achieving occult powers (siddhi) and liberation (mukti, moksa). The . BraY represents such a tradition. It in fact comprises one of the most consequena tial sources of evidence for early tantric Saiva goddess cults, while sexuality has a signicant presence in its systems of ritual. In a model of the canon of Saiva scripture advanced in chapter thirty-eight of
18 Alexis Sanderson argues for the inuence of the Pratyabhij hrdaya of the Kashmiri author a . Ksemar ja, as well as other Saiva works, on the Laksmtantra, in History through Textual Criticism, a . . 35 36. On the Laksmtantra and the role of goddesses therein, see Sanjukta Guptas introduction to her . translation, Laksm Tantra: A P car tra Text. Translated with an Introduction and Notes. a a . 19 On the comparative insignicance of goddesses in the Niv satattvasamhit , an early siddh ntatantra, s a a . a see chapter 3, section 2. Interestingly, sexual ritual does have limited attestation in some of the earliest siddh ntatantras, though even in these it appears marginal. See Sanderson, review of N. R. Bhatt, ed., a Matangap ramevar gama (Kriy p da, Yogap da et Cary p da), avec le commentaire de Bhat. a R makan. ha, a s a a a a a a a .t .t and of Bhatt, ed., Raurav gama. dition Critique, introduction et notes, 565; and also chapter 3, section 2, a in the present thesis. 20 Nonetheless, much of the scholarly literature has assumed an articial distinction that, at times, goes so far as to exclude the Vaisnava P car tra and/or Saivasiddh nta from the category of Tantra. a a a .. Note for instance David Whites recent remarks: a number of works that closely resemble the Tantras in their ritual focus call themselves Agamas or Samhit s. These are works belonging to the a . . . . Saivasiddh nta and (Vaisnava) P car tra schools, respectively. Kiss of the Yogin: Tantric Sex in a a a .. its South Asian Contexts, 17. This view, which suggests that the Saivasiddh nta and P car tra are a a a marginal to the study of tantric traditions proper, is based in part on what Dominic Goodall points out is an articial distinction between tantras and agamas: in fact, early Saiva scriptural sources, whether siddh ntatantras or e.g. bhairavatantras, designate themselves by both these terms, and samhit is attested a . a as well. Goodall, introduction to Bhat. a R makan. has Commentary on the Kiranatantra. Vol. I: Chapters .t a .t . 16, xxxvixxxix. That the P car tra scriptureswhich often do refer to themselves as samhit s a a . a are also, by their own designation, tantras, requires little investigation; note, for instance, the text title Laksmtantra (see above), and e.g. Ahirbudhnyasamhit 11.62ab: pacar tr hvayam tantram moksaikaa a . . a . . . phalalaksanam ([Visnu created] the Tantra called Pacar tra, characterized by having liberation as its a . . .. sole aim). Cf. Y mun c ryas Agamapr m nya, which refers to the status of the P car tra tantras as a a a a a. a a authoritative scripture ( pacar tratantrapr m nya) in the prose immediately following the introduca a a. tory verses. Marginalization of the P car tra and Saivasiddh nta from the study of tantric traditions a a a in effect reinforces the characterization, alluded to above, of Tantra as being intrinsically concerned with goddesses and/or sexuality, for the other most inuential tantric traditionsTantric Buddhism (in its latter phases), and non-Saiddh ntika tantric Saivismprovide much greater attestation of goddess a cults and trangsressive ritual. See also n. 32 below.

8 the BraY , tantras are organized into three primary scriptural streams (srotas): the a siddh ntatantras of the middle stream, distinguished by their focus upon the deity a Sad siva; the v matantras of the leftward stream, regarded as scriptural authorities a a for the cult of the Sisters of Tumburu; and the rightward stream of the bhairavatantras. The BraY identies itself with the latterscriptures of the cult of Bhairava, a Sivas manifestation as the archetypal skull-bearing ascetic or kap lin. Furthermore, a an important early classication of the bhairavatantras attested in the BraY subdia vides these scriptures into four seats or mounds (p. ha): those of ritual gestures or t mudr s; of mandalas, diagrammatic representations of the deities; of [male] mantras; a .. and of vidy s, the lores or [female] wisdom mantras.21 Mantras are gendered, for a they are divinitiesnot infrequently referred to as mantra-deities (mantradevat ).22 a This ontological identication of efcacious sonic formulae with divinities is distinctive to the tantric traditions. The division between the Mantraptha and Vidy ptha is . a . in fact one between male mantra-deities and the female vidy -goddesses. It is within a the Vidy ptha of the bhairava-stream of revelation that the BraY situates itself, an a . a acknowledgement of the preponderance of feminine divinities in the ritual systems it advances. While a variety of terms designate these goddesses, foremost are the synonyms yoginthe feminine equivalent of yogin, i.e. yogiand yoge/yogevar: s s female master of yoga.23 The cult of yogins is central to the BraY . This scripture expounds a pantheon a of mantra-deities headed by the bhairava known as Kap lsa (Lord of the Skulla bearers) and the supreme Goddess, his consort, known primarily as Canda K p lin . . a a
On the BraY s model of the Saiva canon, see chapter 5, sections 4 and 5; see also chapter 4, section a 4 (especially tables 4.24). On the subdivision of the bhairavatantras into p. has, see Sanderson, Saivism t and the Tantric Traditions, 66871; and in the present thesis, chapter 3, sections 2 (in the discussion of the Svacchandatantra) and 3. 22 Note, e.g., BraY xi.43ab, referring to the smarana-mantra (on which see chapter 5, section 3): a . n nena rahit [h] devi sidhyante mantradevat h, without this [mantra], the mantra-deities do not bestow a a . a. siddhi. Cf. Ksemar jas comments ad Svacchandatantra 1.76cd77ab. The latter provides the mantra a . of Kap le a[bhairava], ending with the statement, kap leah prakrtitah in 77b ([this] is proclaimed to a s a s . . be Kap le a); Ksemar ja remarks, ayam kapale a [prakrtitah] uktah mantr nam mantradevataik tmatv t a s a a. . a a s . . . . (this is said to be Kap le a, because of the fact that mantras and the mantra-deities are identical). a s 23 Although not attested in the BraY , yogs also occurs in the literature in the same sense, e.g. a Tantr loka 1.322a. a

9 (Grim Bearer of the Skull), Aghor (Not Terrible), and Bhairav. Several sets of goddesses complete the primary deity mandala: the Four Goddesses (dev) .. or Guhyak s; the Four Attendants (kinkar), also called the Consorts (dut); the Six a Yogins; and the Eight Mothers (m tr), in descending order of status. Their names a. and mantric forms are delineated in chapter 5, section 3 (table 5.1). While possessing distinct identities and degrees of cultic importance, these goddesses belong to a common typology, discussed belowthat of the yogin or yogevar. The latter s terms are multivalent in the BraY , referring, according to context, to female initia ates, i.e. human yogins;24 to specic sets of deitiesthe Six Yogins of the primary mandala, most frequently;25 and to the nameless hordes of ying goddesses said .. to manifest before the practitioner and grant boons, in the culmination of the most arduous rites.26 And although in the BraY the terms dev (goddess), devat (dea a ity), and sakti (power) are preferred, related Saiva sources frequently use yogin and yogevar as generic terms for the myriad goddesses who pervade the cosmos as s expressions of Sivas power (akti).27 In addition, Yoge occurs as an epithet of the s s
A striking case is that of BraY i.12cd13, for which see the critical edition and translation in part a ii; this appears to refer to female practitioners as yogins. The passage is echoed in the opening verses of BraY xiv, quoted in chapter 2 (n. 200). a 25 See, for instance, BraY ii.1114ab (edited and translated in part ii). a 26 References to the yogins granting their daran to the successful practitioner are profuse; cf., e.g., s BraY xvii (f. 92r, line 1): a vidy cakram tu yo vetti yath vat s dhakottamah | a a a . . acirenaiva k lena yogeyo daranam vrajet | a s s .
cakram a a . ] em.; cakr n By 24

yoge yo ] corr.; yogesyo Bya s

The best of s dhakas who truly understands the Circle of Vidy safter a brief while the a a yoges become visible [to him]. s Here the optative singular vrajet apparently has the sense of the plural; cf. BraY iv.358c, devy[o] a pratyaksat m vrajet. . a. 27 Note, for instance, Tantrasadbh va 16.47cd48: a a tattvarup s tu yoginyo j tavy s ca var nane 47 a a a sivecch nuvidh yinyo manoveg mah bal h | a a a a a. . vicaranti samast s ca brahmavisnvindrabhumisu 48 a .. The Yogins should be known in the form of the reality levels (tattva), O fair woman. Carrying out the volition of Siva, as swift as thought and mighty, they all traverse the worlds of Brahm , Visnu, and Indra. a .. Text as quoted by Ksemar ja commenting on Netratantra 19.71 (here numbered as per Mark Dya . czkowskis collation of the mss). In contrast, the BraY refers to goddesses of the tattvas by the term a dev, in particular; in BraY xxxi, note for instance 93cd: tattve tattve sthit devyo adhik rapad nug h, The a a a a a. goddesses are present in each of the tattvas, adhering to their [respective] positions of authority. Cf.

10 supreme Goddess, especially in her role as eighth of the Mother goddesses (m tr).28 a. Yogin hence designates a spectrum of female sacred gures, many of which in different contexts bear distinct names. In the BraY , yogin most frequently refers a to a subset of a broader class of goddesses, while tantric Buddhist traditions often give preference to the term dakin as the generic name for such deities.29 It does . appear, however, that yogin had the widest currency as the class name for the goddesses in question, and the term is adopted accordingly in the present study. This is an amorphous class of deities, all the more so owing to the use of the term yogin in designating tantric adeptsfemale yogisor even the Great Goddess (mah dev) herself. It is nonetheless possible to identify characteristic attributes that a are widely, but not uniformly, shared by this class of deities. The possibility of a polythetic approach to classifying religious phenomena received serious treatment rst, I believe, in an essay of Jonathan Z. Smith. In the polythetic mode, membership in a class is determined by possession of signicant shared properties, no single one of which is necessarily held by all members of the classin rejection of the essentialist idea of perfect, unique, single differentia.30 Such a polythetic approach has been applied in relation to the nebulous category of Tantra, most notably by Douglas Brooks, who advances a polythetic denition Hindu Tantricism.31 While one might disagree with some of the properties identied and their relative priority,32 this approach appears productiveand, further a . 87cd: tattvarup [h] sthit devyo yogaivary hy anindit [h], the goddesses take the form of the tattvas, a s a a . possessing yogic mastery, and irreproachable. 28 In the BraY , Yoge describes the supreme Goddess as eighth of the Mothers in xlv.32b and a s lxvii.56a. Cf., e.g., yogs in M linvijayottaratantra 3.14d. a 29 See chapter 2, section 2 (n. 46); and chapter 3, section 4. 30 Smith, Fences and Neighbors: Some Contours of Early Judaism, chapter 2 in Imagining Religion: from Babylon to Jonestown, 45 (quote on p. 4). 31 a Brooks, The Secret of Three Cities: An Introduction to Hindu S kta Tantrism, 5372. Similar in principle are the lists of dening criteria proposed by Sanjukta Gupta, et al, Hindu Tantrism, 79; and, in reference to Tantric Buddhism, Stephen Hodge, introduction to The Mah -Vairocana-Abhisambodhi Tantra: With a Buddhaguhyas Commentary, 45. 32 It must be objected that the ten dening criteria Brooks adduces for Hindu Tantrism privia lege goddess-oriented or S kta traditions, particularly in their late medieval varietiesone of which is the subject of his monograph, Secret of Three Cities. Note in particular that the third criterion which begins, Tantrics are at once theists and philosophical nondualists (p. 58)excludes from the category Tantra the Saivasiddh nta and Vaisnava P car tra, which typically possess dualist thea a a ..

11 more, offers a useful approach to the category yogin. Below, I attempt a polythetic denition of the yogin, identifying what appear to be key shared properties char acterizing this class of sacred gure. Undoubtedly, there is much scope for further renement. This is not, moreover, the rst such attempt: David White has offered an eight-part descriptive denition of the yogin, which I both draw upon and depart from.33 Multiplicity. Characteristic of yogins is their occurrence in groups. From perhaps the tenth century, they became closely associated with congurations of sixtyfouran association that remains strong today. The multiplicity of the yogins is monumentalized in the stone temples dedicated to them constructed from
ologies. Part of Brooks seventh criterion seems in fact to have little application outside of late medieval, Sm rta tantric traditions. Asserting that tantrism does not differ signicantly from Pur nic a a. Hinduism in the ways it conceives the world and God, Brooks makes the equally surprising claim that a S kta- and Saiva-oriented Tantrics assume the pantheon of classical Hindu deities; he also utilizes Advaitaved ntic terminology for explicating tantric theology. Ibid., 6768. None of this appears apa plicable to early medieval Tantric Saivism of any variety. Criterion nine associates Hindu Tantra with ritual use of conventionally prohibited substances and antinomian practices. Based upon such a criteria, Brooks comes to the problematic conclusion that S kta forms of Tantrism are deemed to be Hindu Tantrism par excellence. Ibid., 72; see also 230 (n. 51). Furthermore, missing from Brooks list is at least one important criterion: the ontological identity of mantras and deities, which is surely a dening characteristic of the Saiva Way of Mantras (mantram rga). a 33 The denition White offers is as follows: The Yogins whose cults were central to Kaula practice had the following features: (1) they were a group of powerful, sometimes martial, female divinities with whom human female witches were identied in ritual practice; (2) their power was intimately connected to the ow of blood, both their own sexual and menstrual emissions, and the blood of their animal (and human?) victims; (3) they were essential to Tantric initiation in which they initiated male practitioners through uid transactions via their mouths; (4) they were possessed of the power of ight; (5) they took the form of humans, animals, or birds, and often inhabited trees; (6) they were often arrayed in circles; (7) their temples were generally located in isolated areas, on hilltops or prominences and were usually round and often hypaethral; and (8) they were never portrayed as practicing yoga for the simple reason that yoga as we know it had not yet been invented. Kiss of the Yogin, 27. In point eight, White presumably has in mind hathayogic bodily disciplines, with . which yogins were little connected. Many of the themes White identies are indeed signicant to conceptions of yogins, while points (2) and (3) seem problematic. There is undoubtedly a sanguinary dimension to the cult of yogins, and to the powers attributed to their most dangerous varieties. But as for yogins sexual and menstrual emissions, the supposed signicance of these remains mysterious to me; while the uids of female practitioners are certainly signicant in some rituals, the women concerned are seldom referred to as yogins, while references in primary sources to the uids of yogins, as goddesses, appear rare and ambiguous. Similarly, I am presently unaware of a Saiva tradition in which yoginspresumably White has in mind female ritual consortswere essential to Tantric ini tiation, and certainly not one in which they transacted with initiands in the manner suggested. On these points, I expect to write in greater detail in the future, examining the evidence on which White bases such claims.

12 the tenth century, enshrining congurations of forty-two and perhaps eightyone, but especially sixty-four goddesses.34 In Vidy ptha and tantric Buddhist a . sources, however, smaller pantheons are typical. Yogins are in fact character ized more by their multiplicity than their individual identities, for there exists remarkable uidity in the composition of yogin sets. No particular name or set of nameswith the exception of the Seven or Eight Mother goddesses, as discussed belowbecomes closely associated with these goddesses, and they are in fact frequently spoken of as an amorphous band or horde (vrnda, gana) . . that pervades the cosmos in innumerable forms and varieties. Even sources that place particular importance upon a group of sixty-four might mention numerous other yogins.

Figure 1.3: Siva enshrined within a circle of sixty-four yogins. Yogin temple, R npur-Jharial, Orissa. a AIIS Photo Archive.

On the temples of yogins, see Vidya Dehejia, Yogin Cult and Temples. A Tantric Tradition. In the present thesis, see chapter 2, section 4; concerning textual references to sixty-four yogins (primarily in Kaula sources), see there, as well as chapter 3, section 3. Ronald Davidson points out the existence of contemporary temples of sixty-four yogins in Banaras and Ujjain. Indian Esoteric Buddhism: A Social History of the Tantric Movement, 18183.


13 Manifestation in/as mortal women. Yogins blur the boundaries between goddesses and women, for through ritual perfection, a female tantric adept might become a yogin. Men, however, may only seek to join the yogins and partake of their powers. Taxonomies reect this phenomenon by positing yogins as a scale of beings, extending from powerful cult goddesses to the mortal yogins who em ulate and even embody the deities. Thus according to the Siddhayogevarmata, s a Vidy ptha scripture, yogins are fundamentally of two types: kulaj or born a . a in clans, called also m nusya, human; and devat h, deities/goddesses.35 a . a. More complex yogin taxonomies add further layers of ambiguity.36 Female di vinization hence lies at the heart of the image of the yogin, and comprises one of the most historically signicant facets of their cult. Organization into clans. Yogins, as both deities and female adepts, belong to clans (kula, gotra) which shape their natures and identities. Taxonomies of yogins exhibit considerable variety; however, in Saiva sources, their organization into clans of the Brahmanical goddesses called the Seven or Eight Mothers (m tr) a. appears fundamental.37 According to this schema, yogins partake in the na tures and appearances of the Mother goddesses, of whom they are considered partial incarnations or manifestations (amsa, lit. portion). Practitioners too . establish kinship with the goddesses, becoming thus their amsas, for initiation .

Siddhayogevarmata 22.5: s

dvividh yoginh prokt h kulaj devat s tath | a . a. a a a m nusy h kulaj h prokt s tesam srnu kulodgatim | a . a. a. a . . . . dvijaksatriyavitsudrakulotpann s tu n yik h a a a. . . Yogins are taught to be divided into two groups: those born in a lineage[,] and deities. Those born in a lineage are taught to be human; listen to their family line. [These] Heroines are born in Br hmana, Ksatriya, Vai ya, or Sudra families. a s . . Translation by Judit Trzsk, editor of the Siddhayogevarmata: The Doctrine of Magic Female Spirits: s A Critical Edition of Selected Chapters of the Siddhayogevarmata(tantra) with Annotated Translation s and Analysis, 171 (Sanskrit text on p. 50). The passage following upon this one is lacunose. This passage has a parallel in Tantrasadbh va 16, beginning with verse 129. a 36 Illustrative of this is a taxonomy of yogins elaborated in Kaulaj nanirnaya 8; relevant passages a . are quoted and discussed in chapter 3, section 3. 37 Buddhist taxonomies of yogins, as might be expected, seldom associate the deities with the Brah manical Mothers. The Laghucakraamvaratantra, for instance, advances several classication schemas s . based upon clans headed by identiably Buddhist deities, in chapters 1619 and 23.

14 effects entry into the clans of the deities.38 Sets of yogins, such as those en shrined in temples and listed in the pur nas, frequently include the Mothers a. among them,39 while the term m tr (Mother goddess) is occasionally applied a. to yogins.40 Classication of the clans of yogins forms an important theme in the literature of the Yogin cult, an example of which is BraY lxxiii, edited a and translated in part ii of the present thesis. The BraY , moreover, places the a Mothers within a broader hierarchy of goddess clans, extending from deities of the most pernicious sortsuch as the dangerous damarto the texts highest . mandala deities.41 .. Theriomorphism. A hallmark of yogins is polymorphism, with theriomorphic forms being especially common. From horses and lions to birds and snakes, sculptural and textual representations of yogins attest a wide variety of animal elements. As a deity typology, a close parallel lies in Sivas ganas (troops, horde), an . amorphous and diverse class of male deity, often theriomorphic, whose imagery ranges from the horric, grotesque, and martial, to comic, exuberant, and musical. Another parallel lies in the multitudinous Mother goddesses de scribed in the Mah bh rata, deities with whom the genealogy yogins is closely a a linked.42 In contrast to the ganas and Mothers, actual shapeshifting is closely . associated with yogins, who are thought to take on the forms of female animals in particular. Tales of yogins also associate them with the power to transform others.43 K p lika cult and iconography. The yogins connection with the cult of Bhairava, the a a
See chapter 2, section 1. Dehejia, Yogin Cult and Temples, appendix I (pp. 187200). 40 Use of the term m tr in the sense of yogin is attested in the Siy n inscription of Nayap la, in a. a a Bengal; see chapter 2, section 4 (n. 299). Dehejia points out several other examples; Yogin Cult and Temples, 31. 41 BraY lv.354. a 42 See chapter 2, section 2 (on the Mah bh rata). a a 43 On the shapeshifting of the yogin, note Kaulaj nanirnaya 23, quoted and discussed in chapter a . 2, section 4 (the text of 23.112ab is given in n. 309). Stories of yogins from the Kath sarits gara are a a discussed in chapter 2, section 3 (subsection on the Brhatkath retellings). a .
39 38

15 archetypal mortuary ascetic (kap lin), nds representation in their k p lika or a a a mortuary iconography, for they frequently bear skulls, bone ornaments, and skull-staves (khatv nga), as well as incorporate other elements of radical tantric . a iconography. Furthermore, yogins have a strong association with cremation grounds: while a variety of liminal places are spoken of as their haunts, their primary locus is the charnal ground (maana), the preferred site for the radical s s practices advanced in Vidy ptha tantras. a . Danger, impurity, and power. Fundamental to conceptions of yogins is their potency as sources of both danger and immense power. In this respect they inherit the ambivalence of early Indian Mother goddesses. While dangerous to noninitiates and fatal to apostates, the wild horde of yogins becomes all-benecient to the greatest of tantric heroes (vra) who succeed in their arduous rituals of propitiation. It is to such rituals that much of the BraY and similar sources a is devoted, and the Yogin cult is distinguished by the aim of achieving bless ings of these beings in direct, transactional encounterscalled most frequently mel pa or melaka (meeting, encounter, union). Effecting and navigating ena counters with yogins thus become subjects to which the literature of the cult devotes much attention.44

A passage from BraY xiv provides a vivid account of the danger posed by yogins in ritual: a japet mantro mah sattvo digv so daksinamukhah | a a . . . saptar trena yoginyo agacchanti mah bhay h 214 a . a a. a raudrarup s tath suddh h sakrodh m ranatmik h | a a. a a . a. tad drstv tu na bhetavyam vrasattvena mantrina 215 .. a . . argham t s m prad tavyam pranip te krte sati | a . a a. . . a . tusyante n tra sandehah s dhake sattvasamyukte 216 a . . a . kathayanti ca tam sprs. v yath rtha ca subh subham | a a . ..t a pram d d yadi ksubhyeta sattvahnas tu s dhakah 217 a a a . . tatksanad devi kh danti yoginyo yogadarpit h | a a. . . na tam raksayitum sakto rudro pi svayam agatah 218 . . . .
suddh h ] corr.; suddh a a 214b mukhah ] em.; mukham Bya 215a rup s ] em.; rup Bya a. a . . Bya 215b tmik h ] em.; tmik Bya 216a prad tavyam ] em.; prad tavyah Bya 217a tam ] conj.; t m a. a a a a. . . . Bya 217c pram d d ] em.; pram d Bya 218a ksanad ] em.; ksana Bya 218c tam ] em.; ta Bya a a a a . . . . .

The [s dhaka] of great spirit should recite the mantra, naked, facing south. After seven a nights, the yogins comehighly dangerous, with terrifying forms, impure, angry, and lethal. But seeing this, the mantrin of heroic spirit should not fear; after prostrating, he should give them the guest-offering. [They become] pleased towards the s dhaka endowed a with [heroic] spirit, without a doubt. And touching him, they tell truly the [prognostication

16 The danger and power of the yogin appear closely linked to engagement with impurity, an important dimension of which is her association with the cremation ground. They epitomize a culture of ritual nondualism (advaita), in which the purity conventions of Brahmanical orthopraxis give way to a visionary mysticism of fearless omnipotence, of unfettered super-agency, in which the s dhaka seeks to assimilate the powers of the hordes of yogins, primarily, in a occult manipulations of impurity.45 Transactional encounters with yogins of ten revolve around conventionally impure substances: practitioners offer wine or their own blood in lieu of the guest-water offering (argha), burn incense of neem oil and garlic, make offerings of esh in re sacrice, or even offer mixed male-female sexual uids. Conversely, a yogin might offer impure rit ual gruel (caru) to the disciple, the unhesitant acceptance and consumption of which becomes a medium for her bestowal of power.46 Protection and transmission of esoteric teachings. Yogins are ascribed the dual roles of protecting and in some cases disseminating esoteric tantric teachings. Often, their bestowal of power manifests in the transmission of secret lineage teachings (samprad ya),47 rather than direct transference of power (siddhi). In some a cases, works of tantric literature link their pedigree to transmission by yogins.48
of] good and bad. If by mistake a s dhaka of weak spirit should tremble, the yogins, a arrogant with their yoga, devour him that very moment. If he came, not even Rudra himself would be able to save him. Sanderson, Purity and Power among the Brahmans of Kashmir, 201. On the offering of caru, note BraY lxxxv.13cd14ab, quoted in chapter 3, section 5; and Kaulaa j nanirnaya 11.7cd10, in the same section (n. 165). a . 47 On the notion of samprad ya as esoteric knowledge transmitted by yogins, see the annotation to a BraY lxxiii.74. In some cases the teachings imparted by yogins are referred to as j nawisdom, a a which, though vague, can have textual connotations. Note BraY xcvii.2526ab: a
46 45

m tryoginik y ni sakinn m kul ni tu | a. a a a . a sidhyanti s dhakendrasya yogen nena suvrate 25 a a kathayanti ca sadbh vam kulajam j nam uttamam | a . . a Through this yoga, O pious woman, the clans of the groups of Mothers and yogins, [and] of sakins, bestow siddhi on the s dhaka; and they speak the true essencethe highest a scriptural wisdom (j na) arising from the Clans. a
48 A noteworthy case is that of the Mah rthamajar of Mahe var nanda, a South Indian author writa s a ing around the beginning of the thirteenth century who attributes his composition to the inspiration

17 According to the BraY s revelation narrative, yogins are said to hide away the a scripture at the end of the Kaliyuga, reecting another aspect of their roles as guardians of the teachings. Furthermore, yogins wreak destruction upon violators of the tantras, including those who break the initiatory Pledges (samaya) hapless individuals who risk becoming yogin food.49 Flight. Yogins are consistently associated with the power of ight, foremost among the powers (siddhi) sought by their votaries. In this they inherit the mantle of the vidy dhara and vidy dhar, the semi-divine sorcerors of early Indic myth.50 a a Taxonomies of yogins suggest that aerial deities represent only one of their numerous varieties, alongside e.g. terrestrial (bhucar) goddesses. However, the archetypal yogin is the autonomous Sky-traveller (khecar), joining whose ranks represents the ultimate attainment for the siddhi-seeking practitionerthe s dhaka. a Discussion is warranted concerning the expression Yogin cult,51 for this ren ders no emic category used in the classication of the tantric traditions, and, furthermore, risks the exotic connotations of cult in its popular usageconnotations that deeper acquaintance with some aspects of the tradition might not immediately
of a visionary encounter with a yogin. See Whitney Cox, Making a Tantra in Medieval South India: the Mah rthamajar and the Textual Culture of Cola Cidambaram, 16. Cox draws attention to several a similar traditions of inspiration by a yogin; ibid., 23. 49 Cf. Vnasikha 329cd21ab: . svayamgrhtamantr s ca n stik vedanindak h 329 a a a a. . . . a. samayebhyah paribhras. as tath tantravidusak h | a . . t .. gurunam vihethanapar s tantras ravilopak h 320 a a a. . yoginbhih sad bhras. ah kathyante dharmalopak h | . a a. . t . Those who take up mantras on their own, atheists, critics of the vedas, breakers of the Pledges, desecrators of the tantras, those intent on harming the gurus, and those who violate the essence of the tantrasthose who violate Dharma are said ever to be ruined by the yogins. On the threat of being eaten by yogins, see above (n. 44). 50 See chapter 2, section 3 (subsection on the Brhatkath and its retellings). a . 51 The expression Yogin cult surfaces in the works of early twentieth-century authors, such as P. C. Bagchi. I have not yet identied nineteenth-century precedents, although these might exist in archeological reports on yogin temples. Contemporary scholars, such as Vidya Dehejia and Alexis Sanderson, have continued to use this terminology, although David White eschews the expression in Kiss of the Yogin: Tantric Sex in its South Asian Contexts.

18 discourage. Cult is nonetheless a productive, if not indispensable category for referring to tantric systems of worship. Though centered on specic deities and often possessing distinct authorizing scriptures, tantric worship systems or cults are not mutually exclusivecertainly not to the point of being distinct sects. Yogins, however, do not gure as cultic focii in the manner of most tantric di vinities: their cult is integrated within those of the high deity or deities who form the primary focus of a given ritual system. While the BraY expounds the cult of a Kap lsabhairava and Aghore or Canda K p lin, ritual practices connected with a s . . a a yogins register a constant presence. This is true of the other major extant Vidy ptha a . tantras as well.52 It would nonetheless be problematic to identify the cult of yogins with the Vidy ptha, for the Yogin cult extends beyond its connes. Most of the sura . viving Saiva literature closely connected with yogins in fact belongs to the corpus of Kaula scripture, which appears to have roots in Vidy ptha traditions.53 Moreover, a . the cult of yogins is by no means restricted to Saivism, for it characterizes Indian Tantric Buddhism in its latter phases, which saw the production of a corpus of scriptural literature frequently referred to as the yogintantras, Tantras of the Yogins. There is even evidence suggesting Jaina engagement in the cult of yogins.54 The Yogin cult hence extends across both cultic and sectarian boundaries, identifying a complex of cognate tantric cults spanning several centuries which placed considerable emphasis upon a multiplicity of female divinities designated, most frequently, by the term yogin and its variants.
52 In the words of Alexis Sanderson, Accessible from the main cults of the Vidy ptha, and undera . . lying them in a more or less constant form, is the more ancient cult of Rudra/Bhairava in association with female spirits (Yogins). Saivism and the Tantric Traditions, 671. 53 On the Kaula and Vidy ptha distinction, see chapter 3, section 3. a . 54 Signicantly, a class of Jaina goddesses appears to be modelled upon yogins: that of the vidy devs a or Wisdom goddesses. Occuring in groups usually sixteen in number, these goddesses nd depiction in numerous temples, perhaps most notably in circular ceiling panels in the temples of Mt. Abu. On the Jaina vidy devs, see U. P. Shah, Iconography of the Sixteen Jaina Mahavidyas; Maruti Nandan Prasad a Tiwari, A Note on the Figures of Sixteen Jaina Goddesses on the Adinatha Temple at Khajuraho; and John Cort, Medieval Jaina Goddess Traditions. A Jaina commentator upon the Yaastilaka (on which s see chapter 2, section 3) in fact explicitly identies yogins as vidy devs; elsewhere he describes them, a drawing on a Jaina taxonomy of divinities, as mah vyantardevyah: great goddesses of the intermediate a . class.

19 Several scholars have, to various degrees, connected the worship of yogins with a sect supposedly called the Yogin Kaula. This designation is dubious, however, being in fact based on the erroneous interpretation of the expression yoginkaula as it occurs in the Kaulaj nanirnaya. A Kaula scripture, this tantra has been cited compara . atively widely by virtue of having been published seventy-ve years ago. The phrase yoginkaula occurs as a titular epithet in almost all colophons of the Kaulaj nanirnaya, a . and twice elsewhere in the text;55 this appears to mean Kaula teaching of [=transmitted by?] the yogins. That yoginkaula refers to esoteric knowledge associated with or possessed by yogins is suggested by the Mrgendr gama, where this term de a . scribes one of eight sub-streams (anusrotas) of scriptural revelationa tradition of secret wisdom maintained by the yogins.56 However, Prabodh Candra Bagchi, the Kaulaj nanirnayas editor, interpreted the term as a sectarian designation; he cona . cluded that Matsyendra was the founder of a new sect of the Kaula school, called the Yogin-kaula.57 In arriving at this he apparently misinterpreted a list of texts as a list of sects: the passage in question seems to list four texts, one of whichthe Kaulaj nanirnayais described as yoginkaula. This appears entirely consistent with the a . colophons.58 The one other non-colophonal occurrence of the term yoginkaula is
Note for instance the colophon of chapter seven, which is typical: iti j nanirnnitiyoginkaulam a .. mahacchrmacchaghnap d vat rite candradvpavinirgate saptamah patalah, as reads Kjncod . The syntax is a a a . . . puzzling: presumably read yoginkaule. 56 Mrgendr gama, Cary p da 40cd41ab: a a a . yoginyo lebhire j nam sadyoyog vabh sakam 40 a . a a yena tad yoginkaulam nottrnam t bhya eva tat | . . a . The yogins obtained scriptural wisdom that immediately makes [the power of] yoga manifest. For this reason, it is [called] yoginkaula (Kaula wisdom of the yogins). It has not [subsequently] emerged forth from them. Bhatta N r yanakantha remarks on this verse that the yogins obtained scriptural wisdom (j na) from aa . a .. .. Siva, which remains among them alone as a secret tradition (sadyah tatksanam eva yogam avabh sayati a . . . yat tath vidham j nam sivabhat. arak d yoginyah pr puh | tac ca tabhya eva sak sad nottrnam n nyatra a a . . a . a . . t a . a . prasrtam asv eva samprad yatay sthitam ity arthah ). a a . . 57 Bagchi, introduction to Kaulaj nanirnaya and Some Minor Texts of the School of Matsyendran tha, 35. a a . 58 The passage in question is Kaulaj nanirnaya 16.4749: a . mah kaul t siddhakaulam siddhakaul t matsodaram | a a a . caturyugavibh gena avat ram coditam may 47 a a . a . j n dau nirnitih kaulam dvitye mahat samjitam | a a . . . . trtye siddh mrtam n ma kalau matsodaram priye 48 a . . a . . ye c sm n nirgat devi varnayisy mi te khilam | a a a . . a etasm d yoginkaul n n mn j nasya nirnitau 49 a a a a a .

20 more problematic, perhaps referring to a specic technique.59 Abhinavagupta, incidentally, cites a text by the name Yoginkaula; yet this does not appear to be the Kaula j nanirnaya.60 Bagchis problematic postulation of a sect called the Yogin Kaula a . has been repeated by V. W. Karambelkar, in an article entitled Matsyendran tha and a his Yogin Cult;61 Devangana Desai, in a discussion of the Yogin temple of Khaju
47b matsodaram ] Kjncod ; mas daram Kjned a 48a nirnitih ] Kjncod ; nirntih Kjned . . . . cod ; matsyodaram Kjned matsodaram ] Kjn 49d nirnitau ] Kjncod nirnntau Kjned (unmetrical) . . . . . 48d

From the Mah kaula comes the Siddhakaula; from the Siddhakaula the Matsyodara; I aca complish the descent [of scriptural revelation] in accordance to the division of the four yugas. [47] In the beginning ( dau) there is the scripture (j na) [called] Nirnitikaula a a . [i.e. J nanirnitikaula = Kaulaj nanirnaya]. In the second [yuga], the one called Mahat a a . . [=Mah kaula]. In the third, the one named Siddh mrta [=Siddhakaula]. In the Kaliyuga, the a a . Matsodara, my dear. [48] And I shall describe entirely those [scriptures?] which emerged from this, this yoginkaulathe J nanirniti [i.e. Kaulaj nanirnaya] by name. a a . . The syntax and interpretation of 49 are especially problematic. It appears to me that the locative nirnitau must agree with the ablative etasm d (49c)a grammatical barbarism not beyond the language of a . this text, in which there is often little distinction between the oblique cases. Bagchi arrived at a rather different interpretation of this passage, identifying matsodara (i.e. matsyodara) as a reference to the gure Matsyendran tha, and Yoginkaula as a sect. He remarks, it appears a from these slokas that Matsyodara belonged to the Siddha or Siddh mrta sect and was particularly con a . nected with the Yogin-kaula, the doctrines of which are explained in the J nanirnti. Introduction to a . Kaulajnana-nirnaya, 35. White too refers to a group called the Yogin Kaula, presumably having the same passage in mind. Kiss of the Yogin, 22. He interprets the passage above as listing the sectarian groups through which the Kaula gnosis was transmitted, down to the Fish-Belly in the present age (Kiss of the Yogin, 103). He offers the following translation of verses 4748: From the Mah kaula [arose] the Siddha Kaula; from the Siddha Kaula the Fish-Belly. It a was uttered by me upon each of the divisions of the four ages (yugas). In the rst [age] the bringing forth (nirniti) [was made] to the Kaula; in the second to the [Kaula] known . as Mahat; in the third, to the [Kaula] named Siddh mrta [and] in the Kali [age] to the a . Fish-Belly. (Kaulaj nanirnaya 16.4748) a . Kiss of the Yogin, 25. His translation of verse 49 is given elsewhere (p. 103): I will now discuss to thee, in their entirety, those [teachings] that were lost [in transmission], O Goddess! [The teaching known] by the name of [the Bringing Forth of the Kaula] Gnosis came through this Clan of the Yoginsi.e. the Yoginkaula. However, this and the others are most certainly texts, not sects. Note for instance that Kaulaj nanirnaya 16.54a refers to what was spoken in the Siddh mrta (siddh mrte tu yat proktam); a a . a . . . the Siddh mrta is moreover mentioned in a list of scriptures in Kaulaj nanirnaya 21. According to a . a . Sanderson, the Siddh mrta is quoted by early Kashmiri exegetes, as was a text by the name Matsyodar a . (cf. Matsyodara in Kaulaj nanirnaya 16.48). History through Textual Criticism, 4. In the same passage a . in Kaulaj nanirnaya 21 are also listed the Mah kaula (21.5a) and the Siddhevara[kaula] (21.7a), the latter a a s . probably identical to the Siddhakaula mentioned in 16.47ab. White in fact appears to contradict himself concerning the interpretation of Kaulaj nanirnaya 21: he once refers to this as containing a list of no a . less than nine clan scriptures (p. 105), but elsewhere an expanded list of the various subclans of the Kaula (p. 25). 59 Kaulaj nanirnaya 14.59b. a . 60 Tantr loka 7.4041. Jayaratha too quotes from the Yoginkaula, commenting after Tantr loka 7.19ab. a a 61 Karambelkar, Matsyendran tha and his Yogin Cult, 365. This article has in turn been drawn a upon, for instance, by R. K. Sharma, who on its authority associates the cult of yogins and yogin temple tradition with the Yogin Kaula founded by Macchendran tha, adding that the principal tenets of a the Yogin Kaula m rga is [sic] revealed in the Kaulaj nanirnaya . . . . The Temple of Chaunsatha-yogin at a a . . Bheraghat. Cf. H. C. Das, T ntricism: A Study of the Yogin Cult, vii, 23. a

21 raho;62 Nilima Chitgopekar, in an article examining yogins from the perspective of gender;63 and David White, in his recent monograph on yogins and Kaula sexual ritual.64 It would appear that a Saiva cult of yogins ourished to the greatest extent in the period circa 7001200 c.e., although its presence extends both before and beyond this period. From the tenth to perhaps thirteenth century, monumental stone temples enshrining yogins were constructed spanning from one end of the subcontinent to another. Yet beyond this period, it becomes increasingly problematic to speak of a Saiva Yogin cult, although tantric practices connected to yogins, and certainly belief in them, nd continued attestation. Latter medieval Kaula sources, such as the Kul rnavatantra, continue to attach signicance to yogins; yet their roles pale in a . comparison to those in earlier Kaula sources. The decline of the Saiva cult of yogins appears to have been gradual, and some indication of this trajectory may be gathered from its eclipse within the N th cult of the second millennium. While well-known a medieval Sanskrit works of N th yoga place relatively little importance upon yogins, a there is reason to believe that the case was different in the thirteenth century; this is evident from the prominence of yogins, alongside e.g. hathayoga, in the lost Amrta . . kunda or K mrubj ksa, a text translated into Persian (and subsequently Arabic) most a a . .. probably in the thirteenth century.65
Desai, The Religious Imagery of Khajuraho, 92. Citing the Kaulaj nanirnaya, Chitgopekar claims that the name Yogin-Kaula refers to a religious a . system which is orally transmitted by a line of female ascetics, the Yogins. The Unfettered Yogins, 93. 64 See above (n. 58). 65 Referring to the Persian redaction, which he believes to represent the earliest Islamic version, Carl Ernst remarks, This eclectic Persian text contained breath control practices relating to magic and divination, rites of the yogini temple cult associated with Kaula tantrism, and the teaching of hatha yoga according to the tradition of the Nath yogis. . . All of this was placed in a context of the supremacy of the goddess Kamakhya . . . . The Islamization of Yoga in the Amrtakunda Translations, 204. Incidentally, Ernst sees the association this text makes between a series of yogins and the planets as a deliberate attempt by the translator to familiarise the subject, in this case by likening the summoning of Indian goddesses to well-known Middle Eastern occult practices involving planetary spirits. Ibid., 219. However, this interpretation probably overlooks medieval Indian astrological conceptions of yogins; the predictive technique known as yogindaa is based upon conceiving of eight planetary bodies as yogins. s Manuscripts on the subject of yogindaa are listed in the catalogs of several collections; I have for in s stance examined one by this title, attributed to the Rudray mala, in the Van Pelt Library, Philadelphia; a Collection of Indic Manuscripts no. 390, item 714. That the system of yogindaa remains in practice s
63 62

22 The present study is concerned with yogin traditions of the rst millennium. Its principal objective is to advance understanding of the content and contexts of an unpublished and little-studied tantric Saiva scripturethe BraY . The thesis has two a parts: the rst consists of studiestwo chapters of which concern the cult of yogins broadly, and two of which focus upon the BraY while part ii presents a critical a edition and annotated translation of selected chapters of the BraY . Following a a review of scholarship on the subject of the BraY , below, chapter two embarks upon a an examination of the early literary, epigraphic, and sculptural evidence for the cult of yogins. This is intended to be comprehensive with respect to pre tenth-century material, while later sources are discussed selectively. Excluded from consideration is tantric literature proper; this instead forms the subject of chapter three, which traces the background and formation of the yogin cult in early Saiva and Buddhist textual sources. Although the scope of chapters two and three is broad, the BraY remains a a constant point of reference. In chapter four, focus shifts to examination of the form, content, and structure of the BraY ; this chapter also addresses the question of the a texts dating and provenance. The fth chapter focuses on interpreting the identity the BraY articulates for itself within its model of scriptural revelation and the Saiva a canon, based upon investigation of its various titles and epithets. The source materials for this thesis are to a large degree unedited and unpublished texts, and the difculties inherent in working with these have dictated an approach that is text-critical in emphasis. Chapter two involves examination of epigraphic and material evidence as well, particularly religious images. As will become apparent from the critical edition of part ii, considerable philological scrutiny is required to yield sense from the BraY in a great number of cases, and even then, the a interpretation often remains provisional. This situation is not exceptional in the cor pus of tantric Saiva texts, the study of which remains at an early stage. In the absence of unambiguous data concerning the BraY s provenance, dating, and authorship, the a
is suggested by modern astrological manuals such as Rajeev Jhanji and N. K. Sharma, Applications of Yogini Dasha for Brilliant Predictions.

23 approach has been to situate the BraY in relation to the available materialstexts, a inscriptions, and sculptureand to probe the texts self-presentation for clues concerning the agents and circumstances involved in its production. Chapters of the BraY included in the critical edition and translationpart ii a were selected on the basis of their relevance to the studies in part i. BraY i and ii are a important sources for chapter ve, along with BraY xxxviii, the inclusion of which a was prevented by time constraints. The other chapters edited concern subject matters central to the cult of yogins: yoginlaksana or the characteristics of yogins; chomm , . . a the secret signs used for communication with the deities and other initiates; and yoginmel pa, encounters with yogins. Further relevant material from the BraY on a a the subject of the clans of yogins could not be included, unfortunately.


the brahmayamala in scholarship

Although acknowledged as important, the BraY has received relatively little schola arly attention. Its oldest manuscript was described more than a century ago by Harapras d S str, in his partial catalog of the collection of the former Durbar Lia a a brary, Nepal.66 S str says little concerning the BraY s content, but provides an a incomplete list of chapter colophons. Decades later, Prabodh Candra Bagchi penned several pages concerning the BraY in an appendix to his 1939 book, Studies in the a Tantras. In this, he summarizes chapters one and thirty-eight,67 providing also the text of several passages.68 I am not aware of further scholarship substantively addressing the BraY in the decades which followed. After a gap of half a century, Teun a Goudriaan wrote on the BraY in his history of the literature of Hindu Tantraan a
A Catalogue of Palm-leaf and Selected Paper Manuscripts Belonging to the Durbar Library Nepal, vol. ii, a 6062. Reprinted in Reinhold Grnendahl, A Concordance of H. P. S stris Catalogue of the Durbar Library and the Microlms of the Nepal-German Manuscript Preservation Project. 67 Following the colophonal numbering in the ms, Bagchi labels the latter chapter thirty-nine. 68 Studies in the Tantras, Part i, 1025 (appendix: Detailed Notices on Manuscripts). This book is a compilation of articles Bagchi published in the early 1930s, and I do not believe a second part ever appeared. His accounts of the BraY are frequently inaccurate. Note, for example, that Bagchi refers a to BraY i, the sambandhapatala, as having the title Aksaravidh na; this he presumably culls from the a a . . s texts epithet Nav ksaravidh na. He would also have I vara rather than Bhairava teaching the tantra to a . a the Goddess.

24 ambitious undertaking considering how little of the early material had been studied carefully. After making some preliminary observations on the BraY and the y malaa a tantras as a genre, Goudriaan provides a precis of the texts subject matters, as well as an excerpt from chapter xxiv in translation.69 Although offering little insight into the historical position and signicance of the text,70 he noted that a closer study of the Picumata, although certainly not an easy task on account of its cryptic ways of expression, rambling style and bad grammar, is necessary for better insight into early Hindu Tantrism.71 The rst attempt at more detailed study of the BraY was made by the late S. a N. Ghoshal Sastri of Vi vabh rat University. Sastri drew heavily upon the BraY s a a in his ambitious multi-volume series, Elements of Indian Aesthetics. Unfortunately, he had at his disposal only a single incomplete ms of the work containing chapters four through seven, which I refer to as the Vi vabh rat ms, and a transcription s a of chapter xlii from an unknown ms. Based upon the limited evidence available to him, Sastri believed the Picumataan epithet of the BraY to be a section of the a a BraY , much as he considered the Pingal mata to be a section of the Jayadrathay mala; a a incomplete mss of both the latter texts were also available to him in the Vi vabh rat s a collection.72 Sastris primary interest in the BraY was its material culture and arts, for a he saw in its mandalas, iconometry, iconography, ritual, and ritual paraphernalia .. evidence for primitive Indian arts and crafts. He considered the Tantric tradition one of the principal streams of ancient Indian aesthetics, alongside the Vedic, and saw the BraY , perhaps correctly, as a uniquely important source for study of early a Tantra.73 Based upon connections of a most tenuous nature, Sastri claimed that the
a Hindu Tantric and S kta Literature, 4044. Cf. for instance Goudriaans rather bland remark that the Picumata is a typical representative of the Bhairava current in Saiva Tantrism. Ibid., 43. 71 Ibid., 44. 72 See the discussion of mss in the introduction to the critical edition in part ii. These manuscripts were all apparently of Nepalese provenance, gifted by the monarch of Nepal to Rabindranath Tagore. 73 Sastri, Elements of Indian Aesthetics, vol. ii, part 1, chapter xi.
69 70

25 BraY represents the cultural legacy of Indo-Tibetan tribal peoples of the northeastern a regions of the subcontinentspecically, the G ro tribe of Megh laya, a region in a a a which a district headquarters bears the name Tur . This he connected with the word tura for skull, used often in the BraY , which appears to have been his only a evidence linking the BraY to the G ro tribe.74 In addition, he considered the BraY a a a improbably ancient.75 The distinction of rst publishing a complete chapter of the BraY is Sastris, the a only chapter to have appeared in print prior to the present dissertation.76 Sastri considered this forty-second chapter of the BraY ,77 the mudr patala or Chapter on a a . Mudr , to be one of the most ancient treatises of Indian Gesturology.78 He puba lished his edition on the basis of Harad s Mitras transcription of a manuscript bea lieved to have once been in the Vi vabh rat collection. I suspect that the manuscript s a in question transmitted BraY xlii independently, whether alone or in a composite a manuscript. It is possible but by no means certain that its readings reect a transmission distinct from that of the oldest Nepalese ms, nak 3-370.79 Sastris edition reproduces this transcription with several proposed emendations, providing also an English rendering which, at times implausible and at times incomprehensible, conElements of Indian Aesthetics, vol. ii, part 4, 41. In ibid., vol. ii, part 1 (p. 98), Sastri opines for a date of the third century or earlier on the basis of a a dubious relative chronology of the N . yaastra, K lid sas Meghaduta, the BraY and Pingal mata, and at s a a a Matsyapur na. Yet in vol. ii, part 4 (p. 3), he claims the BraY was composed between the third and fth a. a centuries, for which he refers the reader back to the above discussion in vol. ii, part 1! 76 Elements of Indian Aesthetics, vol. ii, part 3, 297305. Note that the editions title page misleadingly refers to this as the rst chapter in the BraY /Picumatas Caturtha-satka (4th Part of the 6th unit). The a . . BraY might possess two satkassee chapter 4, section 2but this chapter is the rst of neither. The a . . expression caturthasatka means in fact the fourth division of six-thousand verses. Appparently Sastri . . confuses the BraY with the Jayadrathay mala, the latter of which is divided into four satkas. a a . . 77 Following the colophon, Sastri refers to this as chapter forty-three; but it is forty-second in order of occurrence. The numbering in the ms goes awry from chapter xxix until lxxx. 78 Sastri remarks,
75 74

We nd no other earliest extant evidence of Aesthetic Gesturology than the N tya astra. a. s On the other hand, no earliest Tantra and Pur na than the Brahma-y mala-tantra is yet a. a known to us. On the above postulation, the Mudr dhikarana of the Picumata and the a . Pingal mata, may be considered the prime documentary source of Indian Gesturology a and to that end in the original texts of two Mudr dhikaranas are annexed to the Part iii a . of the present volume. [sic] Elements of Indian Aesthetics, vol. ii, part 3, xiv. 79 See the introduction to part ii.

26 tributes relatively little to the interpretation of the text.80 Mark Dyczkowski makes more meaningful use of the BraY in his 1988 monoa graph on the Saiva scriptural canon.81 Reading from its oldest Nepalese ms, he cites passages from the BraY as supporting evidence on a variety of subjects. In partica ular, Dyczkowski draws upon BraY xxxviiis account of Saiva revelationthe very a chapter which had interested Bagchi. Making several useful observations,82 he does not however attempt a detailed analysis of the BraY or advance a hypothesis on its a historical position.83 As with so much of tantric Saiva literature, signicant strides in the study of the BraY commenced with the works of Alexis Sanderson. In Saivism and the Tantric a Traditions, his monumental 1988 essay mapping the cults and canon of Tantric Saivism, Sanderson advanced a compelling hypothesis concerning the position of the BraY within Saiva traditions. He notes several signicant ways in which the text a appears archaic. Seeing within the Bhairava-stream of esoteric Saivism an historical

Note for example his text and translation of the opening verse, BraY xvii.1: a atahparam pravaksami karasamsk ra-odhanam | . . . . a s mudr nam laksanam caiva sarvak m rthas dhanam 1 a. . a a a . . . Let me now tell how we could sanctify and purify our hands. What is the denition of the mudr (Gesture) and which are their common characteristics. These mudr s may lead a a to the way of success all noble works and fulll all desires of a man.

Elements of Indian Aesthetics, vol. ii, part 3, 297. 81 a The Canon of the Saiv gama and the Kubjik Tantras of the Western Kaula Tradition, especially 3653. a 82 Note for example the following remark: The BY makes use of this [four-p. ha] system of [scriptural] classication, integrating it t somewhat awkwardly with a division of the scriptures into Left, Right and Middle currents. The BYs account of the p. has is sketchy and unsystematica sign that this system t of classication is still at an early stage of development. a Canon of the Saiv gama, 51. While the BraY s p. ha system does seem undeveloped, its juxtaposition a t with a system of streams is not however awkward: the four p. has are divisions of a single stream, t the daksinasrotas of bhairavatantras. I provide an edition of the relevant passages from BraY xxxviii in a . . chapter ve. 83 On one important historical matter Dyczkowskis remarks warrant reconsideration: he suggests that the BraY must be younger than the Nity sodaik rnava, on account of the BraY listing the Yoa a. . s a . a ginhrdaya in its description of the canon; Yoginhrdaya is a name of the Nity sodaik rnava. Canon of the . . a. . s a . a Saiv gama, 4748. While it is true that the BraY lists a text called Yoginhrdaya, it is improbable that it a . refers to the same Kaula scripture of the cult of Tripurasundar which survives by this namea text which, as Dyczkowski points out, mentions several other Kaula scriptures, including the rather late Kubjik mata. (On the date of the Kubjik mata, see Sanderson, Remarks on the Text of the Kubjik mata, a a a 13.) Abhinavagupta makes no reference to a scripture called Yoginhrdaya, and according to Sanderson, . the extant Yoginhrdaya displays substantial inuence from the nondualist Kashmiri exegetical tradition. . See The Visualization of the Deities of the Trika, 37.

27 trajectory towards increasing emphasis on goddesses, he nds in the BraY an early a window into this process. Ostensibly a text teaching the cult of a bipolar Godhead, a y mala or god-goddess pair, in the BraY , the goddess Aghore var in fact transcends a a s Bhairava, for her vidy -mantra contains within its nine syllables the entire mandala of a .. mantra-deities. And in the ritual practices of the BraY , Sanderson identied the rada ical mortuary (k p lika) and exorcistic rites forming the earliest stratum of the Yogin a a cult, which would undergo transformation as the Kaula movement came to permeate most cults of the bhairavatantras.84 Sanderson makes a number of other contributions to the study of the BraY as well. Perhaps most noteworthy are his reconstruction a of its nine-syllable vidy -mantra, and demonstration that the BraY is one of several a a Saiva sources redacted into the Buddhist Laghucakraamvaratantra.85 In addition, he s . rst identied the reference to the BraY in the old Skandapur na, discussed in chapter a a. two, conrming the likelihood of the texts early period of composition. Several other contemporary scholars have begun to draw on the BraY . In introa ducing her ne critical edition of chapters from the Siddhayogevarmata, one of the s a . few other surviving scriptures of the early Vidy ptha, Judit Trzsk makes reference to the BraY and provides a transcription of a portion of chapter xxxviii (xxxix aca cording to its colophon) from its oldest ms.86 Somadeva Vasudeva also makes limited use of the BraY in his signicant monograph on Saiva yoga and the M linvijayottaraa a tantra. He moreover has made available electronic transcriptions from the oldest ms of two chapters of the text: chapter nine, called laksyabheda,87 and forty-two, the . same chapter published by S. N. Ghoshal Sastri.88 A somewhat different case is that of David Gordon Whites Kiss of the Yogin:
Sanderson, Saivism and the Tantric Traditions, 67072, 67980. Ibid., 672; and Sanderson, History through Textual Criticism, 4146. On the relationship between the BraY and Laghuamvara, see also chapter 3, section 5 in the present thesis. a s . 86 Trzsk, Doctrine of Magic Female Spirits, viixx, 237 (appendix 9). Her remarks on the relative chronology of the Siddhayogevarmata and BraY are mentioned in chapter 4, section 3. ) s a 87 On this complex and interesting subject see Vasudeva, Yoga of the M linvijayottaratantra, 25392. a Vasudevas references to the BraY occur in this context. a 88 Along with much other interesting Indological material, these transcriptions are presently available on his website, (accessed August, 2006).
84 85

28 Tantric Sex in its South Asian Contexts, one of the most recent works drawing upon the BraY . White summarizes and partially translates two passages from BraY a a lxiv,89 material outlining radical ritual practices that involve sexual intercourse. His accounts of these two sections are however highly problematic.90 While the BraY is not of central importance in the work of these authors, their a references attest the wide range of subjects it could potentially illuminate. It is hoped
Kiss of the Yogin, 24850. White also makes reference to the BraY on pp. 17, 23, 101, 163, and 322. a White states that he reads from the oldest Nepalese codex, nak 3-370 (see Kiss of the Yogin, 332, n. 171), yet his bibliography confusingly lists instead a late, corrupt, Devan gar-script ms: Brahmay mala. a a Nepal National Archives. mss no. 1-743. Nepal-German Manuscript Preservation Project reel a166/i. 439 folios. Ibid., 337. The folio numbers provided (260b65bb apparently meaning verso) do not match the text he cites from chapter forty-four in nak 3-370 (which I report as Bya ), but rather the Devan gar-script ms (reported as Byc in my critical edition). a The shorter of the two passages White cites is from BraY xliv (numbered xlv in its colophon). After a arguing, quite implausibly, that mudr can have the meaning vulva, White remarks, a
90 89

. . . The term mudr can, however, denote a substance to be eaten, rather than the a vulva as seal. So, for example, a verse in chapter xlv of the Brahmay mala reads: a tvaakty s dhako nityam yath vibhavasa[mbh]av[ t] mudr m caiva yath ny yam madhyacaiva s aa a a a. a a . . prad payet (But the practitioner who is without a consort [should] constantly [offer] aca cording to what is possible for him. One should also offer mudr , according to the rule, a as well as liquor.) Kiss of the Yogin, 82, 295 (n. 87). It is virtually impossible that an odd-numbered verse-quarter would begin with the enclitic particle tu; White moreover prints madhya (middle) while apparently reading madya (wine), and prints and translates aakty s dhako as though it could be a compound (an aluks aa sam sa?). He has in fact been misled by his choice of (the least reliable) manuscriptByc which reads a tvaakty s dhako nityam yath vibhavasamtavam mudr m caiva yath ny yammagha caiva prad payet. The s a a a a. a a . a . . . text of Bya , which appears original, is as follows: is. v y gam yath ny yam naivedy ni prad payet 648 a a . a a .t a a . svaakty s dhako nityam yath vibhavasambhavam | s a a a . mudr caiva yath ny yam argha caiva prad payet 649 a a a a Having worshipped the pantheon (y ga) as is betting, the s dhaka, together with his a a consort (svaakti), should always make the food offerings to the extent of his capacity; and s he should offer both mudr , as is betting, and the guest water. a BraY xliv.648cd49. Substantial internal evidence in the BraY claries that the expression mudr m a a a. prad payet (one should offer the mudr ) means one should bind/display the mudr , mudr here a a a a having its normal sense of gesture, rather than something consumed. The ritual sequence of rst binding a mudr and then offering argha occurs repeatedly, and the phrasing usually leaves no ama biguity. Cf., e.g., xxxii.168cd (mudr m badhv tato devi argham tasy h prad payet), xxxiii.132ab (mudr m a. a a. a a. . badhv yath ny yam argha caiva prad payet), and xvii (mudr bandham tatah krtv argha c sya prad payet, a a a a a a a a . . . exposure 95l). The more ambiguous expression mudr m [. . . ] prad payet occurs thrice in BraY xii, and a. a a twice elsewhere in chapter forty-four. Due to its length, I will not here reproduce and discuss in full the passage from BraY xliv that a White summarizes and partially translates (Kiss of the Yogin, 24950). Sufce it to say that his account of this ritual of viewing ones [past] births (svayonidarana) has numerous problems. Note for example s Whites rendering of xliv.691ab: He becomes a Virile Hero, surrounded by yonis. Kiss of the Yogin, 250. This implies the Sanskrit vro bhavati so devi yonipariv ritah, presumably Whites emendation of a . Byc s reading (f. 265v): vro bhavati s devi yonpariv ritah. This is of course unmetrical, and Bya is surely a a . correct in transmitting vro bhavati so devi yoginpariv ritah: He becomes a hero, O goddess, surrounded a . by yogins.

29 that the critical editions provided in the present dissertation and planned for future publication will facilitate more extensive use of the text.


a note on conventions

In most instances, I provide the text and a translation of primary sources cited; some exceptions are made for readily accessible texts, especially if the passage is long, not especially problematic, and of no special consequence. When the length of a passage appears disproportionate to its relevance, I sometimes provide the text but forgo translation, especially if the source is well-known; also omitted are translations of some short passages that I paraphrase in the course of discussion. This approach, which I hope has helped curb the volume of annotation, has undoubtedly led to a number of inconsistencies. Sources edited from manuscript appear with a critical apparatus; a list of the abbreviations used has been provided after the table of contents. Part i of this thesis quotes substantially from the BraY . In most cases, the quoted text is constituted a based only on the readings of the oldest codex, National Archives of Kathmandu no. 3-370 (reported as Bya ; see the introduction to part ii for a discussion of the manuscripts). I usually resist the temptation to comment upon the (sometime considerable) linguistic peculiarities of the passages cited, and upon some of the problems of interpretation. Such matters are, however, addressed in detail in the critical edition (part ii). Passages quoted from the BraY are cited by verse number rather than a folio, while appendix A provides a concordance of the texts chapters and the folio numbers in the oldest codex (Bya ). Verse numbers from chapters not included in the critical edition are generally determined by mechanical division of the text into verses of four quarters (p da), and it is likely that the numbering will change slightly a in future editions. In very short quotations from the BraY , an orthographic normala ization (especially the correction of s to s and vice versa), a minor emendation (e.g. a for o or vice versa), or the addition of an anusv ra (m) or visarga (h), is indicated by ena . .

30 closing the character in question in square brackets. A critical apparatus is provided for longer passages, and in cases where more signicant emendations or conjectures are proposed. Translations are the present authors unless indicated otherwise. One undoubtedly quirky convention adopted is the use of a pair of question marks, in parentheses, to mark the beginning and end of a translated passage in which I consider the degree of uncertainty concerning the text or its interpretation particularly high. The rst of the pair is invertedi.e. (). In quotations from commentaries, words from the text commented upon are set in bold typeface. Text citations are always from the edition or manuscript listed in the bibliography. In cases in which more than one edition or manuscript is listed, the source is specied. The Tantrasadbh va, referred to a frequently, is cited on the basis of the manuscript collation and draft edition generously made available in electronic form by Mark Dyzckowski;91 in the few cases that I have consulted the manuscripts myself and differed in their interpretation, this is indicated.

91 Available from the Digital Library of the Muktabodha Indological Research Institute,

Chapter 2
early evidence for the cult of yogins: the literary, i sculptural, and epigraphic sources


introduction: yogins and mother goddesses i

A variety of evidence documents what I have described as the Saiva Yogin cult in the early eighth century c.e. It had moreover almost certainly developed earlier; but there are difculties in determining exactly when, for much of early tantric Saiva literature has been lost, and little has been dated with precision. For establishing a plausible chronological framework, other types of data are thus crucial. In the present chapter, I review the early evidence available in non-tantric literary sources, and the epigraphic and art historical records. The emphasis is upon pre tenth-century sources, but important evidence from the tenth century and beyond is also reviewedparticularly the temples of yogins, with which this chapter ends. In the subsequent chapter, the discussion will turn to tantric texts proper, both Saiva and Buddhist, reviewing the background of the Yogin cult in early tantric literature as well as actual tantras connected with yogins. It will be seen that the balance of ev idence points toward the existence of Saiva yogin traditions in the seventh century certainly in the eighth-and suggests that some of the extant Tantric literature was in circulation then as well, including the Brahmay mala (hereafter BraY ), although a a not necessarily in the form we have it today. Reaching back into the sixth and then fth centuries, the evidence becomes increasingly tenuous, but nonetheless remains substantive. 31

32 Signicant elements of the Yogin cult considerably predate its development. It has been widely acknowledged that yogins have roots in early traditions of Mother goddesses, m trs or m trk s,1 whose qualities they inherit in no small measure. In a. a. a seeking a genealogy of the cult of yogins, the present chapter hence delves in some detail into the historical development of cults of Mother goddesses.2 A number of art historians and historians of religion have analyzed the textual, archeological, and epigraphic evidence for Indian Mother goddesses, including J. N. Tiwari,3 N. P. Joshi,4 Michael Meister,5 Katherine Harper,6 Shivaji Panikkar,7 and others.8 The present chapter reviews much of the same material, but also brings to bear upon the subject some of the rich evidence available in early tantric literature and the old Skandapur na, much of which has only recently become available or remains a. unpublished. The specic aim is to identify with greater precision the relationship of
1 Authors of studies on the Mother goddesses have generally preferred the term m trk , for reasons a. a unclear to me, possibly following contemporary usage or the usage of later Pur nic literature. Although a. both terms occur, m tr is far more common in the early sources and in tantric Saiva literature, and a. preferred accordingly in this study. I frequently use the English translation Mother or Mother goddess in precisely the same sense. 2 Other feminine deities feed into the image of the yogin as well; noteworthy are the yaks or yaksin, . . . vidy dhar, and to some extent the apsaras. Perhaps even more signicant are Sivas ganas: a horde or a . male deities whose theriomorphic or otherwise bizarre forms, multiplicity, variety, and engagement in activities such as warfare are highly suggestive of yogins. The most detailed attempt to trace the early roots of conceptions of yogins is that of David G. White, Kiss of the Yogin: Tantric Sex in its South Asian Contexts, 2766. I discuss the role of the vidy dhar and to some extent yaks later in this chapter, a . in the section on the Brhatkath tradition, but do not otherwise delve in detail into this question; my a . concern is with the actual emergence of a tantric Yogin cult, rather than a genealogy of all concepts that went into forming the image of the yogin. 3 Tiwari assembles much textual and other material concerning Mother goddesses in his commendable monograph, Goddess Cults in Ancient India: With Special Reference to the First Seven Centuries A.D. 4 M trk s: Mothers in Kusana Art. Joshis monograph offers a comprehensive evaluation of the Kusaa. a .. . na-period Mother-goddess statuary. . 5 Regional Variation in M trk Conventions. This important article analyses regional patterns a. a in the development of Mother-goddess iconography in central and northwestern India, tracing the depiction of the Mothers from their early shrines to the static temple door panels of the ninth century and beyond. 6 Iconography of the Saptamatrikas: Seven Hindu Goddesses of Spiritual Transformation. An art historian, Harper provides a comprehensive study of temples of the Brahmanical Seven Mothers through the seventh century, primarily. 7 Saptam trk Worship and Sculptures: An Iconological Interpretation of Conicts and Resolutions in the a. a Storied Br hmanical Icons. Panikkars work, published in 1997, carries forward scholarship such as a Harpers on the Brahmanical cult of Mothers. 8 Two works not specically focused on Mother goddesses but nonetheless highly relevant, particularly concerning the early period, are the dissertations of Yuko Yokochi (The Rise of the Warrior Goddess in Ancient India. A Study of the Myth Cycle of Kau ik-Vindhyav sin in the Skandapur na) s a a. and Richard Mann (The Early Cult of Skanda in North India: From Demon to Divine Son).

33 the cult of Mothers to Tantric Saivism and emergent traditions of yogins. While Mother goddesses are of considerable antiquity in India and undoubt edly had non-elite, probably non-Aryan roots, the Yogin cult appears to presuppose the crystallization of their classical Gupta-era form: a set of goddesses called the Seven Mothers (saptam tr), six of whom are named after and iconographically mira. ror important Brahmanical gods, joined by a seventh, C mundathe independent a . . goddess who is leader of the Mothers (m trn yik ). The six normally consist of a. a a Brahm n, M he var, Kaum r, Vaisnav, V r h or Y m,9 and Indr n, female couna. a s a a a. .. a a terparts of Brahm , Siva, Skanda, Visnu, Var ha or Yama, and Indra. Their numbers a a .. are however sometimes augmented by an eighth goddess, particularly Mah laksm, a . while in tantric Saiva sources, they often are joined instead by the supreme goddess herself, Bhairav, who thus usurps C mundas position.10 a . . Early Saiva treatises on the characteristics of yogins (yoginlaksana)11 classify . . these goddesses according to clans (kula, gotra) that have the Seven or Eight Mothers as matriarchs, clan mothers in whose natures the yogins partake as amsas, por . tions or partial manifestations. Tantric practitioners too establish kinship with the Mothers, leaving behind their conventional clan and caste identities and entering during initiation into what I refer to as initiatory kinship with the deities.12 It is possible that the most fundamental initiation mandala of the Yogin cult comprised ..
Probably the most common variation in sets of the Seven Mothers, as described in texts, is the alternation between Y m/Y my and V r h, the former a counterpart of the death-god Yama, and a a a aa the latter of Var ha, avat ra of Visnu. V r h dominates, however, in sculpted sets of the Mothers. a a aa .. Further research is needed to determine the signicance of these variations. In the BraY , Y my (also a a a called Vaivasvat) features in all cases of ritual or doctrinal importance, V r h being mentioned only in aa sections of the text which might represent a later stratum. See the discussion on the BraY s structure a in chapter 4, and the annotation on BraY ii.16. In contrast, sets of eight m trs show more variation, a a. both in text and image; see the section in this chapter on post-Gupta era m tr shrines. a. 10 See the discussion of post-Gupta era Mother-goddess shrines later in this chapter. On the Mothers in tantric literature, see also chapter 3. 11 These include Siddhayogevarmata 29, BraY lxxiii (edited and translated in the present dissertas a tion), and Tantrasadbh vatantra 16. a 12 A yogin of the clan of Br hm/Brahm n is said to be brahm nyamsa, an amsa of Brahm n. a a. a. . a. . Cf., e.g., Tantrasadbh va 16.253cd. An initiate too is said to be connected to or possess (yukta) an a amsa of a Mother goddess; cf., e.g., BraY lxxiii.47cd, brahm nkulaj devi sv msasiddhiprad yik ([She a a. a a. a a . is] a yogin of the clan of Brahm n, O Goddess, who bestows siddhi upon those [s dhakas] of her own a. a [Mother-goddess] amsa). .

34 Bhairava in a circle of the Eight Mothers. Note for instance that in the BraY , while a the Four Devs and Four Duts form the primary mandalas inner circuit ( varana) a .. . of goddesses,13 surrounding Bhairava and Bhairav/Aghore, the initiatory clans re s main nonetheless those of Bhairava and the Eight Mothers. A concordance in BraY a iv provides the Mother-clan names associated with each of the eight Devs and Duts, as though mapping the mandala positions of the Mother goddesses they supplant.14 .. Initiatory kinship with the Mothers links male practitioners to the yogins, their ini tiatory sisters, seeking visionary, transactional encounters (mel pa) with whom they a undertake powerful and radical rituals. Guided by knowledge of the characteris tics of yogins (yoginlaksana), as described in Saiva scriptural sources, they might . . also seek out clan sisters living among ordinary folk, looking out for behaviors and appearances associated with the Mother goddesses yogins embody: A woman dark and malodorous, who has a long neck and ngers, [whose] teeth have a very beautiful shine and whose eyes are very round, always fond of red clothing, draping a garment from her shoulders, always fond of scents and owers . . . These are the characteristics of [yogins] born in the clan of [the Mother goddess] Indr n. After six months of worship, a.
On the basic pantheon of the BraY , see the section discussing the epithet nav ksaravidh na in a a . a chapter 4. 14 See BraY iv.88898, especially 890cd894ab: a rakt y s tu yad p to gotram m hevaram hitam 890 a a a a . a s . kar l y yad p to brahmagotrah sa ucyate | aa a a a . kar ly y yad p to vaisnavgotrako hi sah 891 a a a a a .. . dantur y yad p tah kaum rgotrasambhav | a a a a . a a . a a candaksy y yad p to vaivasvatikulodbhav 892 a a a .. bhmavaktr prap tena m hendram gotram ucyate | a a a . mahocchusm prap te tu carcik gotrako hi sah 893 a a . a . mah bal yad p tah purangotra ucyate | a a a a . .
891b gotrah ] corr.; gotras Bya 892c candaksy y ] em.; candaksay Bya . . . . a a . . . a gotram ] conj.; m hendragottra (unmetrical) Bya a 893b m hendram a .

When the [ower cast into the mandala during initiation] lands on Rakt , the clan of Siva a .. [i.e. M he var] is enjoined. When it lands upon Kar l , he is said to be of the brahm -clan a s aa a [i.e. the clan of Brahm n]. When it lands on Kar l, he is of the clan of Vaisnav. When a. a .. it lands on Dantur , she is born of the clan of Kaum r. When it lands on Candaks, she a a . . . is born of the clan of Vaivasvat [i.e. Y m]. By landing on Bhmavaktr , the clan of the a a great Indra [i.e. the clan of Indr n] is enjoined. If it lands on Mahocchusm , he has the a. . a clan of Carcik [i.e. C munda]. If its fall [indicates] Mah bal , the clan of The One Who a a a a . . Completes [the Mothers] is enjoined [i.e. the clan of Bhairav]. The gender shift with the feminine sambhav in 892b and 892b is suspect; most probably read a sambhavah. Subsequent verses give a concordance of the Mother-clans and the Six Yogins of the . mandala as well. ..

35 successfully propitiated, they bestow an encounter (melaka). [The practitioner] traverses the entire world, working all [kinds of] wonders.15 Chapters from the BraY concerning the classication of and encounters with yogins a are included in the critical edition, part ii of the present dissertation. That Mother goddesses lie in the background of the Yogin cult is evidenced in a variety of other ways as well. Historical continuity is readily apparent through comparing textual and sculptural representations, beginning, in the case of the Mothers, with the Mah bh rata and sculpture from the early centuries c.e. The present chapa a ter begins with this early level of evidence, and then shifts to Gupta-era and early post-Gupta era sources. There we nd our rst glimpses of a tantric cult of Mothers and the emergence of yogins as sacred gures. Their mutual association is close: for instance, the old Skandapur na, a text probably of the sixth or early seventh century a. (see below), speaks of Tantras of the Mother Goddesses (m trtantras) and identies a. these with a list of Saiva y malatantras that includes the extant BraY , a fundamental a a text of the yogin cult. Examination of period sources helps establish an historical framework for discussion of the Yogin cult in early tantric literature, the subject of chapter three.


early sources

the mahabharata, kusana-period statuary, and the early cult of mother goddesses . .

Sources for study of the pre-Gupta period Mother goddesses are multiple and rich. Among texts, the most important is undoubtedly the Mah bh rata, which is supplea a mented by a substantial body of statuary preserved from the Kusana-period Mathur a . . region (circa 13rd centuries c.e.). Typologies of Mother goddesses in the Mah bh rata a a match well to the sculptural evidence surviving from Mathur . In both instances, as a well as in early medical literature, there are strong associations between the Mother
15 Brahmay mala lxxiii.6771. For the text, and problems concerning its constitution and interpretaa tion, see the critical edition and translation in Part iii.

36 goddesses and the deity Skanda, in the context of whose myths the Mothers appear in the Mah bh rata. a a Mah bh rata accounts of the mythology of Skanda and his retinue of grahas (seiza a ers)16 and Mother goddesses are many and highly layered, no doubt the products of a long development.17 This is illustrated by the complex claims made concerning . Skandas parentage, as related in the Aranyakaparvan, chapters 21521. Most directly, he is the child of Agni and Sv h . Agni, who lusted after the wives of the Seven a a Sages, was seduced by Sv h , who approached him taking on in turn the guises of a a six of the rsis wives, sparing from scandal only the devoted and powerful Arundhat. .. Yet Skanda is also the child of Siva and P rvat, who entered into Agni and Sv h , rea a a spectively, and used them as proxies for producing a son. Skanda is, in addition, the child of the Krttik s, the six stars of the Pleiades, which are in this version of the story a . identied with the six wives of the sages. Skandas birth as narrated in Salyaparvan a 4345 adds the river-goddess Gang to the mix: into Agni fell Sivas seed, and nding it difcult to bear, he sought succour by entering the holy river. She too found the luminous embryo difcult to suffer, and placed it on a mountain peak. There it was spotted and nurtured by the Krttik s. In these basic details, the latter account a . agrees with that of Anuanaparvan 84 and 86. Alongside much that is undoubtedly s ancient, such as Skandas association with the Krttik s, these layered myths appear to a . preserve conicting sectarian claims: a legend asserting Skandas origins from Agni
A variety of terms are used in fact to speak of the entourage of Skanda, with gana (member of . . the group) and graha (seizer) the most encompassing. Aranyakaparvan 219.42 categorizes both the Mothers and male retinue of Skanda under the category skandagraha, Skandas seizers: ye ca m trganah prokt h purusas caiva ye grah h | a . . . a. a. . sarve skandagrah n ma jey nityam sarribhih 42 a a a . . These are a diverse lot, including male and female deities and spirits of every conceivable shape and hue. Theriomorphism is common, much as the six-headed Skanda himself is said to sport the head of a goat. 17 The most comprehensive discussion of Mah bh rata passages describing Skanda and the Mother a a goddesses is that of Mann, Early Cult of Skanda. Mann devotes a chapter each to the three Mah a bh rata sections on the mythology of Skanda. See also his article, Parthian and Hellenistic Inuences on a the Development of Skandas Cult in North India: Evidence from Kusana-Era Art and Coins, Bulletin . . of the Asia Institute 15 (2001): 11128. Mann argues that Skanda has origins as the leader of inauspicious grahas, including the Mothers, only later becoming transformed into the son of Siva and warrior god of Hindu mythology, for which both Brahmanical and royal Greco-Persian inuences are responsible.

37 and the wives of the sages, which brings him within the orbit of Vaidika orthodoxy; and a Saiva layer asserting Skandas parentage from Siva and Um . By the time a of K lid sas Kum rasambhava, an epic poem on Skandas birth probably of the fth a a a century, the Saiva identity of Skanda as son of Siva and P rvat would dominate; a and in the somewhat later Skandapur na, the cult of the warrior-child and the Mother a. goddesses is fully assimilated into Saivism. In the Mah bh rata, there is no evidence of a Mother goddess cult tantric in chara a acter, and there is indeed no reason to believe that Saivism of the type attested in the earliest surviving Tantras had developed. In one description, worship of Skanda, the Mothers, and other deities of his retinue has attainment of longevity and vitality for its impetus, and involves bathing [the deity images], offering incense, ointments, food (bali, perhaps non-vegetarian), and other offerings (upah rapossibly in the sense of a sacrice), and performing Skandas worship (ijy ). Taken together, these suggest a a shrine or temple image-worship context.18 The high deities mentioned in association with the cultRudra, Agni, Um , and Sv h are said to be worshipped by a a a those desiring progeny, an important theme in later accounts of the Mother cult.19 There is in general a strong apotropaic dimension, which comes through most clearly . in the early medical literature,20 as well as Aranyakaparvan 218. Skanda and his subsidiary deities afict children with disease if not propitiated, and the Mothers are

. Aranyakaparvan 219.4344: tesam praamanam k ryam sn nam dhupam ath janam | s a . a . . . a . balikarmopah ra ca skandasyejy viesatah 43 a s a s . . evam ete rcit h sarve prayacchanti subham nrnam | a. . .. a ayur vryam ca r jendra samyakpuj namaskrt h 44 a . . a. For them [the Mothers and Skandas other grahas] must be performed ritual pacication: bathing, incense, ointment, the rite of offering bali and gifts [or sacrice], and particularly the worship (ijy ) of Skanda. All of them [the grahas], thus worshipped, bestow good a luck, longevity, and vitality to people, when paid respects with due reverence, O lord of kings.

This description of the constituents of worship bears comparison with the temple cult of Mothers described in the B gh copper plate inscriptions of the late fourth century c.e., discussed in the next a section. Although little is known of the eras of composition of individual sections of the epic, it is possible that relatively little time separates this portion of the text from the B gh copper plates. a 19 Note, for example, in BraY i the brief narrative of Deik (verses 8184), who worships the Mothers a a desiring to have a child. 20 On the graha cult in medical literature, see Mann, Parthian and Hellenistic Inuences, 57.

38 hence intimately associated with fertility and disease, with life and death. Sculptural evidence from the environs of Kusana-era Mathur largely supports a . . the picture of the Mother-goddess cult that emerges from the Mah bh rata. A signia a cant body of statuary survives from a variety of architectural contexts, not all Brahmanical.21 The Mathur Mother goddesses a are diverse, comprising small-scale images both anthropomorphic and theriomorphic whose common iconographic features include carrying infants, displaying the gesture of deliverance from danger (abhayamudr ), and occurrence in groups of varia able size alongside a male guardian gure, such as the spear-bearing Skanda (see gs. 2.12.5).22 There is however a discrepFigure 2.1: Mother goddess bearing lotus and ancy in one signicant respect: while the infant, making the abhayamudr . Kusana-era a . . Mathur . AIIS Photo Archive. a

Mah bh rata links the Mothers almost exa a

clusively with Skanda, the extant statuary preserves an association with Kubera as well, god of wealth and lord of the yaksas (cf. figure 2.3).23 This parallel and pos. sibly more archaic convention illustrates the Mothers close links to the yaksa and .
N. P. Joshi notes evidence for Mother sculptures recovered from both Buddhist and Jaina sites in th Mathur area. M trk s: Mothers in Kusana Art, 11015. a a. a .. 22 Joshi surveys the iconography of the extant statuary, classifying the images into thirteen categories. The most common type depicts a row of seated Mother goddesses holding infants. Ibid., 10328. 23 According to Joshis iconographic survey, images of the Mothers with Kubera are in fact more than twice as common as those with Skanda. See ibid. Interestingly, Meister describes a saptam trk panel at a. a Kekind (Nlakanthe vara temple, mid-tenth century) in which Kuberas association with the Mothers .. s surprisingly resurfaces: anking the Seven Mothers are Gane a and Kubera. Meister however interprets . s this as the overlapping of two conventions: Gane a as head of a M trk set and Gane a paired with s a. a . . s Kubera as good-luck charms. One set of images containing Gane a has suggested the other; there is no . s other reason for Kubera, lord of wealth, to join the M trk s. Regional Variations, 245, and g. 5. a. a

39 yaks, popular non-Brahmanical deities connected to the natural world, who, like the . Mothers, are well represented in pre-Gupta era myth and sculpture.

Figure 2.2: Skanda (left) and ve m trs. Kusana-era Mathur . State museum, Mathur . AIIS Photo a. a a . . Archive.

Two signicant continuities have been insufciently emphasized between early Mother goddesses and the Brahmanical Seven Mothers of the Gupta-period. These bear directly upon the question of the emergence of the Yogin cult. First, as sev eral scholars have noted, the number seven has strong precedent:24 Mah bh rata, a a . Aranyakaparvan 217, lists by name a heptad of goddesses known as the sium trs, s a. Mothers of the Infant[s]. These comprise K k, Halim , Rudr , Brhal, Ary , Pal l , a a a . a aa and Mitr .25 It seems possible that the Brahmanical Mothers directly supplant the a sium trs, who might well have been popular deities of the pre-Gupta period. Note s a. also the Salyaparvans reference to saptam trganah, an ambiguous compound perhaps a . . .
24 Cf., e.g., Mann, Early Cult of Skanda, 3738; Harper, Iconography of the Saptamatrkas, 56; and Meister, Regional Variations, 240. 25 Aranyakaparvan 217.9: . k k ca halim caiva rudr tha brhal tath | a a a a . a ary pal l vai mitr saptait h sium tarah 9 aa a a. s a .


Figure 2.3: Kubera and two m trs. Kusana-era Mathur . AIIS Photo Archive. a. a . .

Figure 2.4: Seven m trs, with Skanda (left). Kusana-era Mathur . State museum, Mathur . AIIS Photo a. a a . . Archive.

41 meaning the groups of Seven Mothers, the plural suggesting multiple groups of seven.26 Furthermore, from the Kusana-era (circa 13rd centuries c.e.) Mathur envia . . rons survives what appears to be a set of seven Mothers with Skanda (figure 2.4); but this unique panel is broken on the right and might hence have contained additional images.27 Second, although there is no clear evidence in the Mah bh rata for the Gupta-era, a a classical set of Seven Mothers, the classication of Mother goddesses into groups corresponding to Brahmanical deities is attested in the Salyaparvan. This is of considerable consequence, illuminating the roots of the Brahmanical Seven Mothers and suggesting an ancient precedent for the Yogin cults organization of female deities into clans having each of the Seven or Eight Mothers as matriarch. Salyaparvan 45 presents a rich account of the diverse Mother goddesses, in the course of which it describes them variously as y myah, raudryah, saumy h, kauberyah, v runyah, m hendryah, a a. . . . a . . a . agneyyah, v yavyah, kaum ryah, and br hmyah.28 These are abstract nouns formed a . a . a . . from the names of the male deities Yama, Rudra, Soma, Kubera, Varuna, Mahen. dra/Indra, Agni, V yu, Kum ra/Skanda, and Brahm , the passage hence providing a a a strong evidence for organization of the Mothers according to deities of the Brahmanical pantheon. It must be emphasized just how much yogins as a deity typology inherit from the Mothers, as described in this Salyaparvan passage. Among the Mothers, some have long claws, fangs, or beaks; some are youthful maidens, while others are eshless or pot-bellied. Having various hues, changing shape at will, and speaking many languages, the Mothers rival the apsaras in beauty, Indra in power, Agni in radi26 Salyaparvan 43.29ab: saptam trganas caiva sam jagmur viam pate. M trganah might however be a a . . a s . a . . . karmadh raya compound, meaning the ganas who are the [Seven] Mothers, or even a dvanda, the a . [Seven] Mothers and the ganas. The context is a list of divinities who come to see Skanda. . 27 J. Bautze claims in fact that all Kusana-era seated m tr-goddess panels so far published are fraga. . . ments, broken at one end or both. A Note on Two M trk Panels, 25. a. a 28 Not accepted in the critical edition are, in addition, the epithets vaisnavyah, sauryah, and v r hyah, aa .. . . . in a verse that would follow 45.36ab. This might have been interpolated to harmonize the passage with later conceptions of the Mothers; the absence of Vaisnav and V r h, in particular, might have been aa .. inexplicable to a Gupta-era or later audience. Yokochi quotes and discusses this Mah bh rata passage a a in Rise of the Warrior Goddess, 101.

42 ance, and so forth.29 They dwell in liminal places such as crossroads and cremation groundsthe same environs enjoined for performing the radical rituals of the Yogin cult, one of the primary aims of which was to effect direct encounters with goddesses. The yogins theriomorphism, shapeshifting, multiplicity, extraordinarily variegated appearances, bellicosity, independence, and simultaneous beauty and danger all nd precedent in these early Mother goddesses. This continuity is readily visible in sculpture. While taking on the powerful iconography of tantric deities, the yogins reect in visual terms clear continuity with the Kusana-era Mother goddess typology (cf. . . figures 2.5 and 2.6).30 Much as there is continuity between yogins and the early m trs, dangerous and a. powerful female deities whom, as Michael Meister suggests, the Brahmanical cult of

Salyaparvan 45.2940: et s c ny s ca bahavo m taro bharatarsabha | a a a a . k rttikey nuy yinyo n n rup h sahasraah 29 a a a a a a. s . drghanakhyo drghadantyo drghatundya ca bh rata | a .. s saral madhur s caiva yauvanasth h svalamkrt h 30 a a a. . . a. m h tmyena ca samyukt h k marupadhar s tath | a a a. a a a . nirm msag tryah svet s ca tath k canasamnibh h 31 a. a a a a. . a . krsnameghanibh s c ny dhumr s ca bharatarsabha | a a a a .. . . arunabh mah bh g drghakeyah sit mbar h 32 a a a s . a a. . a a . urdhvavendhar s caiva ping ksyo lambamekhal h | a a. . lambodaryo lambakarnas tath lambapayodhar h 33 a a. . t mr ksyas t mravarnas ca haryaksya ca tath par h | a a . a a a. . . s varad h k mac rinyo nityapramudit s tath 34 a. a a . a a y myo raudryas tath saumy h kauberyo tha mah bal h | a a a. a a. v runyo tha ca m hendryas tath gneyyah paramtapa 35 a . a a . . v yavya c tha kaum ryo br hmya ca bharatarsabha | a s a a a s . . rupenapsaras m tuly jave v yusam s tath 36 a. a a a a parapus. opam v kye tatharddhy dhanadopam h | a a a a. .t sakravryopam s caiva dpty vahnisam s tath 37 a a a a vrksacatvarav sinya catuspathaniketan h | a s a. . . . guh smaanav sinyah sailaprasravanalay h 38 a s a . . a. n n bharanadh rinyo n n m ly mbar s tath | a a a . a a a a a a . n n vicitravesas ca n n bh sas tathaiva ca 39 a a a a a. . ete c nye ca bahavo ganah satrubhayamkar h | a a. . . . anujagmur mah tm nam tridaendrasya sammate 40 a a . s .

Yogins shed the Mothers maternal associations to a large degree. However, there are numerous examples of their representation with infants; note for instance two of the yogins from Lokhari, U.P., published in Dehejia, Yogin Cult and Temples, 159 (which I reproduce as figure 2.6), 161. Dehejia publishes another image of a theriomorphic yogin from Naresar that carries an infant, labelled Um a Dev. Ibid., 14647. An image (10th century?) from Bundelkhand, M.P., of a four-armed goddess holding a child also appears to be a yogin, presently in the Bharat Kala Bhavan of Varanasi (aiis Photo Archive, accession no. 7175, negative no. 4).



Figure 2.5: Bird-headed Mother goddess carrying an infant in her likeness. Kusana-era . . Mathur . State museum, Mathur . AIIS Photo a a Archive.

Figure 2.6: Horse-headed yogin from Lokhari, U.P., with like infant. Photograph by Vidya Dehejia, published in Dehejia, Yogin Cult and Temples, 159.

Seven Mothers sought to contain,31 the early Skanda, though succesfully transformed into a benevolent warrior god, resurfaces in the gure of Bhairava, lord of yogins, who takes on much of the ancient imagery of Rudra as well. For although a playful, handsome young warrior dominates the later image of Skanda, in the Mah bh rata a a mythology lie clear traces of an ambiguous and potentially dangerous deity, in this respect resembling the m trs he heads. Richard Mann argues that this Skanda has in a. fact historical primacy.32

Meister, Regional Variations, 24445. Mann, Early Cult of Skanda, passim. It is possible however that Mann goes too far in reading historical layers into the Skanda myth. Specically, I see little reason why the Mothers and Skanda could not have been, even in their earliest conceptions, simultaneously auspicious and inauspicious, connected with both fertility and disease, life and death. In other respects, his argument for historical transformation seems entirely plausible.


gupta-era sculpture and inscriptions

Several fourth- and fth-century, Gupta-era inscriptions make reference to Mother goddesses, while the earliest remains of Mother shrines appear to date to the beginning of the fth century. In this period, we nd evidence for the emergence of the Brahmanical saptam trs, Seven Mothers, female counterparts to a series of major a. Brahmanical deities headed by an independent goddess, C munda. There is appara . . ent, moreover, a transformation by which Siva usurps Skandas position as leader of the Mothers, to the extent that Skanda rarely nds place in the iconographic programs of post fth-century Mother shrines. The goddess C munda is normatively a . . depicted as an emaciated and powerful hag whose iconography includes mortuary (k p lika) cult objects such as the skull and skull-staff (khatv nga). This variety of a a . a iconography is characteristic of tantric deities of the bhairavatantras and early Yogin cult, and it is possible that C munda was, from her obscure beginning, a tantric deity. a . . That is to say, the emergence of C munda could in itself be an indicator of the exisa . . tence of a tantric goddess cult, perhaps even some form of the Yogin cult, although this is not at all certain.33 In the elite traditions represented in sculpture and inscriptions of the Gupta and early post-Gupta period, the Seven Mothers appear to eclipse the more diverse Moth ers popular in the Kusana era. Inscriptions associate the Udayagiri Saiva cave com. . plex of the early fth century with the emperor Candragupta ii, a site having multiple sets of the Seven Mothers.34 Another royal Gupta inscription, that of the mid fth-century Bihar Stone Pillar erected by Skandagupta, also appears to include a
It is conceivable that the m trtantras (Tantras of the Mother Goddesses) mentioned in some Saiva a. sources were connected with a tantric cult of C munda and the Mothers. However, perhaps the earliest a . . source to mention thesethe old Skandapur naidenties them with the y malatantras of the Saiva a. a Yogin cult. On the Skandapur na, see the subsequent section. One possible preservation from an early a. tantric cult of C munda is the love magic of Indian erotic literature (k maastra); see Gyula Wojtilla, a a s . . Vakarana Texts in Sanskrit K ma astra Literature, in Teun Goudriaan, ed., The Sanskrit Tradition and s a s . Tantrism, 10916. 34 The two inscriptions associated with the Udayagiri cave temples are published as nos. 7 and 11 in Corpus Inscriptionum Indicarum, vol. iii, as well as in D. C. Sircars Select Inscriptions. The dated inscription ends with a profession of devotion to Siva (bhakty bhagavata sambhor gguh m et m ak rayat, a s a a a He had this cave [temple] made out of devotion to Lord Siva).

45 profession of devotion to the Mothers and Skanda.35 Later, the C lukya monarchs a would claim descent from the ancient Mother goddess H rit, claiming also to have a been made prosperous by the Seven Mothers, who are the mothers of the seven worlds, and to have enjoyed the protection of Skanda.36 Such royal patronage of the cult of the Mothers nds attestation in the numerous cave shrines and stone temples which survive from the fth century and beyond. Alongside these, however, must have persisted more humble forms of Mother-goddess worship; tantric literature speaks of Mother shrines in isolated places as though, like the crossroads, jungles, and cremation grounds they are mentioned alongside, they were an integral part of the landscape.37 Perhaps the earliest unambiguous evidence for a temple cult of the Mothers, as well as their association with Saivism, comes from Gupta-period inscribed copper plates recovered from B gh, M.P. Two plates from the second half of the fourth cena tury mention endowments made in favor of Mother goddess temples.38 One records the gift of revenue from two villages and a plot of land for maintaining the worship of the Mothers at Navatataka, installed by the same royal patron. Recording a land. grant made for the support of a shrine of the Mothers established by a P supata a a Ac rya Lokodadhi,39 the second is dated a few years later to the year 375/76 or 376/77 c.e. The endowment makes provisions for funding ongoing worship of the Mothers, described as involving bali and caru (both normally consisting of food offerings), the ambiguous sa[t]tra, sacrice, and offerings of incense, scents, and owPublished as inscription 49 in Corpus Inscriptionum Indicarum, vol. iii, 7278. a From the Navsari plates of Yuvar ja Sry srayal ditya, found in the Surat district of a s a modern Gujrat: h rtputr nam saptalokam trbhis saptam trbhi[r abhi]varddhit sa [=varddhit n m, ed.] a a. . a. a. a a a. k rttikeyapariraksanapr ptakaly naparampar nam . . . caliky n m . . . . Corpus Inscriptionum Indicarum, vol. a a a. a. . a a . . . iv. A number of other C lukya inscriptions invoking the Mothers were published by John F. Fleet in a Indian Antiquary, vols. vivii. 37 See the discussion of m trs in the Niv satattvasamhit in chapter 3. a. s a . a 38 The relevant plates are numbers ii and x in the collection published by K. V. Ramesh and S. P. Tewari, A Copper-plate Hoard of the Gupta Period from Bagh, Madhya Pradesh, 46, 2123. Scholars who have discussed these include Sanderson, Religion and the State: Initiating the Monarch in Saivism and the Buddhist Way of Mantras (forthcoming); Yokochi, Rise of the Warrior Goddess, 110 (n. 83); and B. D. Chattopadhyay, Reappearance of the Goddess or the Brahmanical Mode of Appropriation: Some Early Epigraphic Evidence Bearing Upon Goddess Cults, 25758. 39 Sanderson, Religion and the State: Initiating the Monarch, 1516.
36 35

46 ers.40 Neither inscription indicates whether the temples housed the Seven Mothers or a more archaic conguration. However, the latter temples P supata afliation a suggests the possibility of a Saiva iconic program, such as comes into sculptural evidence in the fth century. At Udayagiri, in the Vidi a district of Madhya Pradesh, two s Saiva cave temples dating to the turn of the fth century incorporate niche-shrines of the Seven Mothers, and one of the sets might have been headed by an image of Skanda;41 another temple (cave no. 3) does house Skanda as its cult image. This site appears transitional, maintaining the Mothers old associations with Skanda, but within the context of the worship of Siva. Later in the same century, a cliff-cut shrine between Badoh and Pathari, also in modern-day Madhya Pradesh, would eschew all associations with Skanda and Kusana-era guardian gures, containing images of . . the Seven Mothers in the company of a seated, urdhvaretas Siva alone. In K lid sas a a Kum rasambhava, probably also of the fth century, the Mothers feature in the retinue a of Siva with no special connection to Skanda.42 Among the early inscriptions, the most signicant for the history of the Yogin cult is the well-known Gangdh r inscription of the vikrama-year 480, that is, 423/24 a or 424/25 c.e.43 This records the construction of a Visnu temple, a temple of the .. a . Mothers (m tnam veman), and a drinking well by Mayur ksaka, a minister of the a r. . s . monarch Vi vavarman, at the central Indian cite of Gangdh r in western M lwa diss a a trict. The inscription describes the temple as extremely terrible and lled with
According to lines 36 of plate x, the land grant in question is to provide for revenue ( bhog ya) to be used (upayojya) for bali, caru, satra, incense, scented pastes, owers, and garlands a for the Mothers of the temple of Picchik naka village, established by the revered P supata teacher a a Lokodadhi (bhagavallokodadhip supat c ryapratis. h pitapicchik nakagr mam trsth nadevakulasya . . . deva a a a a a. a .t a a a agr h ram tnam balicarusatradhupagandhapuspam lyopayojyabhog ya; puspa is the editors emendation a r. . a a . . . of pujya ). Ramesh and Tewari, A Copper-plate Hoard, 22. 41 Katherine Harper, Iconography of the Saptamatrikas, 7579. Although I am unable to conrm her reading of the iconography, Harper sees characteristic features of Skanda in the gure on the left wall abutting the row of Mother goddesses, in the shrine outside of cave no. 3. Ibid., 76. 42 It is noteworthy that the m trs, mentioned in canto vii (3031, 38) as part of the wedding entourage a. of Siva, are a group of unspecied number. They are followed in the entourage by K l, whose a ornaments are skulls (kap l bharana, 39b). aa . 43 This inscription was rst published by John F. Fleet, Corpus Inscriptionum Indicarum, vol. iii, inscription 17 (pp. 7278). D. C. Sircar published a subsequent edition in Select Inscriptions Bearing on Indian History and Civilization, vol. i, 399405.

47 dakins, while depicting the Mothers as they who make the oceans tumultuous . through powerful winds arising from tantras (tantrodbhuta ).44 This inscription thus associates the Mothers simultaneously with hordes of female spirits (dakins), magi. cal powers, and a temple cult, also providing early and signicant occurrences of the terms dakin and tantra in the context of Mother-goddess worship. . Dakin, probably connected with the Sanskrit verbal root d, to y,45 and the . . basis of the modern North Indian term dain, witch, denotes a class of female spirits . prominent in taxonomies of yogins. While tantric Saiva sources generally speak of the dakin as a pernicious being, the term is often perfectly synonymous with yogin, . especially in the yogintantras of later Tantric Buddhism.46 As for the inscriptions use of the word tantra, this is probably, as D. C. Sircar recognized,47 in the well-attested sense of spell, such as in the expression tantramantra.48 It seems improbable that the word could refer here to Tantric scripturepotentially m trtantras or dakintantras49 a. . for powerful winds (prabalapavana) would not in normal usage be described as having arisen (udbhuta) from texts.50

Verse 23 (on lines 3537): m trna ca [pramu]ditaghan tyartthanihr dinn m a .. a a a tantrodbhutaprabalapavanodvarttit mbhonidhn m a a . gatam idam dakinsamprakrnnam .. . . vem tyugram nrpatisacivo []k rayat punyahetoh 23 s a a . . . .

Adelheid Herrmann-Pfandt discusses the derivation of the word in Dakins: zur Stellung und Symbo. lik des Weiblichen im tantrischen Buddhismus, 11516. The etymological link to the root d or dai is tra. . ditional; for example, Bhavabhatta and Jayabhadra, commentators on the Buddhist Laghuamvaratantra, s . . . both connect the word dakin to dai. See Bhavabhatta ad Laghuamvara 1.2, Sarnath edition, p. 6; and s . .. . . also Jayabhadra commenting on the same verse, p. 107 in Sugikis edition of the Cakrasamvarapajik . a . 46 For descriptions of the dakin as a dangerous variety of female spirit, cf., e.g., BraY lv (12, 43 a . 44), xcviii (3839), and xcix (1012), and the denition Ksemar ja quotes of the rudradakin from the a . . Sarvavratantra, ad Netratantra 2.16. See also chapter 3 of this thesis, n. 29. On the general synonymity of yogin and dakin in Tantric Buddhism, note for example that the scriptural class often referred to as . yogintantras has as one of its earliest and most authoritative texts the Sarvabuddhasam yogadakinj la a . a samvara, dakinj la referring to the matrix of female deities. Cf. Prapacayoginj la and Yoginj la, titles of a a . . a lost Saivas texts mentioned in BraY xxxviii.39cd. The expression yoginj lasamvara, incidentally, occurs a a . several times in BraY lviii, while BraY lvi teaches an observance (vrata) by the same name. a a 47 Sircar, Select Inscriptions, vol. i, 405. 48 Cf., e.g., M latm dhava ix.52, quoted later in this chapter in the discussion of this work. a a 49 On m trtantras, see the subsequent section on the Skandapur na, and chapter 5, in the section a. a. discussing the title Brahmay mala. Dharmakrti makes reference to dakintantras, on which subject see a . Sanderson, History through Textual Criticism, 1112; and chapter 3 of this thesis. 50 My interpretation of this passage undoubtedly has been inuenced by Isaacsons remarks on the subject, in a lecture given at the University of Pennsylvania in January 2003.


48 The Gangdh r inscription does not provide unambiguous evidence for a tantric a cult of goddesses; for this, we would need indication of the ritual practices associated with the Mother-goddess temple, or period textual evidence intimating tantric connections. Nonetheless, the inscription remains highly suggestive. Dakins, as a . variety of goddess or spirit, are in later literature closely associated with Tantra, and the description of the Mothers themselves uses imagery suggestive of powerful, unfettered tantric goddesses,51 not at all in the image of the protective World Mothers (lokam tr) mentioned in other Gupta-era inscriptions. It is accordingly possible that a. the Gangdh r inscription records the existence of a tantric goddess cult in the fth a century, perhaps even a Yogin cult similar to that attested in Tantras such as the BraY . At the least, it shows that characteristic elements of the conceptions of fea male divinities prominent in the Yogin cult had come together by the early fth century. Unfortunately, the Gangdh r inscription is exceptional: we have no other a rmly dated evidence for a cult of Mother goddesses in the company of female spirits in the fth century, which makes the inscription difcult to contextualize. The iconic program suggested, featuring Mothers and a host of minor goddesses, does not come into evidence again in temple contexts for many centuries. It is unclear whether the temple housed the Brahmanical Seven Mothers, while its possible association with a Visnu temple could suggest a non-Saiva cultic context. ..

. the skandapurana: yamalatantras, yoge vars, and the mother goddesses of s i kotivarsa . .

The signicance of the early Skandapur na52 for the history of Saivism and early mea. dieval Indian religion can hardly be overstated, as the recent studies of Hans Bakker
Borrowing an expression from the title of an article of Nilima Chitgopekar, The Unfettered Yogins, in Chitgopekar, ed., Invoking Goddesses: Gender Politics in Indian Religion, 82111. 52 The early or old Skandapur na should not be confused with the better-known published text a. by this name; the latter was in fact somewhat artically assembled by pandits in the colonial period .. from various medieval tracts having the Skandapur na as locus of ascription. See Rob Adriaensen, et al, a. in introducing vol. 1 of the critical edition of the early Skandapur na. a.

49 and Harunaga Isaacson,53 Yuko Yokochi,54 and Peter Bisschop illustrate.55 Preserved in manuscripts that include a Nepalese codex of 810 c.e., scholars working on this text appear at present to concur on the probability of a sixth- or early seventh-century c.e. date.56 Its material is oriented toward an audience of Saiva laity, the m hevaras, a s perhaps communities connected with P supata ascetics.57 While not hence specia cally concerned with tantric forms of Saivism, the Skandapur na nonetheless provides a. important early evidence for the Mantram rga, including the cult of yogins. Signifa icantly, it attests the existence of y malatantras, described as Tantras of the Mother a Goddesses (m trtantras), and lists the BraY among them. This constitutes highly a. a signicant evidence concerning the dating of the BraY , discussed in chapter four, as a well as signicant early evidence for the Yogin cult. The historical importance of the Skandapur nas reference to y malatantras was a. a discussed rst by Sanderson, in correspondence quoted by R. Adriaensen et al. in introducing the Skandapur na critical edition.58 In a subsequent article, Sanderson a. added further reections on the passage in the course of reviewing early evidence for the Saiva Mantram rga, i.e. Tantric Saivism.59 A fuller discussion nonetheless seems a worthwhile, specically examining its relevance to the cults of Mother goddesses and yogins. It is in the second half of Chapter 171 that the relevant material occurs,
See especially the study of the Skandapur nas V r nasm h tyma, in vol. 2 of the Skandapur na a. a a. a a a. critical edition, Bakker and Isaacson, eds. For Bakkers several other contributions, see the bibliography. 54 Yokochis doctoral thesis, Rise of the Warrior Goddess, focuses upon the formation of the Hindu Warrior Goddess on the basis of material from the Skandapur na, of which she also edits several a. chapters. See also Yokochi, Mahisasuramardin Myth and Icon. Studies in the Skandapur na ii. a. . 55 Early Saivism and the Skandapur na: Sects and Centers. a. 56 Announcing the Skandapur na editorial project, Adriaenson, Bakker, and Isaacson had proposed a a. tentative date of the 7th8th centuries. Towards a Critical Edition of the Skandapur na, 328. In the a. rst volume of the critical edition, however, the editorial team suggested the 6th8th centuries as the most plausible range of dates. Skandapur na, vol. i, 4. Yokochi, as will be discussed, on the basis of a. the iconographic type of the Warrior Goddess in the Skandapur na, subsequently argued for the 6th7th a. centuries as the most plausible period of composition. Mahisasuramardin Myth and Icon, 6875. . Taking into account Yokochis assessment, and on the basis of their own studies on the Skandapur nas a. V r nasm h tmya material, Bakker and Isaacson have more recently suggested the probability of the a a. a a sixth or early seventh centuries. Skandapur na, vol. ii, 48. a. 57 R. Adriaensen et al, Skandapur na, vol. i, 4. Concerning provenance, Bakker and Isaacson remark a. that it has a certain probability per se that the Skandapur na was composed either in V r nas, or in a a. a a. (P supata) centre that had close contacts with this city. Skandapur na, vol. ii, 48. a a. 58 Skandapur na, vol. i, 7. a. 59 History through Textual Criticism, 11.

50 within a m h tyma-narrative on the sacred site Kotivarsa, identied by Sanderson a a . . as being in the West Din jpur District of Bengal and identical to Devkota.60 This a . chapter picks up on several themes from the Mah bh rata, beginning with Skandas a a enmity with Indra, whose place as foremost of warrior gods the divine youth usurps. Present also is the tale of Skandas violent assault upon the mountain Krauca, this being perhaps an accretion from the mythology of the famous mountain-splitter Indra. Another continuity is the motif of enemies of the gods becoming powerful by the boon of Brahm , necessitating unusual means for their defeat. But perhaps most a notable in the Skandapur nas narrative is the fact that Skanda himself disappears a. halfway, as the narrative of Kotivarsa commences, much as he disappears from the . . iconic program of Mother goddess shrines by the sixth century. The chapter begins with the story of Skandas rivalry with Indra and his decapitation of the mountain Krauca, which turns out to be an act of fratricide. The rst of its two sections ends with Siva and P rvat returning home to Mt. Meru accoma panied by Skanda, who is described as leader of the horde of Mother goddesses (m tnam ganan yakah, 73d). Yet the next section opens with the sage Vy sa asking a r. . a . . a . how Siva, not Skanda, became leader of the Mothers (m tnam n yakah, 78b); in the a r. . a . . subsequent narrative, Skanda makes no appearance. Below I summarize the episode, Skandapur na 171.78137, translating in full the most relevant section:61 a. Vy sa asks the sage Sanatkum ra to narrate how Siva came to be leader of a a the Mother goddesses; how, why, and by whom the Mothers were created; what powers they possess; and what their locus is. [7879] Sanatkum ra a narrates how Brahm once came upon a pleasant locale on the banks of a the Eastern Ocean (purvamahodadhi) and performed his sandhy -prayers a there for ten-million years. [8083] Admiring the surroundings, he decided to create a beautiful, gilded city there, to whose fortunate residents he guaranteed immortality and power. [8491] The city is named Koti. varsa since Brahm , best of the gods, showers (varsati) a crore (koti) of a . . . desired wishes (is. an m k m n m) upon its happy population. [9293ab] . t a . a a a
60 . According to Sanderson, other names for this site include Devkotta, Srptha, Srkota, and So . . . nitapura, this being located on the bank of the Punarbhav river. Its association with Siva as Hetuka a . or Hetuke vara is attested in several sources, including BraY iii. History through Textual Criticism, s a 7. Cf. Yokochis discussion of the location of the site in Rise of the Warrior Goddess, 1067 (n. 79). 61 Compare with Yokochis summary of verses 100cd16ab, in Rise of the Warrior Goddess, 107.

51 He then returns to his heavenly court, after which the city is overrun by Asuras, led by Asukrodha, who thus became unintended beneciaries of Brahm s blessing. [93cd97] Learning of what the Asuras were doing, all a the gods convened to inform Brahm . Brahm pleaded inability, and they a a proceeded thence to approach Siva at the Himalayan forest where P rvat a had performed penance. [98100] When the gods entered there, led by Brahm , they suddenly became women a by the power of P rvats penance.62 Siva asks why they have come, in a response to which they tell of the torment inicted by the Asuras. They also plead to be rid of the miserable condition of being women (krcchram . . strbh vam). [1014ab] Siva however tells them to approach the Asuras as a they are, for the proud demons may be slain only by women. Afterwards, the auspicious Mothers will return to their prior forms. [104cd6] The gods then bow to Siva and request that he too join them as a woman, with whom they would slay the demons. Siva then created the auspicious goddess Rudr n, as well as an ugly Mother called Bahum ms (Very a. a. a Fleshy), the embodiment of universal destruction (jagatsamh rarupin). . a . [1079] Brahm , Siva, Skanda, Visnu, and Indra create the Mother goda .. desses Brahm, Rudr n, Kaum r, Vaisnav and V r h, and Indr n, re a. a aa a. .. spectively, with Sivas Bahum ms their leader, the Great Vidy -mantra.63 a. a a And from all the other gods too emerged Mothers, possessing their natures and power, slayers of demons: V yav, V run, Y my , Kauber, a a . a a Mah k l, Agney,64 and others by the thousands. [11014] The Mothers a a proceed to the beautiful city and render it free of demons. When the demons had been slain, Siva came to the city to grant boons to the Mothers. Pleased, he told them,65 Having become Mother goddesses,

Emphasis added. Skandapur na 171.101: a. pravis. as tatra te dev brahm dy h sarva eva hi | a a a. . t . . a striya ev bhavan turnam p rvaty s tapaso bal t 101 a a a Skandapur na 171.112: a. sarvatejomay dev m tnam pravar subh | a r. . a a . bahum ms mah vidy babhuva vrsabhadhvaj t a. a a a a .. 112


I.e. Mothers corresponding to V yu, Varuna, Yama, Kubera, Mah k la, and Agni, respectively. a a a . This passage and its list bears comparison with the classication of Mothers according to Brahmanical deities in Mah bh rata, Salyaparvan 45, quoted and discussed earlier. a a 65 From verse 118cd forward, the remaining text is a direct translation. Skandapur na 171.116cd36, a. Bhattar edition: a .. atha daityair hatais sarvair devadeva um patih 116 a . a ajag ma pradesam tam m tnam varaditsay | a . . . a r. . . t n drs. v nihat n sarv n daity n amaravidvisah 117 a ..t a a a a . . paritus. as tad t s m var n pr d d vrsadhvajah | a a a. a a a .. .t . . a. a jagato m taro yuyam m trbhut bhavisyathah 118 a . . yusm kam ye bhavisyanti bhakt h purusapumgav h | a. a. . a . . . . striyo v pi mah bh g na t n himsanti himsak h 119 a a a a a a. . . mrt mama ganas c pi bhavisyanty ajar mar h | a a. . a . a .


52 you shall be the Mothers of the world. Those who will be devoted to you, whether the best of men or fortunate women, pernicious spirits will not harm; and after death, they shall become my ageless, immortal ganas. [11519ab] This place of yours shall become world-famous, known . as Kotivarsa, which frees one of all sin. And since I am your cause . . (hetu), because you were created by me, I will remain here by the name Hetuke vara, granting boons. I shall dwell with you as your leader. s [119cd22] One who will worship you properly, together with me, shall, free of all sin, attain to the highest heavenly destination. Since the d navas a were slain with a spear (ula) by Bahum ms , this sacred ford (trtha) shall s a. a be known by the name sulakunda (Pool of the Spear). And that best of .. men who drinks here from the Pool of the Spear and prostrates before Bahum ms shall be unassailable by all harmful spirits. The beautiful a. a a river Mand kin shall be known here as Pratikul (River Contrary); she a 66 [12326] will always be brimming with blood for you.
bhavatn m idam sth nam kotivarsam iti srutam 120 a a . . . . bhavisyati jagatkhy tam sarvap papramocanam | a . a . aham hetur hi yusm kam yasm t srs. a mayaiva ca 121 a . . t . . a . hetukevaran mn ham sth sy my atra varapradah | s a a . a a . yusm bhih saha vatsy mi n yakatve vyavasthitah 122 a a . a . . yas tu yusm n may s rdham vidhivat pujayisyati | a a . a . . sarvap pavimukt tm sa par m gatim apsyati 123 a a a a. d nav nihat yasm c chulena bahum msay | a a a a a. a sulakundam idam n mn khy tam trtham bhavisyati 124 a . .. . a a . . iha sulodakam ptv bahum ms m pranamya ca | a. a. . a . adhrsyah sarvahimsr nam bhavisyati narottamah 125 .. . . a. . . . pratikuleti vikhy t ramy mand kin nad | aa a a rudhiraughavat seha bhavatn m bhavisyati 126 a. . aham brahm ca visnu ca rsaya ca tapodhan h | a s .. s a. . .. m trtantr ni divy ni m tryajavidhim param 127 a. a. a a. . puny ni prakarisy mo yajanam yair av psyatha | a . a . a . br hmam sv yambhuvam caiva kaum ram y malam tath 128 a a . a a . a . . . s rasvatam sag ndh ram aianam nandiy malam | a a a s . a . tantr ny et ni yusm kam tath ny ni sahasraah 129 a. a a a s . . a . bhavisyanti nar yais tu yusm n yaksyanti bhaktitah | a . . a . . nar nam yajam n n m var n yuyam prad syatha 130 a. . a a a. a . a divyasiddhiprad devyo divyayog bhavisyatha | a a . y s ca n ryah sad yusm n yaksyante sarahasyatah 131 a a . a . a . . yogevaryo bhavisyanti r m divyapar kram h | s a a a a. . chagalah kumbhakarna ca madyau ganan yakau 132 . . s . a yusm kam dv rap latve sth syatas tau mam jay | a . a a a a a . kotivarsam idam sth nam m tnam priyam uttamam 133 a . a r. . . . . . smaanam pravaram divyam bhavisyati sukhapradam | s . . . . var n evam hi t ml labdhv m taro lokam tarah 134 a a. a a a . . bhakty pranamya deveam mumudur bhrsam ardit h | a s . a. . . tatahprabhrti t h sarv h sahit h saimaulin 135 a. a. s a . . a. kotivarse vasanti sma sarvalok bhayaprad h | a a. . . evam sa bhagav n devo m tnam ganan yakah | a a r. . a . . . . abhavac chankaro vy sa yan m m tvam prs. av n asi 136 a a. . ..t a
66 The interpretation of bhavatn m (yours) in 126d is not certain. 126cd might alternatively be a. understood to mean brimming with blood, she [the river] will always here belong to you. Isaacson,

53 Myself, Brahm , Visnu, and sages rich in penance shall create holy Mother a .. Tantras through which you shall receive the highest worship, the rites of sacrice to the Mothers (m tryajavidhi): the Brahmay mala, Sv yambhuvaa. a a sanay mala, and y mala, Kum ray mala, S rasvatay mala, G ndh ray mala, I a a a a a a a a a Nandiy malathese Tantras of yours, and others too by the thousands, a through which men shall worship you with devotion.67 [12729ab] You shall grant boons to the men worshipping. You shall become goddesses who bestow divine siddhi, possessing divine yoga. Those women who always worship you, secretly, shall become yogevars, lovely women of s divine valour. [129cd31] And my gana-lords, Chagala (the Goat) and . Kumbhakarna (Pitcher-ears), shall by my command remain with you . as door guardians. This excellent place, Kotivarsa, dear to the Mothers, . . shall become the chief (pravara), divine cremation ground, which gives happiness. [13234ab] Having obtained these boons, the Mother goddesses, Mothers of the worlds, prostrated before the Lord of the Gods and rejoiced, extremely excited. From that point on, all of them dwelt in Kotivarsa together with . . Siva, granting freedom from danger to the entire world. Thus did the divine lord Sankara become leader of the hordes of Mothers, O Vy sa, a which is what you had asked me. [134cd36] Skandapur na 171 appears to present a transitional picture in several respects. In a. cultic terms, it juxtaposes a tantric Mother cult with an older, non-Mantram rga cult, a illustrating also how they coexisted with shared cultic centers such as Kotivarsa. . . From the eighth century, the non-tantric Mother cult would wane, while tantric goddess cults ourished.68 On the level of myth, the Skandapur na provides a transitional a. link between the cycles of Skanda and the Mothers in the Mah bh rata, and myths of a a the Mothers inscribed in later texts such as the Devm h tyma of the M rkandeyapur a a a .. a
e-mail communication, Oct. 2006. In either case, the implication appears to be that the river runs red with the blood of sacricial offerings to the Mother goddesses. 67 On the interpretation of this passage, Sanderson remarks, I propose that y malam in 128d is to a . be understood with all (sarvaesatay ), so that the titles indicated are Brahmay mala, Sv yambhuvay mala, s . a a a a s Kum ray mala (=Skanday mala), G ndh ray mala, Ianay mala and Nandiy mala. For of these seven all but a a a a a a a a the Sv yambhuvay mala and the G ndh ray mala are found in the scriptural lists of y malatantras known a a a a a a to me . . . . History through Textual Criticism, 7 (n. 4). I consider this proposal all but certain. In ibid., 7, Sanderson presents the seven lists of y malatantras which had come to his attention in Saiva a literature. 68 As Meister discusses, shrines of the Mothers become increasingly rare, while on the other hand static doorway panels of the Seven Mothers become integral to temple iconography in central India, appearing rst in the eighth century. Regional Variations in M trk Conventions, 24143. However, a. a quite different is the case of Orissa; according to Thomas Donaldson, temples of the Mothers are attested from the seventh century, become common in the tenth century, and continue to be constructed even in the thirteenth. Orissan Images of V r h, Oddiy na M rc, and Related Sow-Faced Goddesses, 170. aa a .. a

54 na.69 As Yokochi points out, the Skandapur na preserves the older, Kusana-era and a. . . . Mah bh rata conception of countless diverse Mother goddesses alongside the Hina a duized Gupta-era Seven Mothers. In the Skandapur na, the latter in fact appear only a. in this account of Kotivarsa.70 This Skandapur na chapter also provides clear links a. . . between the cult of Mothers and cult of Yogins, which otherwise appear lacking in early non-tantric literature. There is moreover some evidence suggestive of a transitional iconic program for the Mother goddess shrine of Kotivarsa. . . Described in verses 11826 is the laukika, i.e. non-tantric cult of the Mothers, the rituals associated with which involve pilgrimage, worship of the images of the Mothers and Siva-Hetuke vara, partaking of the sacred waters, and perhaps animal s sacrice (suggested by the reference to the river brimming with blood). The aims are correspondingly of the variety advanced in Saiva pur nas: deliverance from harmful a. spirits, going to heaven, and joining Sivas entourage of ganas after death. Verses . 12734ab describe, however, a Mantram rga cult of the Mothers. Its rituals are those a taught in the m tr- or y malatantras, and its aim, for men, magical powers or siddhi. a. a For women, the secret rites promise more: the possibility of becoming yogins, pow erful and beautiful Mistresses of Yoga (yogevars). In the tantric cult, the goddess s Bahum ms is the Great Vidy -mantra (mah vidy ). And for adepts of the Mother a. a a a a Tantras, Kotivarsa is not merely an holy ford and place of pilgrimage, but the best . . of cremation grounds, suggesting a possible k p lika orientation to the tantric Mother a a cult. Kotivarsas Mothers appear in the standard group of seven attested from the fth . . century or a little earlier,71 with Bahum ms , apparently the cultic focus, representa. a
69 On the dating and signicance of the Devm h tmya, see the section on post Gupta-era Mother a a shrines below. 70 Yokochi, Rise of the Warrior Goddess, 99113, especially 11011. Although not strong evidence in isolation, this fact might suggest that the seventh century would be too late a date for the Skanda pur na. Even in the sixth century, the paucity of references to the Seven Mothers is surprising in a Saiva a. pur na. a. 71 Viz., Brahm n, Rudr n, Kaum r, Vaisnav, V r h, Indr n, and Bahum ms . Skandapur na a. a. a a. a. a a. .. a a 171.110cd12:

abhut pit mah d br hm sarv n sankar d api 110 a a a a. a kaum r sadmukh c c pi visnor api ca vaisnav | a. . a a .. ..

55 ing C munda, leader of the Mothers (m trn yik ). While C munda is most commonly a a. a a a . . . . depicted as a gaunt, ferocious hag, in this case the name Bahum ms suggests corpua. a lence.72 There are in fact early representations of C munda as a full-bodied woman a . . (cf. figure 2.7), although I am not presently aware of a corpulent example.73 In the present case, Bahum ms might have been a goddess of local importance suba. a sequently incorporated into the Mother cult and identied with C munda.74 The a . . Skandapur na does in fact attest a process by which important local goddesses were a. given Saiva identities through incorporation as Mothers. Chapters sixty-four and sixty-eight enumerate the names and locales assigned to the numerous Mother goddesses who emanate from the goddess Kau ik, linking the Mothers with an emergent s
v r h m dhav d dev m hendr ca purandar t 111 aa a a a a sarvatejomay dev m tnam pravar subh | a r. . a a . bahum ms mah vidy babhuva vrsabhadhvaj t 112 a. a a a a .. As Yokochi points out, this sequence shows clearly that the standard set of Seven m trs is intended, a. for the m trs had already been enumerated out of sequence in 108cd110ab. Rise of the Warrior a. Goddess, 1078. 72 Cf. the goddess Brhodar (i.e. brhadudar, She of the Big Belly) mentioned in BraY i.6062. Not a . . specically identied with C munda, she appears to fulll the dual functions of tutelary goddess of a . . her namesake village, and tantric deity, bestowing the vidy -mantra upon Yajasomas son. Yokochi a suggests that bahum ms is a euphemism and reinforces the goddesss identity as the emaciated C mua. a a nda. Although speculative, this too might be possible. Rise of the Warrior Goddess, 1089. . . 73 Clear cases of early (pre eighth-century) images of C munda as a full-bodied woman include those a . . of the m tr shines of the Siva cave temple at Aihole, that of the Aurangabad Buddhist caves, and the a. two early Ellora cave shrines (R me vara and R van-k K ). a s a . a a In Elloras R van-k K cave m tr shrine, the full-bodied C munda at the right end of the row a . a a a. a . . of seven m trs is clearly distinguished by her owl-v hana. In this shrine, a skeletal divine couple of a. a problematic identity (discussed subsequently) occur to the right of Gane a along the main wall. On the . s left wall is installed a four-armed Vnadhara with a bull v hana. In the R me vara cave at Ellora (cave . a a s 21), an elaborate dancing Siva is instead enshrined on the left wall, while the Vnadhara begins the set . of m trs on the left of the main wall. On the right wall are a skeletal pair of male and female divinities. a. The female gure is unlikely to be C munda, for C munda appears to be among the full-bodied m trs a a a. . . . . along the main wall. In the Aihole Mother shrine of the Siva cave temple, on both the left and right walls stand three m trs, a. with two more anking a central, ten-armed dancing Siva along the main wall. A diminutive Gane a . s stands on the viewers left between Siva and the Mother goddess, with a damaged gure of a diminutive two-armed male in the corresponding position on the rightperhaps the gana Vrabhadra, whom one . might expect to complement Gane a, although Meister suggests Skanda. Regional Conventions, 240. . s Closing the set at the end of the right wall is a full-bodied C munda. a . . Although the BraY and many other early sources describe C munda as emaciated, there is at least a a . . one textual description of a full-bodied C munda: that of the Kashmiri Brhatk lottara, in which Yoge, a a s . . . eighth of the Mothers, is visualized as emaciated and C munda as full-gured. Alexis Sanderson, a . . Religion and the State: Saiva Ofciants in the Territory of the Kings Brahmanical chaplain, 267 (n. 92). 74 Cf. Yokochis discussion of Bahum ms in Rise of the Warrior Goddess, 109. a. a

56 theology of the Great Goddess as source of all goddesses.75 This rhetoric of emanation and localization appears almost as a precursor to the theology of sakti, as adapted with great success from Tantric Saivism, through which any and all goddesses and their sacred sites, declared saktip. has (seats of sakti/power), would be subsumed t within the identity of the one Goddess.76 The iconic type described for the Mothers of Kotivarsa might . . suggest a pre sixth-century date. Siva as Hetuke vara joins Bahus m ms in heading the Mothers, a. a while he appoints his ganas Chagala . (the Goat) and Kumbhakarna . (Pitcher-ears) as door-guardians (dv rap la). This suggests that cult a a images of Siva and the Seven Mothers alone appeared in the central shrine (garbhagrha),77 with the ga. nas Chagala and Kumbhakarna in. . stalled as guardians at the base of the vertical panels on either side of the entry door. Saiva Tantras refer
Figure 2.7: C munda and Vin yaka/Gane a, R me vara a a a s . . . s Cave (no. 21) m tr shrine, Ellora. Sixth century. a.

to the installation of such gana-lords . as door guardians. For instance, in

the Svacchandatantras description of ritual entry into the y gagrha (worship shrine), a . a. one rst offers homage to the directional Mothers (dinm trs) outside the shrine, then
Skandapur na 64 narrates the emergence of manifold goddesses from Kau iks limbs. In Skandaa. s pur na 68, Kau ik assigns the various Mother goddesses which emerged from her to locales, including a. s Bahum ms to Kotivarsa. See Yokochis discussion in ibid., 99100, 11112. a. a . . 76 See below in the brief discussion of the Devm h tmya. a a 77 Although not impossible, it seems highly unlikely that the text would mention dv rap las yet omit a a reference to other cult images in the garbhagrha. .

57 to Gane a and Sr[laksm] on the overhead door lintel; one then worships Sivas ga. s . na-lord Nandin on the left of the outer door frame, together with the river goddess . a Gang , and on the right, Mah k la with Yamun .78 Netratantra 3 corroborates this a a a procedure, its commentary mentioning also the ganas Mesanana (Ram Face) and . . Ch g nana (Goat Face) as door guardians.79 a a What appears archaic about this iconic program is the placement and identity of the male deities. In Kusana-era m tr sculpture, the Mothers were often depicted with a. . . images of the young, spear-bearing Skanda (figure 2.2, 2.4) or the robust yaksa-lord . Kubera (figure 2.3), usually positioned to their side.80 However, from the Gupta era Siva himself often features as leader of the Mothers, particularly in the form of Na te a, Lord of Dancers.81 As head of the Brahmanical heptad of Mothers, Siva thus . s

Svacchandatantra 2.2225: a sakalkrtadehas tu puspam ad ya suvrate | . . a. dinm trbhyo namaskrtya dv ram samproksya yatnatah 22 a . . . . . siv mbhas stramantrena vighnaprocc . anam bhavet | a a at . . dv raakhordhvato devam ganeam ca sriyam tath 23 a s a . . s . . a sampujya gandhapusp dyair dhup dibhir anukram d | a . . a arghyap dyopah rai ca tato dv rasya cottare 24 a a s a nandigange samabhyarcya mah k lam ca daksine | a a . . . k lindm caiva sampujya yath nukramayogatah 22 a . a . O pious woman, with ones body transformed into [the mantra-body of] manifest (sakala) Siva, one should take up the ower, bow to the directional Mother goddesses, and sprinkle the door, carefully. Through the siva-water empowered by the weapon-mantra, obstruct ing forces would be driven away. After worshipping Lord Gane a and Sr[laksm] [on the . s . lintel] above the double-doors with scents, owers, incense, etc., in sequence, to the left of a the door one should worship Nandin and Gang with the guest-water, water for washing the feet, and the offerings, and also Mah k la and Yamun on the right, in the correct a a a sequence.


Netratantra 3.9: a r asam tr ganam laksmm nandigange ca pujayet | . . . . . mah k lam tu yamun m dehalm pujayet tatah 9 a a . a. . .

According to Ksemar jas comments on this verse, in the system of the v ma-stream of revelation, i.e. a a . the cult of the Four Sisters of Tumburu expounded in the archaic v matantras, Mesanana (Ram Face) a . and Ch g nana (Goat Face) serve as additional (adhika) dv rap las in shrine contexts, a fact perhaps a a a a relevant given the Skandapur nas reference to The Goat as a door guardian at Kotivarsa. He remarks, a. . . a r. bahir dinm th, dv rordhve ganapatilaksmyau, p rvadvaye nandigange mah k layamune, v me dehalm praa as a a a . . . . navacaturthnamahsabdayogena pujayet | asya nayasya sarvasahatv t siddh ntadrsa nandigange daksine pujye, a a . . . . . mah k layamune v me | v masrotasy evam mesasyacch g syau tu adhikau daksinav mayoh | bhairavasrotasi a a a a a a . . . . a . samh rapradh natv d daksine mah k layamune v me nandigange | sadardhe tu dindimahodarau adhikau 9 a a a a a . a . . . . .. Cf. Ksemar jas remarks on 18.56cd. a . 80 Joshi, M trk s: Mothers in Kusana Art, 10328. a. a .. 81 See the tables Meister provides for the iconic programs of Gupta and early post-Gupta m tr sets, a. as well as later sets from central and western India. Regional Variations, charts a and b.

58 replaces Skanda and Kubera. Yet in only one extant Mother shrine does Siva alone appear in the Mothers company: that of the fth-century cliff shrine between Badoh and Pathari, in present-day Madhya Pradesh. In this case, an image which appears to be Siva anks the Mothersa two-armed, urdhvaretas yogin complemented by C munda at the other end (figure 2.8). Kotivarsas iconic program might hence a . . . . have resembled that of Badoh-Pathari, the cult images being those of Siva, Bahum ms , and six other m trs, with Chagala and Kumbhakarna placed outside the a. a a. . central shrine as door-guardians. With the exception of the early Badoh-Pathari shrine, the Seven Mothers are nor mally depicted in the company of multiple gana-lords, or with Siva and a gana-lord. . . Most common as anking gures are the elephant-headed Vin yaka or Gane a, and a . s one of two anthropomorphic male gures usually identied as Vrabhadra (Aus picious Hero) and Vnadhara (Bearer of the Vna). Problematically, both of the . . latter are often described as forms of Siva.82 Vrabhadra, however, is in Saiva tex tual sources of the period considered a gana or gana-lord (ganea, ganevara, etc.) of . . . s . s Siva, or else a prominent rudra.83 Hence, one common saptam tr iconic program pairs a. the gana-lord Vrabhadra with another gana-lord, Vin yaka, frequently anking the a . . Mothers on either side (figure 2.9). While Vrabhadra was undoubtedly included
For example, Meister remarks, Siva sits as Vrabhadra or Vnadhara at the head of the M trk set . a. a . . . , or dances as Nate a in their midst. Regional Variations, 241. s . 83 In some early sources, Vrabhadra is a prominent rudra; note e.g. Sadyojyotis, in the Moksak rik , . a a referring to the hundred rudras as headed by Vrabhadra ([maheano nugrhnati] icchay satarudr ms s a a. . . ca vrabhadrapurahsar n, 79b). In cosmology, Vrabhadra presides as a bhuvanevara over the high s . a est of the ascending series of bhuvanas (planes or worlds). Cf., e.g., Abhinavaguptas Tantras ra, a 8.2: yat tu katipayakatipayabhed nugatam rupam tat tattvam yath prthiv n ma dyutik . hinyasthauly dirup a a . a at a a . . . k l gniprabhrtivrabhadr ntabhuvaneadhis. hitasamastabrahm ndanugat . The Matangap ramevara, which aa a s . t a. . a a s . refers to Vrabhadra as one of the lokan yakas, world lords (vrabhadr daya caiva brahm nt lokan yak h, a a s a a a a. 23.26cd), also describes the ganas as led by Vrabhadra (vrabhadrapurahsara, 23.47b). His role as a ga . . na-lord is prominent in the early Skandapur na. And in the BraY and other sources, Vrabhadra is a. a . credited with revealing the Vrabhadratantra. BraY xxxviii.61cd62: a

svaro vijayov ca dev niv sam eva ca 61 a s a brahm sv yambhuva caiva vrabhadras tathaiva ca | a a s vrabhadram mah devi prov ca vidhivistaram 62 a a .
ni v sam ] em.; visv sam Bya s a a

The association between Vrabhadra and the Seven Mothers is attested in a number of tantric Saiva sources, such as the Vratantra, the seventy-third chapter of which describes the installation (sth pana) a of the Seven Mothers, Vrabhadra, and Gane a. . s


Figure 2.8: Siva and the Seven Mother goddesses (detail). Saptam tr cliff shrine between Badoh and a. Pathari, Madhya Pradesh. AIIS Photo Archive.

Figure 2.9: Saptam tr Panel, with Vrabhadra a. (left) and Gane a (right). State Museum, Asha. s puri, Madhya Pradesh. AIIS Photo Archive.

60 among the Mothers as a gana-lord, originally, I am less condent about the early . identity of the Vnadhara image type; no gana by this description is mentioned in . . connection to m trs in period textual sources known to me. This probably points a. towards him being a form of Siva, who is indeed associated with the vna.84 In any . case, a shift in the perception of both, from gana-lords to forms of Siva, might at some . point have transpired, much as Gane a undergoes a shift in identity from gana-lord . s . to child of Siva and P rvat. a It appears that the Mother shrine at Kotivarsa predates the association of the . . Mothers with Vrabhadra or the Vnadhara, and the elephantine Vin yaka, all unmis . a takably Saiva sacred gures whose images accompany the goddesses in post fthcentury shrines. In Kotivarsas iconic program, the ganas Chagala and Kumbhakarna . . . . appear instead, stationed outside the shrine proper, with Siva himself joining the Mothers in the sanctum. This pantheon suggests a transitional picture, for although described in the Skandapur na as ganas of Siva, Chagala and Kumbhakarna have clear a. . . ties with the older cult of Skanda. Perhaps not uncoincidentally, they moreover resemble Skanda and Kubera, respectively, the primary deities depicted in connection with the Mother goddesses in early, Kusana-era statuary. . . Chagala, the Goat, is in all probability a name of Naigamesa, a goat-headed . deity prominent in the mythology and cult of Skanda, with whom Skanda is sometimes even identied.85 Such theriomorphic gures are attested in this period as
There might however be some indication in early sculpture of a status similar to Vrabhadras. Note that in the Mother shrine of the Ellora R me vara cave, the vna-bearer anks, on the left, the row a s . of m trs on the main shrine wall; on the left wall is installed a grand Siva image, Nate a. This could a. . s suggest that the vna-bearer was a gana-lord, for Siva-forms would otherwise appear twice, moreover . . in adjacent images. It is conceivable that the vna-bearer was a gana-lord of a type similar to Nandin, . . described in the Skandapur na as a mini-Siva who in appearance mirrors his divine lord. a. 85 In the Mah bh rata, Naigamesa is also called Ch gavaktra, Goat Face;cf. Mah bh rata, Araa a a a a . nyakaparvan 215.23, where Agni is said to become Naigameya or Ch gavaktra, Skandas companion. a . . In Aranyakaparvan 217, mention is made of Ch gavaktra, and then, connecting the two, one of Skandas a six faces is said to be that of a goat, this being his primary face, dear to the Mothers: esa vr s. akah proktah skandam trganodbhavah | a. . . a. t . . . ch gavaktrena sahito navakah parikrtyate 11 a . . sas. ham ch gamayam vaktram skandasyaiveti viddhi tat | . .t . a . . satsiro bhyantaram r jan nityam m trganarcitam 12 . . . a . a . . . sannam tu pravaram tasya srsanam iha sabdyate | . . . .. . a In another passage, Salyaparvan 43.3740ab, Skanda makes himself fourfold, becoming Skanda, S kha,

61 door-guardians; cf. figure 2.11. Kumbhakarna, apparently a yaksa-type, is less read. . ily identied. A r ksasa by this name features in the R m yana as R vanas brother, a a . a a . a . deadly warrior and voracious eater and sleeper. I suspect that in the Skandapur na, a. Kumbhakarna (Pitcher-ears) is an alias of Ghantakarna, Bell Ears, the name of a . . . . gana of Siva mentioned in the Skandapur na and in other sources.86 It is also possible a. . that in Kotivarsa, an assimilation has been made between a local yaksa and a gana . . . . of Siva. As a set, the Kotivarsa Mothers and their two gana-guardians appear to re. . . . place the group of Nine Heroes mentioned in the Mah bh rata, Aranyakaparvan 217: a a the seven sium trs (Mothers of the Infant[s]), their ferocious son, perhaps named s a. Lohit ksa,87 and Ch gavaktra. In the Skandapur na account of Kotivarsa, Chagala or a . a a. . . Ch gavaktra is carried over from the older sium tr set, while Sivas gana Kumbhakara s a. . na replaces Lohit ksa, both apparently erce yaksa-types. Pitcher- or bell-shaped ears a . . . suggest the massive earrings commonly depicted on yaksas; cf. figure 2.10, and com. pare also the bust of a fanged, large-eared yaksa (5th6th century) found at Pawaya, . M.P., near to Gwalior (figure 2.12). Continuity is thus apparent between the Mah a bh ratas Nine Heroes and the Kotivarsa Mothers and their guardians, Chagala and a . . Kumbhakarna, the new set of nine presided over by Siva instead of Skanda. .
Vi akha, and Naigamesa. s . 86 Cf., e.g., Skandapur na 164.61, quoted in the next note. Sats hasrasamhit 4.136a lists Ghantakarna a. . . . . . a . a in a series of p lakas, guardians, a list which includes gana-lords such as Gaj nana (136b). a a . 87 The name of the son of the sium trs is not unambiguous: he is described as endowed with s a. vitality, very terrible, born by the grace of Skanda, having red eyes, terrifying, and a little . child: Aranyaparvan 3.1011: et s m vryasampannah siur n m tid runah | a a. a a a . . . . s skandapras dajah putro lohit kso bhayamkarah 10 a a . . . . esa vr s. akah proktah skandam trganodbhavah | a. . . a. t . . . ch gavaktrena sahito navakah parikrtyate 11 a . . Here n ma probably has an emphatic sense, for it seems probable that the proper name of the gana is a . Lohit ksa. Salyaparvan 45.22 lists the latter among four attendants (anucara) of Skanda: a . nandisenam lohit ksam ghan. akarnam ca sammatam | a . . . . . . t . . . caturtham asy nucaram khy tam kumudam linam 22 a a . a . Cf. Skandapur na 164.61, in a chapter closely parallel to and drawing upon Salyaparvan 45: a. ghan. akarnam surakt ksam nandisenam ca durjayam | a . . . t . . . . . caturtham balin m sres. ham khy tam kumudam linam 62 a . .t . a . a . On this passages dependence upon the Mah bh rata, see Yokochi, Rise of the Warrior Goddess, a a 100. Note in these lists the presence of Ghantakarna, whom I suggested above could be identical to . . . Kumbhakarna at Kotivarsa. . . .


Figure 2.10: Yaksa at pillar base, Nagarjuna. konda. National Museum, New Delhi. AIIS Photo Archive.

Figure 2.11: Theriomorphic gana or dv rap la. Dea a . orani temple (circa 6th century); Tala, near Bilaspur, M.P. AIIS Photo Archive.

At Kotivarsa, the emphasis on Chagala/Naigamesa and Kumbhakarna instead of . . . . Gane a and Vrabhadra thus appears archaic, and could suggest transitional iconog . s raphy of the fth century or somewhat earlier. The possibility of unknown regional variations must however be taken into account, especially given that Kotivarsa lies . . in a peripheral zone of Brahmanical culture. Moreover, what appears in the textual description as archaic could simply be a continuity of narrative, the effort to link contemporary cult and revered myth. That is, the Skandapur na might not represent a. the iconic program at Kotivarsa accurately. The possibility is nonetheless signi. . cant that the site possessed a comparatively ancient shrine of the Mothers. While the Skandapur na, composed somewhat later, cannot be taken as a reliable indicator of a. the earliest cultic practices at Kotivarsa, it does at least indicate that in the sixth or . .

63 early seventh century, if not earlier, a Yogin cult had developed possessing close ties to the cult of Mother goddesses at important Saiva temples and pilgrimage spots. The Skandapur nas description of a. the deities of Kotivarsa nds some . . corroboration in tantric literature. Kotivarsa features in the BraY s sacred a . . geography as one of eight cremation grounds (maana);88 and in the elabs s orate initiation mandala described in .. BraY iii, these are represented in a
Figure 2.12: Head of a fanged, large-eared yaksa. . Pawaya, Gwalior, M.P. 5th6th century. AIIS Photo Archive.

the eight directions around the man. dala perimeter, with Kotivarsa in the . . .

northeast. Its depiction includes the Pool of the Spear, but not Bahum ms and a. a Mother goddesses. Hetuke vara however presides in a circle of eight rudras, beyond s whom lie six yogins headed by Hetuk. In the outer deity circuit are four r ksasas, a . and three of Sivas ganas: Nandin, Chagala, and Kumbhakarna.89 The Tantrasadbh va, a . . a text probably postdating the BraY ,90 also mentions Kumbhakarna as the ksetrap la a a . . (sacred eld guardian) of Pundravardhana, the very region of Bengal where Koti.. . . varsa is situated.91 At Kotivarsa itself (kotakhye) preside the goddess Karnamot and . . . . . . the eld-guardian Hetuka.92 Although not conrming the importance of Mother

BraY lxxxiv.81: a pray g varuna koll at. ah s jayantik | a a a .t a a a . caritraik mraka caiva kotivarsam tath s. amam a a. t . . . .
attah s ] em.; hattah s .. a a .. a a Bya


Sanderson discusses the probable locations of these sites in History through Textual Criticism, 7 (n. 4). Tantrasadbh va 15.21 is identical to this verse, offering as substantive variants at. ah s (adopted above) a .t a a and the corrupt caritrek mbukam (81c; mss as reported by Dyczkowski). a . 89 BraY iii.12027. a 90 See chapter 4, section 3. 91 Tantrasadbh va 19.5758. On the location of Kotivarsa, see Sanderson, History through Textual a . . Criticism, 7, and Yokochi, Rise of the Warrior Goddess, 1067. Yokochi reports that Kotivarsa is . . referred to as a Visaya of the Pundravardhana Bhukti in a grant of Mahp la i, issued on his ninth regnal a .. . year, which was discovered among some ruins called Bangarh in the West Din jpur district, Bengal. a Note however that Tantr loka 15.8788 lists Devkotta and Pundravardhana as separate p. has. a .. t .. 92 Tantrasadbh va 19.3233: a

64 goddesses at Kotivarsa, these scriptures of the Yogin cult do associate the site with . . tantric practice, linking it moreover to some of the same deities as Skandapur na 171. a. Elsewhere in the Skandapur na, evidence for the Yogin cult appears meager. a. Chapter 155 mentions a class of female beings called yogevars, a synonym of yogin, s in the context of the never-ending battle of sura and asura. From drops of the blood of Andhaka, impaled by Siva, sprang warriors who rout Sivas ganas. Siva then, it . seems, creates the Mistresses of Yoga (yogevars),93 and orders them to slay and eat s the great asuras who were born from drops of blood. They oblige. Noteworthy here is that as a class of beings, the yogevars take on the martial role which had been s the domain of Skandas demon-slaying m trs and other ganas in the Mah bh rata, a. a a . illustrating through myth an important historical continuity.94 Note also chapter sev enty, which contains the m h tyma-narrative of the sacred mountain Srparvata or a a Srgiri, a site early tantric Saiva texts mention as an important sacred site (p. ha).95 t Its close association with the Yogin cult is attested in an early eighth-century drama, the M latm dhava of Bhavabhuti, as discussed subsequently. However, in the Skandaa a pur na account, the description of religious practices at Srparvata appears largely to a. reect P supata Saivism, with liberation the goal and yoga the means; mentioned a also are pilgrimage practices of the Saiva laity.96 This accords in general with the
karnamot tu koty khye mah balakulodbhav | a a . . . a sulahast sthit dev sarvayogevarevar 32 a a s s tasmin ksetre sthit dev vatavrksasam sritah | a . . . a . . ksetrap lo mah k yo hetuko n ma n matah 33 a a a a a . .
32b kulodbhav ] G; kulodbhav h K Kh a a. 32d svare var ] K Kh; svarai var G s s

mss as reported by Dyczkowski. 93 In this section, 10a13b, there is loss of about four verse quarters in the editio princeps, including the object of [mah devo] srjat (Siva created . . . ). The beings he creates are however identied as yogevars a s . in 15b. 94 . The bellicosity of the early Mothers is illustrated for example in Aranyakaparvan 215 (1622). 95 Cf., e.g, Svacchandatantra 9.37a. 96 Note Skandapur na 167, which describes Srparvata as a siddhiksetra (a sacred eld for attainment) a. . with hundreds of sivalingams, where Brahmins devoted to the practice of yoga attain spiritual success (siddhi). This Saiva, Brahmin sect of yogins is surely the P supatas. Skandapur na 167.10307: a a. tatra srparvato n ma parvatah srniketanah | a . . siddh maraat krnam siddhiksetram tad uttamam 103 a s a . . . . a a. sulino yatra ling n m puny n m varad yin m | a a . a a. sahasram sth pitam vy sa sil dena mah tman 104 a a a a a . . ye payanti tam sanam srparvataniv sinam | s . a

65 sectarian orientation of the Skandapur na. Yet the text might make allusion to the a. Mantram rga as well, telling the tale of how P rvat became a great perfected yoa a gin (siddh mah yogevar) by doing penance on Srparvata.97 This terminology is a a s suggestive of the Yogin cult, particularly in association with Srparvata. Thus while the Saiva Yogin cult does not appear prominent in the Skandapur na, a. chapter 171 provides unambiguous evidence for its existence in the eighth century, fairly secure evidence for it in the seventh century, and a signicant possibility for the sixth, depending upon the dating of the Skandapur na itself. Vague reference a.
janmamrtyubhayam tesam n sti pums m kad cana 105 a . . . . a . a. anenaiva sarrena tasmin ksetre bhav tmake | . a . yog bhy sapar vipr h siddhim y nti yathepsit m 106 a a a a. a . a manas py abhigacchanti ye nar h srgirim mune | a a. . na te y nti yam v sam mrty v api sam gate 107 a a a . . a a After Sanatkum ra tells the tale of the asura Hiranyaka ipus penance and defeat of the gods, a s . Vy sa asks him about the mountain, Srparvata, where he practiced his austerities. The Dev, while a sporting with Siva on the mountain, had noticed powerful yogins there circumambulating [an image or lingam of] Siva, perfected through austerities and possessed of great yoga. She questions the lord about the yogins, and he describes them as great ones perfected in yoga, abiding in the P supata a yoga (70.48ab). The Dev too desires this attainment, and performs penance there; as boon, Siva declares that she will become the Great Perfected Mistress of Yoga (siddh mah yogevar), knowing the a a s highest doctrine or meaning (par rtha) of the entire adhy tmatantra [spiritual treatise?] taught by Siva. a a The mountain upon which she performed penance will be given the name Srparvata. Skandapur na a. 70.4259, Bhattar edition: . . a hiranyakaipur daityo yatra tepe mahat tapah 43 s . . girau tasmin mah punye devadevo vrsadhvajah | a . .. . umay sahito devy reme nityam mah dyutih 44 a a a . . tis. hatas tatra devasya dev girivar tmaj | a a .t tapahsiddh n mah yog jvalan dityavarcasah 45 a a a a . . krtv pradaksinam sambhoh sapranamam mah mune | a . a . .. . . . gacchatas t n muhur drs. v papraccha bhuvanevaram 46 a s ..t a ka ete deva samsiddh yogsas tv m pranemire | a a. . . tato devah prahasyain m uv ca paramevarah 47 a a s . . yogasiddh mah tm no yoge p supate sthit h | a a a a a. ya ih r dhya m m devi jahuh pr nan narottam h 48 aa a. a. . a. ta ete siddhadeh rth h svacchandagatic rinah | a a. a . . moksasiddheh par m nis. h m gant rah paramam padam 49 a . .t a . a . . . . atha samcintya suciram vismay yatalocan | a a . . devam prov ca sarv n vacah paramapujitam 50 a a. . . yath ham api devea pr pnuy m siddhim drsm | a s a a. . a karisy mi tath yatnam esa c smi gat vibho 51 a a a . . . tatah krtv sriyo rupam dev paramaobhan | s a . . a mahat tatra tapas tepe sahasram parivatsar n 52 a . deva c sy varam pr d t taih sam nam mah dyutih | s a a a . a . a a . . krtsnasy dhy tmatantrasya may proktasya bh mini 53 a a a a . mah yogevar siddh par rthaj bhavisyasi | a s a a a .

66 to Tantras by the thousands suggests that even at this early date there may have existed a large and perhaps developed corpus of Mantram rga literature. This textual a situation nds mutual corroboration in the BraY , one of the scriptures the Skandaa pur na lists, which as I shall discuss in chapter three describes a sizeable body of a. Saiva scripture in its account of the canon. It is signicant, moreover, that the Skandapur na places the Yogin cult in close association with the cult and shrines of Mother a. goddesses, for in this period, monumental shrines of the Mothers such as may have existed at Kotivarsa were constructed in considerable numbers. It is to more of this . . evidence that we turn next.
post-gupta era temples of the mothers

A large number of Mother-goddess shrines survive from the sixth and seventh centuries,98 and the circa mid sixth-century Brhatsamhit of Var hamihira speaks of the a . . a temple cult of Mothers alongside major sectarian denominations of the period, including Buddhism, Jainism, and the Vaisnava Bh gavatas.99 Nonetheless, the surviva .. ing Mother shrines have clear Saiva orientations, occuring primarily in association with Siva temple complexes. While these hence attest the existence of a widespread
See Katherine Harper, Iconography of the Saptamatrikas, 10149. Brhatsamhit 59.19: . . a visnor bh gavat n mag ms ca savituh sambhoh sabhasmadvij n a a a. a .. . . m tnam api mandalakramavido vipr n vidur brahmanah | a r. a . .. . . saky n sarvahitasya santamanaso nagn n jin n m vidur a a a a. ye yam devam up srit h svavidhin tais tasya k ry kriy 19 a a . a a a a . Edition of A. V. Tripathi. The Saiva brahmins with ashes (sabhasmadvij n) are in all likelihood a P supatas. While this passage does not specify the Mothers identities, elsewhere the text speaks of cona structing the images of Mothers in accordance with the appearances of the deities they are named after (m trganah kartavyah svan madev nurupakrtacihnah, 57.56ab). It therefore seems likely that Var hamihira a. . . a a a . . . knows of and refers to the Brahmanical Seven Mothers. His terminology for describing specialists in the cult of Mothers, knowers of the mandala-sequence (mandalakramavidah) or knowers of the .. .. . Mother-mandala (m trmandalavidah, edition of H. Kern), has been taken by Harper as an indication of a. .. .. . a tantric cultic orientation. Iconography of the Saptamatrikas, 122. However, the mere occurrence of the term mandala does not warrant this; here it probably means the group/set [of Mothers]. Note the .. same terminology in the inscription of the contemporaneous saptam tr shrine at Deogarh, which in its a. benedictory verse refers to the enshrined deities as the mandala of Mothers: .. sth n m jagadraksaksamaujas m a a. a. . . m tnam lokam tm(t)nam mandalam bhutaye []stu vah 1 a r. . a r. r .. . . . .. . .
99 98

Epigraphia Indica xxx.15 (pp. 12527).

67 Saiva temple cult of the Mothers, they bear an unclear relationship to developments in tantric forms of Saivism. That the two were in some contexts linked is suggested by the Skandapur na account of Kotivarsa. In the present section are discussed two a. . . facets of Mother shrines that could provide insight into this connection: cases of an eighth goddess in some sculpted sets of the Mothers, and the presence of k p lika a a deities and iconography. Tantric Saiva sources of the Yogin cult sometimes speak of Mother goddesses as an octad, rather than the earlier and more widely attested Brahmanical heptad, even when afrming their identity as the Seven Mothers.100 In tantric literature, the eighth, additional Mother is sometimes Mah laksm, or else the supreme Goddess of a . the bhairavatantras herself, often by the name Yoge (=Yogin).101 Hence in the BraY , s a Bhairav or Aghore is called the one who completes the Mothers (m trpuran).102 s a . . Pur nic accounts evidence more variety; the Devm h tmya of the M rkandeyapur na, a. a a a .. a. for instancea text which marks the entry of tantric Saiva notions of sakti into the pur nic theology of goddesses103 adds N rasimh to the Mothers, the embodiment a. a .
Thus chapter 1 of the Svacchandatantra, in the published Kashmiri recension, distributes the phonemes of the alphabet in vargas connected with eight Mothersthe usual heptad plus Mah laksmafter enumerating whom the text states, these are the seven Great Mothers, situated a . in the seven worlds (et h sapta mah m trh saptalokavyavasthit h, 36cd). Sanderson points out that this a. a a .. a. reference to the Mothers is absent from the recension of the Svacchandatantra preserved in Nepalese manuscripts (personal communcation, January 2007). 101 For Yoge as the name of the eighth Mother, cf., e.g., BraY xxxxv.32, Tantrasadbh va 14.155b, s a a Tantr loka 29.52d, and the Brhatk lottara, mentioned previously (n. 73). a a . 102 E.g. BraY ii.18b. a 103 The Devm h tmya is frequently spoken of as a text foundational to the formation of Hindu goddess a a traditions, providing, for instance, an early example of the textual depiction of the Mothers. Note e.g. Harper, who assumes a circa 400600 c.e dating of the text. Iconography of the Saptamatrikas, 91, citing R. C. Hazra, Studies in the Pur?n?ic Records on Hindu Rites and Customs, 1012. Cf. Thomas Coburn, who accepts a fth- or sixth-century dating of the Devm h tmya, citing D. R. Bhandarkar, Epigraphic a a Notes and Questions, Journal of the Bombay Branch of the Royal Asiatic Society xxiii (1909): 7374; he also refers to V. V. Mirashi, A Lower Limit for the Date of the Dev-M h tmya, Pur na vi 1 (Jan. 1964): a a a. 18184. Coburn, Dev-M h tmya: The Crystallization of the Goddess Tradition, 1, 63 (n. 204). However, the a a Devm h tmya is probably not as old as has been supposed: Yuko Yokochi questions on strong grounds a a the minimal evidence adduced in support of a sixth-century (or earlier) dating. She instead proposes the second half of the eighth century, primarily on the basis of the iconic type described for the Warrior Goddess. Rise of the Warrior Goddess, 2123 (n. 42). Yokochi notes that the Devm h tmya is the rst source of its type to draw upon the tantric Saiva a a conception of sakti in relation to the nature of goddesses more generally. She points out that Skanda pur na 171 does not describe the Mother-goddesses emitted by male gods as their saktis, in contrast to a. the Devm h tmya. Ibid., 15 (n. 31.). Her revised dating of the Devm h tmya ts well with the emerging a a a a picture of the chronology of tantric Saiva literature, for the tantric Saiva goddess cults that might have

68 of the feminine power (akti) of Visnus Man-lion avat ra. Sculpted sets of Mother s a .. goddesses do not adopt equivalent iconographic schemes, although several add an eighth goddess. The mid sixth-century shrine in the Saiva cave temple at Aihole, mentioned earlier (n. 73), depicts eight Mothers and might include Mah laksm in a . place of Vaisnav, adding P rvat to the heptad as well.104 Meister mentions two a .. other cases of an eighth goddess among the Mothers: the late sixth-century Elephanta cave east of the main shrine, and the eighth-century Vait l Deul temple of a Bhuvane var, Orissa.105 In both cases, the iconographically ambiguous eighth gods dess appears subordinate to C munda, rather than transcending her in the manner a . . of Aghore/Bhairav in the Saiva Yogin cult.106 The eighth goddess at Vait l Deul s a bears a trident in one of her two arms, suggesting an association with Siva, but has no v han marking her identity; Thomas Donaldson suggests Mah laksm, but on a a a . unclear grounds.107 Another case of an eighth goddess joining the heptad of Mothers is the m tr a. shrine of the little-studied Aurangabad Buddhist cave complex, perhaps of the sixth century.108 This small cave temple presages textual evidence of the seventh century
inspired the theology of sakti alluded to in the Devm h tmya seem likely to have been widespread by a a the eighth century. 104 The central wall of this shrine features two goddesses framing a dancing Siva, with diminutive images of Gane a and either Vrabhadra or Skanda standing below them. Meister identies the goddesses . s . as P rvat and Sr (i.e. Srlaksm or Mah laksm). Regional Variations, 240. The latter deity holds a a . a lotus, and could instead be (an unusual) Vaisnav, while the image identied as P rvat possesses a .. a crescent moon upon her headdress, like M he var on the wall to the viewers left. This set could a s thus include both Mah laksm and the supreme Saiva goddess, the two most common additions to the a . heptad in tantric accounts of the Mother goddessesor else it simply adds P rvat to the usual heptad. a 105 Ibid., 23738. 106 While I follow Meister in identifying the emaciated image adjacent to Gane a, on the left, as . s C munda, Harper identies the image as K la, a male deity, in which case the standard heptad of a a . . Mother-goddesses is present. Iconography of the Saptamatrikas, 113. This would to some extent bear comparison with the R van-k k shrine at Ellora; there, however Mah k la is depicted to the right of a . a a a a Gane a, rather than within the row of Mothers, and is accompanied by a smaller, emaciated goddess . s presumably K l. a 107 a Donaldson, Tantra and S kta Art of Orissa, vol. 1, 109 (figure 211). Perhaps Yoge/Bhairav? s 108 This is the so-called Brahmanical cave, located near to cave six. See Carmel Berkson, The Caves at Aurangabad. Early Buddhist Tantric Art in India, 217, 22528. Katherine Harper suggests a close stylistic relationship between the Aurangabad Mothers and the m tr sets of Elephanta and Ellora, R me vara a. a s cave, hence situating the shrine in the sixth century. Harper, Iconography of the Saptamatrikas, 114. My description of the temple is based upon personal observation, as well as the images published by Berkson, ibid., passim.

69 pointing toward limited incorporation of the Mother-goddesses in Tantric Buddhism, discussed in chapter 4. Along its left wall are situated a row of six standing Mothers, anked near the entrance by a four-armed male guardian gure, presumably Vrabhadra. On the central wall, opposite the entrance, is situated an imposing seated Gane a, anked on either side by goddesses: the seventh Mother, C munda, to the a . s . . viewers left, and a four-armed goddess on the right who bears a triula, sword, bell, s and pitcher, possibly with a lion as v han Durg , it has been suggested.109 Along a a a with the case of Aihole, this points toward a convention of representing the spouse goddess of Siva as eighth of the Mothers, in this case in the guise of the Warrior Goddess facet that rose to great prominence in the period, eulogized in the early seventh-century Candsataka of B na.110 On the right wall are present two nondea. . . script seated Buddhas (one badly damaged) with attendent gures, displacing Siva and providing a Buddhist identity to the Mothers. K p lika deities and iconography gure in representations of the Mothers from a a as early as the mid-sixth century. This development appears rst, it seems, in the Mother shrines of the R me vara and R van-k k cave temples of Ellora, of the a s a . a a mid- or late-sixth centuries; these depart from earlier models by the addition of a skeletal divine couple adjacent to the Mothers (figure 2.13). The identity of these deities is problematic: art-historians have suggested K la (i.e. Mah k la) and K l, a a a a which is plausible.111 As discussed subsequently, early seventh-century literature evidences the existence of a tantric cult of Mah k la, whose association with the Mother a a goddesses is described in chapter fty-four of the BraY . Inclusion of Mah k la in a a a the iconographic programs of Elloras sixth-century Mother shrines hence suggests a degree of congruence with period texts. The identity of the female deity seems less certain; B na speaks of Mah k la as the consort of Candik , a deity whose identity a. a a .. a
Berkson attributes this identication to Ramesh Gupte. Caves at Aurangabad, 227. It should be pointed out, however, that this eighth goddess is not contiguous with the other seven, for the image of Gane a intervenes; she might therefore not have been considered one of the Mothers, . s per se. 111 Cf. Harper, Iconography of the Saptamatrikas, 11314, 11617.
110 109

70 may subsume any number of erce Saiva goddesses, especially Durg and C mua a nda.112 As consort of Mah k la, this erce goddess could be viewed as an eighth a a . . Mother, completing a pantheon akin to that described in BraY livthe Seven Motha ers headed by Mah k la (identied with Bhairava) and Bhairav or Yoge var. a a s It does not seem possible to situate with precision the Mothers, as represented in sixth-century sculpture, with the goddesses as described in extant Saiva yogin literature; the latter might indeed belong to a somewhat later period. However, the evidence for a convention of including an eighth goddess and the presence of Mah k la in particular suggest possible correlations with developments in Tantric a a Saivism. On the other hand, it seems highly likely that the k p lika iconic program a a of the Vait l Deul temple of late eighth-century Orissa is informed by contemporaa neous tantric pantheons and iconography, a period and region in which tantric Saiva goddess cults were undoubtedly prominent.113 This extraordinary temple enshrines a cult image of C munda in a program that a . . includes seven other Mother goddesses and seven male deities. Here we nd the Mothers in a cultic context that is clearly saktaC munda, rather than Siva, presides a . . as supreme deityand manifestly k p lika in iconography. Besides Vrabhadra and a a Gane a, who normally accompany the Mothers, the male deities include Kubera, . s Var ha (who holds a skull-bowl), a seated yogin with a canopy of hooded serpants a identied as N gar ja (King of Serpants), and two Saiva, k p lika deities: a skelea a a a tal (Atirikt nga) Bhairava holding a skull-cup and knife, astride a corpse; and a a skeletal, ithypallic deity seated in yog sana on a corpse, whom Donaldson identies a as Gajasamh ramurti: Siva as Slayer of the Elephant demon (figure 2.14).114 An . a Orissan-provenance text refers to the presiding C munda of the temple as K p lin a a a . . (Skull-bearer), suggestive of the principle epithet of Bhairav in the BraY : Canda a . .
112 On B na, see the subsequent section. In the BraY , the name Candik occurs as a synonym of a. a .. a C munda; cf. BraY ii.16d. a a . . 113 While the outer structure contains an inscription of the thirteenth or fourteenth century, the core shrine has a short inscription apparently dating to the late eighth century, on paleographic grounds. Krishna Chandra Panigrahi, Archaeological Remains at Bhuvaneswar, 3233. 114 Donaldson, vol. I, 10812.


Figure 2.13: Mah k la and K l? Ellora, R me vara cave temple, west wall of Mother-goddess shrine. a a a a s AIIS Photo Archive.

K p lin, Grim Bearer of the Skull.115 Signicantly, the temples decorative reliefs a a include carvings of male Saiva ascetics bearing skull-staves (khatv nga) and engaging . a in erotic activities, while there is evidence of sacricial cultus as wellboth sug115 Svarnadrimahodaya, cited by Panigrahi, Archaeological Remains at Bhuvaneswar, 233. Cf. Tantrasad. bh va 21, which provides a vidy -mantra addressing Aghor as C munda K p lin. Tantrasadbh va a a a a . . a a 21.154cd56ab:

siddhac mundakalpedam kathitam sarvak madam 154 a .. a . . atah parataram vaksye aghory bhmavikram | a a . . . c munde []ti padam adau k p lini atah param 155 a .. a a . sv h ntam pranav dyam ca mulamantram idam subham | a a . . a . .
vikram a

] conj.; vikram m mss a

k p lini ] em.; kap lini mss a a a

This kalpa of Siddhac munda, which grants all wishes, has been spoken. Next, I shall a . . teach something further: [the kalpa] of Aghor, the erce indomitable one. The word camunde at the beginning, then kapalini, ending with svaha and beginning with om .. . [i.e. om camunde kapalini svaha]: this is the auspicious root mantra. . .. As I mention in chapter 4, this mantra appears modelled upon the nine-syllable vidy -mantra of a . Aghore in the BraY : [om] hum cande kapalini svaha. s a . ..


Figure 2.14: Siva Gajasamh ramurti, Vait l Deul temple, Bhuvanesvar. AIIS Photo Archive. a . a

gestive of a Vidy ptha ritual context.116 The Tantrasadbh va, a Vidy ptha scripture a . a a . perhaps also of the eighth century,117 describes tantric ritual centered upon C muna . da and performed in temples of the Mothers (m trgrha), in one case with the aim of a. . . encountering yogins.118
Donaldson publishes a relief from Vait l Deul of three k p likas engaged in amorous activities; a a a a Tantra and S kta Art, vol. 3, g. 627. Panigrahi notes remains of a stone sacricial altar (yupa) outside of the temple. Archaeological Remains at Bhuvaneswar, 234. 117 On the Tantrasadbh va, see chapter 3, section 3, and chapter 5, section 3. a 118 Both of the references identied belong to Tantrasadbh va 21, referred to above (n. 115); one has a as its context the vidy -mantra and worship of Red C munda and the other, those of C munda as a a a . . . . a a Aghore or Canda K p linsupreme goddess of the BraY . In the latter case, the ritual is said to s a .. bring about direct encounter with the yogins. Tantrasadbh va 21.211cd13ab: a



the later literary evidence

. works of bana: the kadambar and harsacarita i .

In B nas K dambar and Harsacarita, prose works of the rst half of the seventh cena. a . tury composed by a contemporary of king Harsa (r. 60647119 ), we nd evidence . for tantric Saiva ritualists and practices characteristic of the bhairavatantras.120 However, while the divinities and varieties of ritual B na depicts are in no small measure a. consistent with the Saiva Yogin cult, yogins themselves nd no mention. In the K dambar, a love tale in ornate prose, the hero Candr pda happens upon a a . a jungle temple of the goddess Candik while journeying to the city of Ujjayin. Of .. a ciating over her worship is an elderly Dravidian ascetic (jaraddravidadh rmika), a . whose grotesque appearance and dubious magical practices are described in some detail. The Harsacarita too, B nas chronicle of the early life of King Harsa, proa. . . vides substantial evidence for the Tantric Saivism of the bhairavatantras. As with the K dambar, this work features in its third chapter a tantric ritualist: a Southerner a (d ksinatya) and great Saiva by the name Bhairav c rya, in this case an exhalted a . . a a and powerful guru rather than macabre magician. The association between Bhairav a ac rya and Harsas ancestor, the king Puspabhuti, culminates in the king serving as . . an assistant in vet las dhanathe zombie rite by which Bhairav c rya ascends to a a a a the skies as a semi-divine wizard (vidy dhara). a
m trgrham pravis. v tu pujayitv tu mandalam 211 a. . . a .t a .. japed yogevarm devm supat. as tadgateksanah | s . . .t . . . bhramam nam iv k se t vat tam nadate grham 212 a. a a a . . agacchanti tato devyo yoginyo vikrt nan h | a. . a After entering a temple of the Mothers and worshipping [their] mandala, one should .. incant the [Vidy -mantra of the] goddess Yoge var, () having a good cloth, ones gaze a s xed on thatup until the temple resounds, as though roaming through the sky (?). Afterwards come the goddesses, yogins of grotesque visage. The interpretation of 212bcd is somewhat uncertain; is the temple supposed to ascend into the sky? For supat. as, perhaps read svapat. as (ones ritual cloth), referring to a cloth inscribed with a ritual .t .t diagram. 119 D. Devahuti, Harsa: a Political Study, vii. . 120 B nas accounts of Saivism have been discussed by Lorenzen, K p likas and K l mukhas, 1623; and a. a a aa briey but with much substance by Sanderson, History through Textual Criticism, 11, 13 (n. 11). The latters remarks are referred to below.

74 The activities of the Dravidian ascetic in the K dambar are highly suggestive of a the power-seeking practices of the s dhaka taught in bhairavatantras. These include a . the use of magical ointments (siddh jana),121 powders (curna),122 minerals (dh tu),123 a a herbal salves (ausadh jana),124 bindis or forehead dots (tilaka),125 and mustard seed a . (siddh rthaka);126 alchemy (ras yana);127 the effort to nd hidden treasures and to ena a ter the netherworlds;128 practice of mantra-propitiation (mantras dhana) for attaining a invisibility;129 and seeking power over yaksa maidens.130 While these and similar . magical practices are not restricted to the bhairavatantras, they receive their greatest elaboration in the latter scriptures.131 In addition, B nas Dravidian ascetic is said to a.
He is referred to as having had one eye rupture because of a magical eye-ointment given by a quack (kuv didattasiddh janasphutitaikalocanatay . . . ). a a a . 122 anyadeagatosit su jaratpravrajit su bahukrtvah samprayuktastrvakaranacurnena (one who had often s a s . a . . . . . employed powder for controlling women on elderly nuns, who were staying [in the Candik temple .. a precincts] after arriving from another land). 123 samj tadh tuv dav yun (one in whom the wind-humour disease (v yu) of belief in [magical] piga a a a a . a ments/metals had arisen). 124 ingudkoakrtausadh janasamgrahena (one having a collection of herbal salves made from the bark s . . a . . of the Ingud). 125 duhsiksitaramanadis. atilak baddhavibhavapraty sena (one whose hope for power is bound to a [maga a . . s . .t ical] tilaka taught [him] by a poorly tutored ascetic). 126 asakrdabhimantritasiddh rthakaprahatipradh vitaih piacagrhtakaih karatalat danacipitkrtaravanaputena a a a. . . s . . . . s . . (one the inside of whose ears have been attened by slaps of the palms of those possessed by goblins, who had been chased away by [his] blows of sesame seeds enchanted by mantras no few times). 127 asamyakkrtaras yan nt k lajvarena (one who has an untimely fever brought on by improperly pera a a a . . formed alchemy). 128 avirbhutanidhiv davy dhin (one in whom the disease of belief in hidden treasure was manifest), a a a and lagn suravivarapraveapiacena (one whom the goblin of [desiring] entry to the netherworlds of a s s the Asuras has latched onto). 129 vardhit ntardh namantras dhanasamgrahena (one whose collection of rites of mantra-propitiation a a a . . for [achieving] invisibility had grown large). 130 pravrttayaksakanyak k mitvamanorathavy mohena (one in whom has commenced an infatuation a a a . . with the desire to be the lover of a yaksa-maiden). . 131 Note for example BraY xv.1314, which contains a typical enumeration of magical attainments: a adhun sampravaksy mi mah vet las dhanam | a a a a . a yena tv as. avidh siddhih s dhakasya praj yate 13 a a .t . a khadgam rocanap t lam vidy dharapadam tath | aa . a a . . . p dukau ajana caiva uttis. h ntardh nakam tath 14 a a a .t a . I shall now teach the great zombie rite, through which the eightfold siddhi arises for the s dhaka: [magical] sword and pigment, [ability to enter] the netherworlds, the state of a being a vidy dhara, [magical] sandals and eye-ointment, levitation (? uttis. ha), and invisia .t bility. This list includes three of the fruits sought by the Dravidian ascetic: the ability to enter the netherworlds, invisibility (antardh na), and magic ointment (ajana). Uddamarevaratantra 14.10 links the latter a s .. two, mentioning a siddh jana having invisibility as its purpose. Siddh jana is also mentioned as a a a magical attainment in BraY lxiv.71cd (gudik khadgavet lasiddh janaras ni tu). The BraY has a chapter a a a a a . a . ostensibly devoted to the subject of ajana: the ajanayogapatala, number lxviii. Regarding magical .

75 be versed in thousands of wonder-tales of Srparvata,132 a mountain closely asso ciated with esoteric Saivism, and not to waver in his self-identifcation with Siva (aiv bhim na).133 Although hence a Saiva, his devotion to Candik suggests engages a a .. a
. powders (curna), BraY lxxx.80cd mentions a recipe for one that, as is the case in B nas description, a a. . is used for bringing another person under ones power (curnena strpum m v pi y vajjvam vaam nayet). a. a a . s . Concerning the magical dh tus mentioned in connection with B nas Dravidian, note for instance the a a. mineral pigments harit la (yellow orpiment) and manahsil (red arsenic) in a list of siddhis in BraY a a . a xlv.5859ab: p dukau rocan caiva harit lam manacchil m | a a a . a. yog janaras devi khadgam cint manis tath 58 a a a . a . . upatis. anti vai tasya s dhakasya tu siddhayah | a .t .
58a p dukau ] em.; p duk Bya a a a 58c ras ] em.; ras m Bya a a. 58d manis ] em.; manin Bya . .

. As with curna, ajna, osadhis (herbs), etc., dh tus are substances used in ritual, which, it appears, a . are also thought to manifest in siddhi-bestowing varieties as fruits of ritual. Ras yana, which appears to a refer to the potent material product of alchemical processes, is listed alongside these as a siddhi-inducing substance, manifesting as the fruit of ritual. See for example Tantrasadbh va 20: a ksubhyanti purav sinyah s dhakam tu tato naghe 317 a . . a . pr rthayanti praveanti dadanti manasepsitam | a s rasam ras yanam divyam ausadhyo baladarpit h 318 a a. . . . . ajanam p dalepam ca p duko tha manahsil | a . a . . a gudik sastram l v yad anyam siddhik ranam 319 aa a a . . a .
319a ajanam ] em.; ajana Tsk Tskh Tsg .

319c gudik ] Tsg ; gutik Tsk Tskh . a . a

The women of the city get aroused; then, O sinless woman, [they] beseech the s dhaka, a give him entry, [and] proud of their power, grant him what his mind sought: divine elixir or ras yana, herbs, ointment, foot balm, sandals, red arsenic, pills, a set of weapons a (astram l ), or else another cause of siddhi. s aa (mss as reported by Dyczkowski.) Forehead marks or tilakas comprise a related category. Chapter ve of the BraY , for example, provides recipes for preparing tilaka compounds that bestow siddhi when a enchanted by mantra and applied in ritual. On the other hand, siddh rthaka or white mustard seeds do a not as far as I know appear to manifest as a fruit of ritual, although they are certainly used in magical rites. B nas reference to nidhiv da apparently refers to seeking hidden treasure by magical means. Coma. a pare for instance chapter nine of the Uddamarevaratantra, which makes several references to obtaining s .. wealth or hidden treasures as the result of ritual. For more detailed accounts, see the Buddhist Majurs mulakalpa, e.g. chapter 55. As for what B na refers to as asuravivarapravea, entry into the netherworlds a. s of the Asuras, tantric sources commonly call this p t lasiddhi, power [to enter] the netherworlds, aa or simply p t la. Cf. xv.1314 above. Regarding the yaksa-maidens the Dravidian is said to lust after, aa . these (yaksakany , yaksin, etc.) are frequently mentioned among the various females a s dhaka might a a . . . seek erotic power over. On such vakarana, see chapter 5 of the present dissertation. Note also the s . practice of yaksins dhana, rites specically aimed at gaining the control of a yaksin, and not simply for . . a . . erotic, but also magical ends. Cf., e.g., bori ms no. 503 of 189598, Yaksins dhana. In the BraY , a . a . an entire chapter is devoted to this subject: lxiv (labelled lx), the yaksins dhanapatala. B nas list of a. . . a . magical powers and substances bears comparison with that of the Buddhist Sub hupariprcch , discussed a . a by Ronald Davidson, Indian Esoteric Buddhism, 20001. 132 srparvat scaryav rtt sahasr bhijena (One well-versed in thousands of wonder tales of a a a a parvata). Sr 133 Although this phrase might mean pride in being a Saiva, abhim na also has in tantric literature a the technical sense of meditative identication with a deity. Note that the expression saiv bhim na a a occurs with this meaning in Tantr loka 13.252a, where Abhinavagupta quotes or paraphrases the a Nandiikh tantra. s a

76 ment in a tantric goddess cult, which is wholly consistent with the bhairavatantras of the Vidy ptha.134 a . Although the Harsacaritas Bhairav c rya is not referred to as such, his ritual a a . practices too are those of the siddhi-seeking specialist or s dhaka of the bhairavatantras, a and B na utilizes the technical terminology of tantric ritual. Bhairav c ryas practice a. a a of purvasev , preliminary worship of a mantra-deity, is described in some detail.135 a Additionally, the vet las dhana Bhairav c rya undertakes bears afnity in several rea a a a spects with this ritual as described in BraY xv, a chapter devoted to the subject. This a rite marks the culmination of his propitiation of the heart-mantra of Mah k la.136 On a a the fourteenth night of the waning moon, the king, duly initiated,137 joins three other disciples in taking up positions as guardians of the quarters in an ash-drawn mandala .. in a deserted temple near the cremation ground.138 In the center sits Bhairav c rya, a a upon a corpse. He performs re sacrice (homa) with black sesame seeds using a re lit in the very mouth of the body. As he performs homa, incanting mantras, spirits attempt to disrupt him, until nally a powerful n ga emerges from a ssure in the a earth near the mandala. Angered at not being made bali offerings, the n ga attacks, a .. but suffers defeat at the hands of the king, who nonetheless spares the snake lords life on spotting his sacred thread. At this point the ritual comes to fruition: won over by the pious king, Laksm herself manifests in his enchanted sword, at. ah sa. .t a .
Besides presiding over a temple of Candik , note that he is described as pat. ik likhitadurg stotrena a .. a .t a . (one who by whom a eulogy-hymn of Durg has been copied onto a small cloth), and jar m gaten pi a a. a daksinapath dhir jyavarapr rthan kad rthitadurgena (despite having grown old, he aficts Durg with a a a a a a . . prayers for the boon of overlordship of the Deccan). 135 Bhairav c rya states, in Sandersons translation, I have completed the preliminary service a a (purvasev ) of the great Mantra called the Heart of Mah k la by muttering it ten million times in a a a a great cremation ground while wearing a garland of black owers, a black robe and black unguent, with all the adornments prescribed in the Kalpa (bhagavato mah k lahrdayan mno mah mantrasya krsa a . a a .. nasragambar nulepen kalpena kalpakathitena mah smaane japakoty krtapurvasevosmi). History through a a a s . . a . Textual Criticism, 13 (n. 11). 136 tasya [mah k lahrdayan mno mah mantrasya] vet las dhan vas n siddhih. a a . a a a a a a a . 137 It is possible that the king takes samayadksa, the initiation of the samayin or pledge holder, by . which he becomes a Saiva neophyte. This is suggested by B nas terminology, in which niyamav n a. a (possessing/observing the rules) might be synonymous with samayin: ath tikr ntesv ahahsu pr pt y m a a . a a a. . ca tasy m eva krsnacaturday m saivena vidhin dksitah ksitipo niyamav n abhut (Then, when the days had a s a. a . . . a .. . past and that very [designated] fourteenth of the dark fortnight arrived, the King, initiated by Saiva procedure, became niyamav n). a 138 mah smaanasampabh ji suny yatane. a s a a

77 Granted boons, he requests only siddhi for his guru, Bhairav c rya, who thereupon a a joins the ranks of the vidy dharasying, semi-divine wizards. Of her own accord, a the goddess grants that Puspabhuti shall become progenitor of a line of great kings. . B nas description of vet las dhana matches in basic details with the more elaba. a a orate account of BraY xv.139 In the BraY s description, the ritual takes place in a a a cremation ground on the fourteenth or tenth night of the waning moon, with the aid of assistants. The basic sequence of action is identical: the s dhaka, seated on the chest a of the ritually prepared corpse, performs the re sacrice in its mouth, during the course of which various obstructing forces (vighnas) and spirits appear to distract or frighten him. Among the eight siddhis promised is the state of being a vidy dhara a (vidy dharapada), Bhairav c ryas aim in undertaking the ritual.140 In both descripa a a tions, the successful ritual ends with the s dhaka ying off into the sky.141 a Both the K dambar and Harsacarita make reference to tantric manuscripts and a . Saiva texts. Bhairav c ryas lips droop slightly, as though laden with the weight of a a the entire Saiva canon on the tip of his tongue.142 For his part, the K dambars a Dravidian ascetic possesses a collection of manuscripts that include a stotra of Durg a copied onto a cloth (pat. ik ),143 and palm-leaf manuscripts with red-lac lettering con.t a
This chapter contains, to the best of my knowledge, the most detailed account of such a practice in tantric literature. Its description of vet las dhana begins with verse 13. a a 140 See BraY xv.1314, quoted above. Those seeking only petty (ksudra) or middling (madhyam ) a a . siddhis discontinue the ritual when the sought attainments are achieved. 141 In the BraY , one attains a magical sword and becomes lord of emperors, with mastery over a a crore of aerial vehicles. BraY xv.63cd64: a y vad dhaste bhavet tasya t vat khadgam mah prabham | a a a . . tatas tam grhya mantrajo utpate gagan ngane 63 a . . . uttis. hati vim nastho vim naih pariv ritah | a a . a .t . tatr ruhya mah dhra cakravarttsvaro bhavet | a a s koty caiva vim n n m adhipatyam av pnuy t 64 a a a. a a . a
63b t vat ] em.; t vad Bya a a 63d gagan ] em.; gagana Bya a 64a uttisthati ] corr.; uttistati . .. .. Bya vim nastho ] em.; vim nastham Bya a a 64b pariv rtitah ] em.; pariv ritam Bya a a 64d . . . svaro ] corr.; svaro Bya 64e vim n n m ] corr.; vim n n mm Bya a a a. a a a.

At that time, there would appear in his hand a sword of great brilliance. Then, grasping this, the knower of mantras would y up into the vault of the sky. He rises up in an aerial vehicle, surrounded by aerial vehicles. Mounting that, he, very wise, he would become the lord of emperors, and would obtain mastery over aerial vehicles, by the crore.
142 143

jihv grasthitasarvaaivasamhit tibh reneva man kpralambitaus. ham. a s a . a . a .t As mentioned previously, he is described as pat. ik likhitadurg stotrena, One who by whom a stotra a .t a .

78 taining bogus spells and mantras.144 A text he possesses is moreover identied by name: Mah k lamata, the Doctrine of Mah k la, copied according to the instruca a a a tions of an elderly P supata ascetic (mah p supata).145 Similarly, Bhairav c rya is a a a a a said to have performed purvasev (preliminary service) of the great heart-mantra a of Mah k la according to the instructions laid down in its kalpa. It is possible a a that the Mah k lamata mentioned in the K dambar is the same kalpa of Mah k la a a a a a followed by Bhairav c rya, a kalpa being the manuscript of a text setting out the a a procedure for the propitiation of a Mantra.146 Sanderson remarks that neither the [mah k lahrdaya] mantra nor its kalpa can be identied now but the details of the proa a . cedure given by B na tally closely with what is laid down in Tantric Saiva sources.147 a. Providing parallels, Sanderson also points out that chapter fty-four of the BraY is a a kalpa of Mah k la.148 It should be mentioned, furthermore, that one of this chapters a a titles is in fact Mah k lamata, although this is not necessarily the same Mah k lamata a a a a B na appears to have known of.149 a. B na associates what we might consider tantric ritual proper with a variety of a. other exoticized practices. This nexus is particularly apparent in descriptions of the rites and penances undertaken by those seeking to avert the death of prince Harsas .
of Durg has been copied onto a small cloth. a 144 dhumarakt laktak ksarat lapatrakuhakatantramantrapustik samgr hina (One having a collection of a a . a a . a . small manuscripts of bogus (kuhaka) spells (tantra) and mantras on palm leaves with letters in smokey red lac). It might be possible that kuhaka is here a noun rather than adjective, in the sense of quack the source of the manuscripts of tantras and mantras. The term occurs in a different sense in Netratantra 18.89b, where Ksemar ja glosses it as yantrakrty di, rites involving yantras, etc. However, a . a . in the Svacchandoddyota, he instead glosses, kuhakam vism pakam mitahrdayapratyayak rndraj lapr yam a a a a . . . . (kuhaka means something astonishing which causes faith in those of limited awarenessvirtually magical trickery). 145 jrnamah p supatopadealikhitamah k lamatena, One by whom the Mah k lamata had been copied, . a a s a a a a as instructed by an old Mah p supata. a a 146 T ntrik bhidh nakoa, vol. ii, 78, citing Sanderson, History through Textual Criticism, 13. a a a s 147 History through Textual Criticism, 13. 148 Ibid. Following the colophon, Sanderson refers to this as chapter 52. 149 Although the colophon of BraY liv provides as the chapter title mudr p. h dhik ra (the Seat a a t a a of Mudr s chapter), verse 11cd gives its title as Mah k lamata (mah k lamatam hy etad yat surai[h] a a a a a . . paripujitam). Kalpas outlining the propitiation of a specic mantra-deity probably circulated in multiple versions. Nothing in this chapter specically matches B nas description of the purvasev of Mah k la, a. a a a although vet las dhana is mentioned in 194c as one of the applications of the mandala of Mah k la and a a a a .. the Mothers. As I discuss in chapter 4, some of the kalpa texts included as chapters of the BraY , such a as the Mah k lamata, bear tenuous relationships to the text as a whole and could conceivably have been a a independent in origin.

79 ailing sire, and, in the K dambar, in the practices Queen Vil savat of Ujjayin engaged a a in hoping to conceive a son. In the former case, for instance, the kings relatives fast indenitely while laying in the presence of the deity Ahirbudhan,150 noblemen propitiate the Mother goddesses by burning themselves with lamps,151 a Dravidian prepares to offer his head to the god Amardaka,152 a native of Andhra prays to the goddess Candik keeping his rampart-like arms (b huvapra) uplifted,153 servants a .. a propitiate Mah k la by burning incense resin upon their heads,154 intimates of the a a king offer slices of their own esh in re sacrice,155 and princes openly perform barter in human esh.156 While mantra-practice and other distinctive characteristics of tantric ritual do not gure in this account, B na associates both Tantra and a. these austerities with the same deitiesMah k la, Candik , and the Mother goda a .. a desses. In addition, he links both with South Indians and tribal peoples.157 For her part, Queen Vil savat of the K dambar engages in a wide range of orthodox and a a
The royal skandh v ra, military campthe capital, according to E.B. Cowell and F.W. Thomas (p. a a 135)is thus described: kvacit pratiayitasnigdhab ndhav r dhyam n hirbudhnam (in which, in one place, s a aa a a Ahirbudhan is being worshipped by close relatives [undertaking the vow of indenite] laying before the deity). Ahirbudhan or Ahirbudhnya, a minor and archaic Vedic divinity, is listed in Mah bh rata a a i.60.23 among the eleven rudras who are mind-born sons (m nas h putr h) of Siva (sth nu). a a. a. a. 151 kva cid dpik dahyam nakulaputrakapras dyam nam trmandalam ([where,] in one place, the group of a a a a a. .. Mothers is being appeased by sons of good families being burnt by lamps). 152 kva cin mundopah r haranodyatadravidapr rthyam n mardakam. Lorenzen, following the translation aa a a a .. . . of Cowell and Thomas, interprets this line as referring to a Dravidian ready to solicit the Vampire [Vet la] with the offering of a skull. K p likas and K l mukhas, 17. In all likelihood, this is however a a a a aa reference to a dravida preparing to offer a head to the deity Amardakaperhaps his own, as would t . the context of extreme self-mortication. Saiva sources speak of Amardaka or Amardakabhairava as a deity, but mention also a class of spirits called amardakas. For example, BraY lxii.19 lists amardakas a alongside the semi-divine siddhas, gandharvas, vet las, and kinkaras, while BraY lxxvii is a kalpa-manual a a of Mah mardakabhairava (cf. BraY lxxii.2cd: mah mardakadevasya kalpam vaksy mi tattvatahI shall a a a . . a . now teach the kalpa of lord Mah mardaka, as it truly is). Amardaka is, furthermore, one of eight a Bhairavas mentioned in Svacchandatantra 2. 153 kva cid andhroddhriyam nab huvapropay cyam nacandikam (where in one place Candik is being a. a a a .. a .. worshipped by a man of Andhra holding up his rampart-like arms). 154 anyatra sirovidhrtavilyam nagugguluvikalanavasevak nunyam namah k lam (where elsewhere a a a a a . Mah k la is being propitiated by young attendants deformed by the melting incense resin held on the a a head). 155 aparatra niitaastrnikrtt tmam msahomaprasakt ptavargam (where elsewhere intimates [of the king] s s . a a. a are engaged in offering homa of their own esh cut off by sharp knives). 156 aparatra prak sanarapatikum rakakriyam namah m msavikrayaprakramam (where elsewhere the proa a a. a a. cess of bartering human esh is being performed by the kings princes, openly). On mah m msavikraya, a a. see the subsequent discussion of the M latm dhava. a a 157 On the connection with tribals, note for instance the K dambars mention of Sabara women apa plying tilaka of red powder (sindura) to the image of Cand (p. 225, lines ines 2122). The introduction . . draws similar connections between Sabara tribesmen and the erce goddess.

80 unorthodox practices in her quest to conceive a child. These include numerous austerities and worship of Candik and the Mothers, among other deities. Some of her .. a practices involve mandala and mantra.158 .. In B nas works we thus nd abundant evidence for characteristic ritual systems a. of the bhairavatantras. Absent, however, is reference to the sacred gure of the yogin or yogevar. An association between tantric ritual and goddesses is nonetheless s evident: the Dravidian ascetic of the K dambar, for instance, ofciates as priest of a a Candik temple, and his engagement in tantric practices appears linked to devo.. a tion to the Goddess. This might suggest tantric Saivism as known to B na was in a. a stage where ritual forms characteristic of the bhairavatantras and a tantric cult of Candik had developed, but in which focus upon the yogin was not yet central or .. a well-known. In period literature, especially B nas Candsataka and the Devm h tmya of the a. a a . . M rkandeyapur na, the terrible warrior goddess Cand or Candik becomes one of the a .. a. . . .. a principle ciphers for emergent conceptions of a singular Mah dev, whose identity a subsumes the myriad manifestations of feminine divinity. Although thus linked to all goddesses, Cand might in this period have been identied in particular with C mua . . nda, leader of the Mothers, one indication being the synonymity of these names . . in the BraY .159 The Harsacarita links Candik to the god Mah k la as consort,160 a a a .. a . forming a divine couple whose cult is not however well-represented in surviving tantric literature. One detailed and early treatment of the tantric cult of Mah k la a a
158 The description begins, yad yac ca kimcit kuta cic chur va garbhatrsnay tat tat sarvam cak ra (And s sa a . .. . a . she did eveything she heard from anywhere out of her yearning for [conceiving] a fetus). Note for instance that she engaged in lustration and auspicious rites in the crossroads on many fourteenth nights [of the lunar fortnight], standing in the middle of a mandala drawn by the great king [i.e. .. her husband], through which the deities of the directions were gratied by gifts of various food offerings (mah narendralikhitamandalamadhyavartin vividhabalid n nanditadigdevat ni bahulapaksacaturdaa a a a s .. . a niasu catuspathe snapanamangal ni bheje). It is also said that she bore bamboo mantra-caskets contains . ing birch-bark written on with yellow pigment, and that she fastened [on herself] strands of herbs having protective cords (gorocan likhitabhurjapatragarbh n mantrakarandak n uv ha | raksapratisaropet ny a a a a .. a . a. osadhisutr ni babandha). . 159 In the BraY , the seventh Mother is variously called Candik , Carcik , and C munda. On this a a a .. a . . matter, see my annotation on BraY ii.16. a 160 Candik is described as mah k l bhis rik vesavibhramam bibhratm, exhibiting coquettry with the a aa a a . . .. a . guise of a women on a night rendevous with Mah k la. a a

81 is BraY liv, the Mah k lamata. This describes worship of the deity in a mandala a a a .. of the eight Mothers, a conguration unusual in the BraY and probably archaic. a In addition, two late sixth-century shrines of Ellora depict a skeletal divine couple in the company of the Seven Mothers, a couple who have been, with a degree of plausibility, identied as Mah k la and K l.161 The seventh-century tantric cults of a a a Mah k la and Cand might therefore have emphasized Mother goddesses; but B na a a a. . . does not intimate this association, nor does he make clear reference to a tantric cult of Mothers. These goddesses are nonetheless mentioned: Bhairav c rya, for instance, a a is said to dwell near a deserted temple of the Mothers, the queen Vil savat visits a Mother shrines, and reference is made to forest temples of C munda.162 a . . Taken as a whole, B nas works attest the existence of characteristic rituals and a. deities of the bhairavatantras, as well as the emergence of goddess-centered tantric practice. It is noteworthy that B na associates t ntrikas with the Deccan and South a. a India. While I am hesitant to place excessive value upon this regional association, it at least suggests Tantric Saivism was associated with areas and peoples marginal to the North-central heartland of Brahmanical culture, the Middle Country (madhyadea). s It is not clear whether tantric cults of the Mothers were yet widespread, nor whether the yogin had emerged as a sacred gure. Hence, while B nas early seventh-century a. references to Tantra are by no means incompatible with the Saiva yogin cult, they could instead imply a phase of Saivism that predates its development. There might, in addition, be historical signicance to the divergent characterizations of B nas two a. t ntrikas: while Bhairav c rya is presented as a respected and powerful gure, the a a a Dravidian ascetic is described as an exotic, morally ambivalent, and rather inept sorceror. It is tempting to read into this distinction an emergent divide between a
See the discussion in the previous section. In Harsacarita iii, when the king asks the whereabouts of Bhairav c rya, he is told, asya jra a . nam trgrhasyottarena bilvav . ik m adhy ste (he sits in a grove of bilva trees to the north of a decrepit at a a . a. . . Mother-goddess temple). Chapter seven makes reference to forested areas having temples of C muna . da erected in dense groves (gahanatarusandanirmitac mundamandapair vanapradeaih). Cf. Yokochi, Rise a .. .. s . . . .. of the Warrior Goddess, 108 (n. 81). In the K dambar, Queen Vil savat went to temples of the Motha a ers in the vicinity, where faith is displayed [or perhaps, where portents/signs are shown] (daritas pratyay ni samnidh nam trbhavan ni jag ma). a a a. a a .
162 161

82 well-established tantric Saiva cult of Bhairava, on the one hand, and newer, marginal, and more esoteric goddess cults. It is within the latter that the roots of the Yogin cult would lie.
the gadavaho of vakpati .

In the Gadavaho, a Prakrit poetic work composed by V kpatir ja, a contemporary a a . of B na, there occurs a reference to Kaula women (kalan ro) in a hymn to the a. a goddess Vindhyav sin, She Who Dwells in the Vindhya Mountains. Travelling a through the Vindhya range of central India, the king Ya ovarman approaches a juns gle cave temple and offers eulogy, making vivid allusion to the sanguinary cult of the presiding Goddess. The hymn links her identity with that of the singular Warrior Goddess who is at once Durg , K l, Candik , and so forth, as well as the a a .. a pacic P rvatprimary locus of conceptions of the one Mah dev.163 As does B na, a a a. V kpati associates worship of the erce goddess with exotic, macabre practices and a peoples marginal to the civilized Middle Country (madhyadea)in this case forests dwelling sabara tribals.164 Signicantly, the description of the Goddesss cult suggests a tantric, k p lika ritual dimension, and makes reference to a human sacrice viewed a a eagerly by throngs of female practitioners, described as kaula women.165 This seems to be the earliest use of the word kaula in literary works to describe a Saiva sect or its members; the reference is, furthermore, somewhat surprising given that the term is not employed in this sense in Vidy ptha texts such as the BraY and a . a Siddhayogevarmatathe earliest Saiva literature connected with yogins.166 Kaula s
163 Gadavaho, verses 285338. Cf. B nas Candsataka, where the conation of the Warrior Goddess a. . . . with P rvat is prominent. For a discussion focused on the iconic dimension of the Goddesss descripa tion in Gadavaho, see Yokochi, 14651. . 164 Cf. B nas K dambar, p. 65, in the description of the general of the Sabaras. Quoted by Yokochi, a. a ibid., 148. 165 The k p lika ritual context is suggested by reference to the temple environs as a cremation ground a a (mas na), at which heroes (vra) sell human esh, in verse 327. This is briey discussed by Yokochi, a. ibid., 147. On the reference to Kaula women, see Sanderson, History through Textual Criticism, 11 (n. 9). He translates the relevant passage as follows: The Kaula women seem to form a shrine in the air as they clamber over each other in their eagerness to watch a victim of human sacrice being carved up (visasijjantamah pasudamsanasambhamaparoppar rudh | gayane cciya gamdhadim kunamti kalanaro). a a . a . . . . . . . . . . 166 On this distinction of Kaula and Vidy ptha, see chapter 3. a .

83 could in this case have the sense of [women] of/belonging to the Mother-goddess clans, referring to female tantric initiates, but not necessarily implying the ritual systems known from extant Kaula scriptures. Nonetheless, the context of corporate worship by women belonging to clans of the Mothers does suggest the existence of a Yogin cult connected with the erce Goddess. This might, conceivably, be an historical development related to identication of the Warrior Goddess with C mua nda, leader of the Mothers, with whom Skandapur na 171 connects the cult of yoa. . . gins. Though vague, the Gadavahos reference to female practitioners thus provides . a valuable addition to the evidence from B na, who in other respects provides more a. detailed accounts of Tantric Saivism.
the da akumaracarita of dandin s ..

A narrative probably of the late seventh or early eighth century,167 Dandins Daas .. kum racarita has an episode of passing signicance to the yogin cult. A young rogue, a Kalahakantaka, plots to entrap the beautiful, faithful wife of a merchant by having .. her accused of witchcraft. He claims before the merchants guild to have seen her at night in the cremation ground dragging a corpse from a funeral pyre; upon seeing his contrived evidence for this, the townspeople believe her to be a sakina dreadful female being described in Saiva typologies of yogins. Abandoned by her husband, she falls into the hands of the rogue.168 Despite the ironic and undoubtedly false etymological connection to saka (leafy greens), it requires little imagination to guess what business the decidedly nonvegetarian sakin might have with a half-roasted human corpse. Ksemar ja quotes a . the following denition from the Tantrasadbh va: a A female who, for the purpose of shapeshifting, ever drinks the uids of
Dandin has been linked with the Pallava court of the late seventh and early eighth centuries; .. Walter Smith provides a brief bibliography on this matter in The Visnu Image in the Shore Temple .. at M mallapuram, Artibus Asiae 56 (1996): 22. Isabelle Onians suggests the period of 690725 for Daa ndin, placing Daakum racarita within the earlier part. Onians, trans., introduction to What Ten Young s a .. Men Did, 25. 168 Daakum racita, chapter 11. I have consulted the text and translation given in Onians, ibid., 44049. s a

84 living beings, after drawing them close by artice, and who after obtaining [that uid] slays the creaturesshe should be known as a sakin, ever delighting in dreadful places.169 The Daakum racaritas reference is the earliest I am aware of to the sakin in nons a tantric literature, occuring in the context of an episode similar in nature to accounts a a of sakins and dakins in the much later Kath sarits gara. Yet here the description . so lacks in detail and context as to be unclear whether it reects inuence from a tantric cult of yogins; belief in witches and female cremation-ground spirits are undoubtedly older phenomena. However, the period and terminology in question suggest the possibility that this sakin tale is directly inspired by tantric typologies of female spirits and contemporary perceptions of cremation-ground ritual.
the malatmadhava of bhavabhuti i

Substantial evidence for the Saiva Mantram rga, and more particularly the Yogin a cult of the bhairavatantras, emerges in Bhavabhutis M latm dhava, an early eightha a century play.170 While not referring to specic texts,171 it portrays ritual practices which reect the cultic milieu of the Vidy ptha bhairavatantras. This drama revolves a . around the clever efforts of a Buddhist nun, K mandak, to bring about the marriage a of M dhava and M lat, despite the latter having been promised to a favorite of a a the king. Events take a dramatic turn when M lat is abducted in the night by a a erce Saiva s dhaka named Aghoraghanta, and his female attendant, a yogin named a .. Kap lakundal , who are intent on offering her in sacrice to the goddess C muna a .. a .

Netroddyota, quoted in the commentary on Netratantra 2.71: chalen krsya pibati ksudr pr nipayah sad | a .. a a. a . . rupaparivartan rtham labdhv p tayati paun | a a a s . sakin s tu vijey raudrasth narat sad | a a a a a

With minor variants and corruptions, this corresponds to 16.163cd64 in Dyczkowskis draft edition of the Tantrasadbh va. a 170 On the dates of Bhavabhuti, see V.V. Mirashi, Bhavabhuti, 111. 171 Siva is, however, described as nigamanidhi, the repository of scripture. Act 9, verse 4. It seems probable that nigama refers to the Saiva Agamas/Tantras, for no other large body of scripture assigns its origin to him.

85 da.172 . This turn of events forms the subject of the M latm dhavas fth act, set in the a a environs of a large cremation ground (maanav . a) in the city, near to which stands a s s at temple of Kar l , a local C munda. Distressed at the prospect of M lats imminent aa a a . . marriage, M dhava sees no recourse but to adopt the desperate measure of selling a human esh (mah m msasya vikrayah) at night to the spirits of the cremation ground, a a. . in return for his cherished boon.173 A sword in the right hand and hunk of esh in the left, M dhava sets out for the cremation ground. Simultaneously, Aghoraghana . ta, a skull-bearing s dhaka from the sacred mountain Srparvata who dwells now in a . a nearby forest,174 also avails himself of the fourteenth night of the waning moon to fulll his pledge to sacrice a jewel of a woman to the goddess.175 While he proceeds to abduct the sleeping M lat, the yogin Kap lakundal ies off to the a a .. a cremation ground to gather there items needed for the ritual. M lat is then led a before the goddess adorned in the red garments and garlands of a sacricial victim, while Aghoraghanta and Kap lakundal offer worship. Hearing his beloveds a .. .. a a piteous adieu to the world, M dhava rushes to the temple and slays Aghoraghanta, .. rescuing M lat but incurring the wrath of Kap lakundal . a a .. a
172 David Lorenzen has discussed the M latm dhava with the aim of highlighting evidence for a lost a a Saivite sect called the k p likas, of which he assumes Aghoraghanta to be a representative. The K p likas a a a a .. and K l mukhas. Two Lost Saivite Sects, 5357. The limitation of this perspective is that the term k p lika, aa a a assuming a Saiva sect by this name in fact existed, can also be applied as a descriptive term to any and all carrying out mortuary observanceswhether Atim rga Saivas, Tantric Saivas, or Buddhist t ntrikas. a a Caution is called for in reading more than this into the term in any particular instance. Lorenzen, whose work was rst published in 1972, was then unaware of the existence of a large body of tantric Saiva scripture attesting k p lika practices, although he addresses this in a cursory fashion in an appendix to a a the second edition. 173 In Act Four, M dhava says to himself, hanta sarvath samsayitajanmas phalyah samvrtto smi | tat kim a a . a . . . id nm kartavyam | na khalu mah m msavikray d anyam up yam pay mi | (Alas, my lifes fulllment has a . a a. a a . s a become completely jeopardized. What should I hence do now? I see no recourse at all other than the sale of human esh). Prose preceding verse 8. Sanskrit text as printed in Franois Grimal, ed., Hariharaviracit M latm dhavatk . a a a . a 174 In Act One, K mandaks disciple Avalokit describes him thus: a s dhaka named Aghoraghanta a a a .. who has come from Srparvata, who roams at night, dwelling in the forest not far away and bearing a skull (. . . sirpavvad do aadassa rattivih rino nadidur rannav sino s hassa mundadh rino aghoraghan. a a a . . a .. a . a a . .. .t namadheassa . . . ). . 175 Kap lakundal remarks to herself, kathitam ca me guruna vatse kap lakundale adya bhagavaty h a a a. .. a . . .. kar l y h pr g upay citam strratnam upahartavyam (And my guru said to me, my dear Kap lakuna a a. a a a . . dal , today the jewel of a woman previously promised to the goddess Kar l is to be sacriced ). Act aa . a Five, prose following verse 4. Cf. verse 25.

86 The vengance of the yogin bears fruit when she abducts M lat, whisking her a off to the sacred mountain Srparvata to offer up in sacrice. In this terrible deed she is foiled by Saud min, K mandaks rst disciple, who has herself attained a a the powers of a yogin and undertakes the Skullbearer-observance (k p likavrata) on a a . . Srparvata.176 Saving the day, Saud min then applies her power called aksepin siddhi a to shuttle the heroine and hero through the sky back into the presence of K mandak. a The good yogin then ies off to rescue M lats father from suicide, returning again a to magically revive the swooning maiden. M dhavas appearance and activity make it clear he was carrying out a tantric a vrata or special observance. Although I am unable to nd a description in tantric Saiva sources corresponding precisely to the vrata M dhava undertakes, almost all a aspects have parallels. While the practice of offering human esh in re ritual (homa) is more common, M dhavas esh-for-boons barter with cremation-ground spirits is a attested in tantric sources, and mentioned, for instance, by B na in the Harsacarita.177 a. . M latm dhavas twelfth-century commentator Harihara attributes such a practice to a a the authority of the Siddhayogevarmata, and quotes a passage to this effect which s does not occur in the texts surviving short recension.178 Svacchandatantra 4, moreover, in the context of initiatory dream prognostication, lists among auspicious portents the sight of heroes, zombies, and siddhas trading human esh in the cremation ground.179 This cremation-ground practice is apparently mentioned in the Buddhist
In Act One, Avalokit says to K mandak, Your reverence, Saud min now carries out a k p lika a a a a a observance on Srparvata, having obtained wondrous powers of mantrasiddhi (bhavadi s sod min ahua a . na sam s didaccariyamantasiddhippah v sirpavvade k v liavvadam dh redi). aa a a a a a . . 177 A relevant passage is quoted above (n. 156). Note also a reference in Harsacarita vi to the son of . Prayota, described as mah m msavikrayav dav tula, mad with the doctrine of selling human esh; he a a. a a loses his life to a vet la called T lajangha. a a 178 The passage attributed to the Siddhayogevarmata reads, vrahast n mah m msam grhtv vr y s a a a. . . a a a bhimatavarad nam (Having accepted human esh from the hand of a hero, there is the bestowing of a the desired boon to the hero). Comments ad prose preceding verse 8 (p. 154). While the rst eight syllables could form an odd-numbered p da, the remainder is unmetrical. a 179 In the Kashmiri recension of the Svacchandatantra, as reected in the ksts edition, dreaming of a cremation ground and dreaming of the sale of human esh appear to be separate portents. However, in the Nepalese recension, as well as this passage as it was redacted into Tantrasadbh va 9, the cremation a ground forms the setting for this esh trade. Comparing the ksts edition (Stksts ), three mss of Tantrasadbh va 9 (as reported by Dyczkowski), and an early Nepalese Svacchandatantra codex (Stcod = nak1a 224/ngmpp reel b28/18), all of which have independent value for constituting the text, the following

87 Sub hupariprcch as well.180 As for timing, the fourteenth night of the waning lunar a . a fortnight (krsnapaksa) on which both M dhava and Aghoraghanta undertake their a .. .. . . rituals is customarily favored for cremation ground rites, including those involving esh offerings.181 M dhava is described as dark, his body nonetheless grey.182 This could be a refa erence to the otherwise dark M dhava being pale from love-sickness, which is in fact a alluded to elsewhere in the play.183 However, this might possibly refer to his dark body being smeared with ashes, for vratas described in Saiva sources usually enjoin wearing particular colors, especially black or red, or else going naked, smearing the body with ashes, etc. He has bound his curly hair upwards in the fashion of the
is proposed as the original form of the verse in question (Svacchandatantra 4.14cd15ab, in the ksts edition): jvalatpitrvane ramye vravreibhir vrte 14 s . . vravet lasiddhai ca mah m msasya vikrayam | a s a a.

vane ] Stcod Tsk, kh, g ; vanam jvalat ] Stcod(pc) Stksts Tskh ; jvat Stcod(ac) ; jvalan Tsk, g . ksts cod Tsk,kh ; ramyam Stksts ; ( - - ?) Tsg vre ibhir ] Stksts Tsk,kh ; vre abhir St ramye ] St s s . siddhai ] Stksts ; siddhi Stcod Tsg ; Stcod ; (?) Tsg vrte ] Tsk,kh ; vrtam Stcod Stksts ; (?) Tsg s s . . siddh s Tsk,kh a vikrayam ] Stcod Stksts Tsk ; vikramam Tsg ; (?) Tskh

And in a lovely, aming cremation ground, surrounded by heroes and heroines, [one sees in dream] the trading of human esh by heroes, zombies, and siddhas. Ksemar ja, however, understands siddhai in the Kashmiri recension as an associative instrumental a s . (vravet lasiddhai ceti sah rthe trty ). a s a . a 180 Davidson, Indian Esoteric Buddhism, 203. 181 For example, in BraY xv, the practice of vet las dhana, a cremation-ground rite involving a corpse, a a a is enjoined on the eighth or fourteenth night of the waning moon (xv.15). Siddhayogevarmata 13 pros vides a further example of an heroic rite (s dhana, performed after completion of purvasev ) on the a a fourteenth dark lunar night. In this case, the hero fasts three nights and offers homa of human esh before preceding on the fourteenth to the cremation ground, naked and alone. Standing erect, he recites the mantra until yogins surround him, to whom he offers blood from his left hand as the guest-offering (argha). Tantrasadbh va 17 attests a rite in which one proceeds to the cremation ground on the krsnacaturda, a s .. . bearing a human skull and reciting a mantra of C munda. One offers there, in front of the cremation a . . ground, animal esh as bali. om c munde tiri tiri cchinda chinda vidh tre hiri hiri sphura sphura tis. ha tis. ha vividha vividha a . a .. .t .t sv h | anena smaanam gatv m nusyakap lam dh rayitv krsnacaturday m paum msena a a s . a a . a . a a .. . s a. s a. a tasy grato balim datv t vaj japed y vat sphutati | ak sena gacchati yatra smaane tis. hati tatra a a a a s . . .t grham payati | s . . mss as reported by Dyczkowski; s/ silently regularized. The sequence of events envisioned in the last s sentence is unclear (he goes via the sky[;?] in the cremation ground where he stands[,?] there he sees a building). 182 . Act Five, verse 5a: [ya esah] kuvalayadalay mo py angam dadhat paridhusaram. s a . . . 183 In Act Three K mandak refers to M dhava having a body by nature dark like the priyangu vine, a a now pale and emaciated through his love-sickness (priy nguy m ngaprakrtir api c p ndumadhuram vapuh a s a a a a. . . . . ksamam ksamam vahati ramanya ca bhavati, 9cd). . . . . . s

88 matted, upward-fastened locks of the Saiva ascetic.184 In addition, his bearing of a sword has parallels in cremation-ground observances. The Kulapac sik , for exama a ple, refers to great heroes who, devoted to worship and meditation, swords in hand and lusting to drink human blood, wander the cremation grounds seeking encounters (melaka) with yogins.185 Compare also the mahocchusm vrata of BraY xxi, fourth a . a of ve vratas corresponding to the Four Devs and Bhairava. In this observance, the s dhaka wears black clothing and garlands and a variety of ornaments, applies red a lac on the feet, and bears a skull, skull-staff (khatv nga), a variety of weapons, and . a a damaru-drum, etc. Meditating and practicing penance in solitude, he wanders by . night through places that include the cremation ground.186 In addition to this portrayal of a radical vrata, Bhavabhutis drama attests tech nical knowledge of a number of other aspects of Saiva ritual. Aghoraghanta is de.. scribed as a s dhaka, the specialist engaged in solitary practices in pursuit of siddhi. a He had practiced his mantras dhana in this very temple of Kar l , near the cremaa aa tion ground,187 and dwells in the forest nearby. His practices are said to involve roaming by night, the sacrice of living beings, and carrying out the Skullbearerobservance.188 His close connection to a female disciple is also noteworthy: her role
Act Five, prose preceding verse 5: tat ko yam atigambhramadhur krtir uttambhitakutilakuntalah krp a . . . . a nap nih smaanav . am avatarati (spoken by Kap lakundal ). at a .. a . a . . s 185 Kulapac sik 3 (f. 4v): a a a puj dhy narat ye tu nar srkp nalampatah | a a a . a . .. nityodyukt mah vr h khadgahast h pratis. hit h a a a. a. . . t a. smaananilaye deva r trau niyatane rat | s a a . katham nu labhate nityam yogmel pakam hara a . . .
lampatah 184


a . . ] em.; lampatam ms nityodyukt ] conj.; nityodrakto ms . . . nu ] conj. (Isaacson); na ms hara ] em.; harah ms .

pratisthit h ] em.; pratisthit .. a . .. a

O Hara, how do those great heroes who are devoted to worship and meditation, lusty for drinking human blood, endeavoring constantly, standing with swords in hand, devoted to wandering the cremation grounds at night, O god, always obtain mel pa with the Yogins? a nak 1-1076 (ngmpp a40/13); transcription courtesy of Somadeva Vasudeva. 186 BraY xxi.83cd100. Owing to length, an edition of this material is not provided here. a 187 Act Five, prose following verse 4: smaanav . asya nedyah kar l yatanam yatra paryavasita s at . aa . a . a mantras dhanasy smadguror aghoraghan. asy jay saviesam adya puj sambh ro may samnidh panyah. On a a a s . a . a . .t a mantras dhana, the practices following the s dhaka-consecration for achieving mastery of ones mantraa a deity, see Brunner-Lachaux, Le S dhaka. It appears that it was the norm to practice mantras dhana in a a a single, isolated place, as Aghoraghanta is said to have done in the temple of Kar l . Ibid., 431. aa .. 188 Avalokit describes Aghoraghanta as one who wanders by night and a skull-bearer, as quoted a .. in n. 174 above. In Act Five, verse 29, Aghoraghanta refers to himself as hims ruci (fond of violence), .. . a

89 might be restricted to that of a student and assistant, but the implication could be that she is his ritual consort as well. This possibility lacks explicit intimation. While mortuary, k p lika elements were present in the Saiva Atim rga, cremationa a a ground rites focused upon control of the powerful and dangerous forces which congregate there appear unique to the Mantram rga, and particularly characterize a the Vidy ptha bhairavatantras. Bhavabhuti vividly describes the cremation grounds a . frightful night denizens, which include beings such as ulk mukhas (meteor mouths), a putanas (stinkers), and piacas (goblins), as well as female varieties of the same.189 s M dhava bears his food-offering in the left hand, as bets radical tantric ritual. He a advertizes this esh as uncut by weapons, a criterion that apparently makes it accept able to the spirits.190 Saiva sources however sometimes praise the corpse of one fallen in battle as particularly suitable for ritual use.191 Beyond the already extreme practice of making offerings of human esh, scriptures such as the BraY and Yoginsac ra a a mention the rite of actual human sacrice as well,192 whether or not this actually took place with any frequency. Given its sensational quality, Bhavabhuti is not alone among medieval poets in utilizing this theme for dramatic purposes. Attempted human sacrice to a erce goddess comprises, for example, the dramatic focus of the
and pr nyupah raketanajus (frequenting places where there is sacrice of living beings). a. a . 189 Act Five, verses 1118. The ulk mukha is a somewhat uncommon category in Indian spiritology. a In his commentary on the Netratantra, Ksemar ja identies these with the piaca (goblin): glossing a s . piacaih in 2.14b, he states, smaan div sina ulk mukh h piac h. Cf. his comments ad 18.99d (piac aucis . s a a a a . s a . s a s sth n div sina ulk mukh h). Tantrasadbh va 19 mentions a ksetrap la named Ulk mukha. a a a a a. a a a . 190 M latm dhava, Act Five, verse 12: a a bho bhoh smaananiketan h kataputan h | a. . a. . s aastraputam avy jam purusangopakalpitam | s a . . vikryate mah m msam grhyat m grhyat m idam a a. . . a. . a 12

O Corpse Stinkers, denizens of the cremation ground, the greatest of meats is for sale, not puried by [being slain in battle with] weapons, unadulterated, prepared from the body of a man; please take it, take it!
191 Cf. Svacchandatantra 13.24cd: ranaastragh tapatitam narapiitam trimadhusamyutam juhuy t, One a s . a . s . . . should offer oblations of human esh felled by the blows of weapons in battle, mixed with the three types of honey. 192 Note Yoginsac ra 6.26cd, a somewhat random example: sarvayogsvarsiddhi[r] m nusat paun d a a . s a bhavet (Due to a human as sacricial victim, there would be [attainment of] power over all yogins). But note also, in contrast to M latm dhava, that this source (6.100ab) appears to prohibit sacrice of a a a woman: str pautve niyukt tu mah siddhivigh tak[ ] (but a woman employed as a sacricial victim s a a a a ruins the major siddhis).

90 Yaastilaka of Somadevasuri as well, and nds mention in numerous other works of s the medieval period. Particularly relevant are the descriptions and activities of the two yogins, Kap la a kundal and Saud min. Both carry out Skullbearer-observances (k p likavrata), and a a a .. a the formers k p lika attire is described in vivid detail. Both frequent the sacred a a mountain Srparvata, and possess the spectacular power of ight. Although the source of Saud mins ability to y is unstated, it perhaps stems from the same means a by which she has the power to draw others along through the air. When the bewildered M dhava and Makaranda inquire as to her identity, she replies, this you shall a surely understand, and rises: I now shall for your benet use this power of drawing [through the air], which arises from my dedication to the guru, the observances, penance, tantras, and mantras.193 This list bears comparison with accounts of the practices s dhakas carry out in order a to attain encounters (mel pa) with yogins.194 a Kap lakundal introduces herself and her powers with the same phrase, iyam a .. a aham id nm. But the source of her ability to y differs considerably: the pernicious . a . practice of extracting the essences of living beings by invasive yogic methods. She describes herself as untired by a ight fueled by extraction of the ve nectars through the bodys channels (n d). This she does while absorbed in meditation upon the Self a. as Siva in the heart lotus,195 in a cakra on which the six ancillary mantras (anga) have been installed.196 She praises Siva as lord of the saktis (aktin tha), surrounded by s a

Act Nine, verse 52: j syathah khalv etat | (utth ya) iyam aham id nm, a a a . .

gurucary tapastantramantrayog bhiyogaj m | a a a im m aksepinm siddhim atanomi siv ya vah 52 a . .. a . . . Although I cannot locate parallels for the expression aksepin siddhi, the ability of yogins to enable men to y pervades the literature, indeed characterizes in no small measure their popular image. It is worth noting that here, tantra seems unlikely to mean text, but instead spell or something of this nature. Cf. the previous discussion of the Gangdh r inscription. a 194 Cf Svacchandatantra 15.32cd33, BraY lxxiii.40ab, and BraY xcix.25. a a 195 a . Cf. Yoginsac ra 6.56ab: hrccakramadhyam atm nam dhy tv sivasamaprabham. a a a . 196 Act Five, verse 2:

91 whom he is meditated upon in the heart in a cakra of sixteen n ds, bestowing siddhi a. upon s dhakas.197 This conception of the deity with six ancillaries (sadanga) in the a . . center of a cakra of n ds visualized in the heart is consistent with terminology of the a. BraY , which attests similar congurations of sixteen.198 Matching more precisely, the a unpublished Yoginsac ra of the Jayadrathay mala describes as what appears to be its a a basic conguration the deity and six ancillaries (sadanga), who are the Six Yogins, . . in a cakra of sixteen n ds.199 The latter correspond to the vowels, as do the sixteen a. mantra-deities of the BraY s bhautikacakra and kulacakra, the latter moreover having a a special association with yogins.200 It seems probable that Kap lakundal speaks of a a .. a
iyam aham id nm a nityam nyastasadangacakranihitam hrtpadmamadhyoditam . . . . . . . payant sivarupinam layavaad atm nam abhy gat | s s a a a n dn m udayakramena jagatah pac mrt karsanad a. a a . a . . . . apr ptotpatanaram vighatayanty agrenabho mbhomucah a s a . . I have now arrived, ever viewing by the power of meditative absorption the Self, in the form of Siva, arisen in the heart-lotus and placed in a cakra where the six ancillary [mantras] have been installed. Parting the clouds in the sky before me, I am untired by my ight, [caused] by extraction of the ve nectars of living beings (jagat) via the ascending sequence of n ds. a.

Act Five, verse 1: sadadhikadaan dcakramadhyasthit tm s a. a a . . . hrdi vinihitarupah siddhidas tadvid m yah | a . . avicalitamanobhih s dhakair mrgyam nah a. . . a . sa jayati parinaddhah saktibhih saktin thah 1 a . . . . Victorious is the Lord of the Saktis, surrounding by saktis and placed in the middle of a cakra of sixteen n ds, his form placed in the heart bestowing siddhi upon those who know a. him, sought after by s dhakas with unwavering minds. a

198 Perhaps the most signicant cakra of sixteen mantra-deities is the kulasodaa cakra, the subject of . . s BraY xiv, which does not, however, use the terminology of n ds. It encodes a conguration of the a a. sixteen vowels around the central deity. The bhautikacakra of chapter nineteen, on the other hand, explicitly correlates its sixteen n ds with the vowels. a. 199 On the sixteen n ds as the vowels, cf., e.g., Yoginsac ra 11.47ab: n dsodaabh gotth [n] kram [d] a. a a . . . s a a a varnan samuddharet. Chapter ten narrates how the Six Yogins arose from the ancillary (anga) mantras . of the vidy , protecting Skanda from the Seven Mothers, who had become proud of their power after a slaying the demons and receiving Sivas boon. The Six arise as deities of the k lacakra, the same deity a conguration mentioned in chapter 5 in connection with extraction of the ve nectars. 200 This connection is suggested in the Goddesss opening question of BraY xiv, which teaches the a kula[sodaa]cakra or khecarcakra: . . s

yoginyah svalpabuddhyas tu svalpacitt lpasattvik h | a a. . bhartuh surusanapar gurubhaktisamanvit h 1 a a. . s . . t s m siddhir yath deva tam me bruhi sam satah | a a. a a .
1a yoginyah ] em.; yoginyoh Bya buddhyas ] em.; buddhy s Bya 1b s ttvik h ] em.; sattvik h a a a. a. . . Bya 1c bhartuh ] corr.; bhartu Bya su rusana ] em.; sva rusana Bya s . . s . . .

92 comparable conguration of sixteen n ds of the Sanskrit vowels encircling Bhairava a. and his six ancillary mantras. Sanderson has, in addition, drawn attention to the a fact that the process Kap lakundal alludes to of extracting the vital essences is de.. a scribed in considerable detail in the Yoginsac ra.201 As described in its fth chapter, a this involves placing the victim (pau) before one and yogically drawing out the necs tars via the n ds, through coordinated application of breath control and mantra.202 a. In this chapter, meditation (dhy na) is dened as the conjoining of bja-mantras a and n ds.203 While ight is mentioned among the siddhis acquired through such a. practices, it is by no means the exclusive aim, however. In the dichotomy of the cruel Kap lakundal and virtuous Saud min, Ronald a a .. a Davidson sees the contrast of the (Saiva) K p lika propensity for violence and a a the specically Buddhist contribution to extreme ascetic practice of restraint in the service of a moral direction. In his interpretation, Saud min is the rst Buda dhist siddha to be represented in non-Buddhist literature.204 A number of questions might be raised concerning this formulation, however. Saud mins involvement with a k p lika observances might indeed attest emergent Buddhist participation in this doa a main of extreme Saiva practice, for such begins to be described in eighth-century Tantric Buddhist sources. Yet the characterization of Saud min as Buddhist wara rants further consideration, all the more so with regard to some of the plays other major characters. Although one would expect the disciple of a Buddhist nun to follow her in religious orientation, the only testament of Saud mins devotion is proa vided when, while ying, she spots the naturally-formed image of Siva called SuvarYogins, of very little intellect, very little brains and little heroic spirit, [but] intent on serving their husbands, and possessing devotion to the gurustell me succinctly, O god, how they may have siddhi. Cf. BraY i.12cd13, and the annotation thereon. The implication appears to be that the pantheon a and practices of the kulacakra are specically for yogins, female practitioners emulating their ying exemplars. 201 Purity and Power, 213 (n. 89). 202 Yoginsac ra 5, especially from verse 41. a 203 5.37ab: dhy ne tu n dibj n m samyogo j napurvakah (but in the case of meditation, there is the a a. a a . a . . conjoining of the bja-mantras and n ds, which depends upon knowledge). a. 204 Indian Esoteric Buddhism. A Social History of the Tantric Movement, 203.

93 nabindu, to whom she offers homage and a verse of praise. It would thus appear that . her k p lika practices involve her in devotion to Siva. Of the young ladies and men a a who comprise the remaining castthe Buddhists Davidson speaks of Saud min a aiding205 little indication of religious afliation is in fact provided. K mandaks a acolytes Buddharaksit and Avalokit have, unsurprisingly, Buddhist names. Howa . a ever, the heroine M lat is said to visit the temple of Siva on the fourteenth day of the a waning moon; this she does escorted by none other than the Buddhist nun.206 Another reason why it might be unwarranted to read the contrast between Kap laa kundal and Saud min in sectarian terms is that the dichotomy of the pernicious, a .. a cruel sorceress and benevolent ying demi-goddess reects yogin typologies found throughout Saiva sources.207 Moreover, Bhavabhutis theme of the evil versus good witch might have been inspired by tales from the lost Brhatkath the source from a . which Bhavabhuti in fact drew the basic plot of his drama.208 If Davidson is nonetheless correct in seeing in Saud min evidence for Buddhist engagement in k p lika a a a practice, it should still be pointed out that his interpretation reects a one-sided, blood-and-power characterization of Saivism that little captures the complexities of this tradition, echoing dramatic, polemical, or exoticizing treatments of extreme Saiva practice in medieval literary texts. This attitude seems to extend even beyond Davidsons depiction of k p likas; note, for instance, his characterization of the medieval a a
Davidson remarks, She [Saud min] has gained the siddhis, most particularly that of ight (khecari a [sic]), and has come to assist the Buddhists in their struggle with the evil K p lika siddha, Aghoraghaa a nta, and his female companion, Kap lakundal . Ibid., 203. a .. .. a 206 Act Three, p. 103: ajja kasanacaddasi tti bhavade samam m lad samkaragharam gamissadi tado evam . . a . . . a kila sohaggam vaddhadi tti devad r hananimittam sahatthakusum vacam uddisia lavangi dudam m ladim aa . . a . a . .. . . . bhavad jevva kusum arujj nam anassadi tado annonnadamsanam bhodu tti (Since today is the fourteenth a a. . . .. .. . . . of the waning moon, M lat will go with Her reverence [K mandak] to the temple of Siva. Then, as it is a a believed conjugal fortune (saubh gya) thus increases, Her reverence shall herself bring M lat, attended a a by Lavangik , to the garden Treasury of Flowers, with the aim of [her] gathering owers with her own a hands for the purpose of worshiping the deities. Thus may [M lat and M dhava] behold each other). a a 207 See the critical editions and annotation of BraY lxxiii. a 208 See the introduction to M.R. Kales edition of the M latm dhava, 2427. On the theme of the a a good versus wicked witch, note in particular an episode in Kath sarits gara xii.1, a Sanskrit retelling a a of the Brhatkath , in which a womansecretly an evil sakintransforms her husband into a buffalo a . when he catches her in a compromising position with a buffalo herder. He is later rescued through the kindness of a benign yogin, who restores his form, provides him her daughter in marriage, arms him with magical mustard seeds to enact revenge, and instructs him nally in the vidy -mantra of a K lasamkarsin. I discuss related material in the subsequent section on the Brhatkath retellings. a a . . . .

94 representation of Siva: a killer divinity with a permanent erection.209 What I see this fascinating play providing evidence for is severalfold. First, the a practices described are of course those of the Mantram rga, and not of a non-tantric k p lika sect.210 This is evidenced in particular by Aghoraghantas stated engagea a .. ment in mantras dhana, and the above descriptions of the two yogins sources of a ritual power. Signicant parallels in descriptions of ritual and its aims, as well as the importance of C munda, leader of the Mother goddesses (m trn yik ), point morea a. a a . . over towards the k p lika- and goddess-oriented cults of Vidy ptha scriptures. Also a a a . present is a fundamental formulation of the Yogin cult: the possibility of women becoming ying Mistresses of Yoga through ritual perfection. That this tantric and k p lika cult had its centers of activity at sacred p. has such as Srparvata, was viewed a a t primarily as a domain of Saivism, and invited Buddhist participation, might also be suggested. Altogether, the evidence from the M latm dhava suggests that by the a a early eighth century, the Yogin cult described in bhairavatantras of the Vidy ptha a . was prevalent, corroborating references to this literature in the old Skandapur na. a.
the haravijaya of ratnakara

While the M latm dhava shows general awareness of the k p lika cult of yogins dea a a a scribed in Vidy ptha sources, the early ninth-century Haravijaya of Ratn kara,211 in a . a contrast, contains clear references to the Trikathe cult of the goddess triad Par , a Par par , and Apar probably in its developed Kaula variety. This Kashmirian a a a mah k vya echoes specic siddh ntatantras, as identied by Sanderson,212 and in a a a a hymn to the goddess Cand, displays technical knowledge of the Trikas system of . . deity visualization.213 This hymn makes passing reference to yogins as well, describ
Indian Esoteric Buddhism, 90. In contrast, note that the Mattavil sa, a seventh-century drama which parodies the antinomian and a k p lika activities of a Saiva ascetic, provides no conclusive indications of Mantram rga practices. a a a 211 Concerning the date of composition, Sanderson suggests around 830, based upon his assessment of the dates of Cippatajay pda, the Kashmiri monarch in whose court Ratn kara apparently wrote the a . a . Haravijaya. History through Textual Criticism, 6 (n. 3). 212 History through Textual Criticism, 56. 213 See Sanderson, Mandala and Agamic Identity in the Trika of Kashmir, 169 (n. 1); and History .. through Textual Criticism, 1819 (n. 21). See also David Smith, Ratn karas Haravijaya: An Introduction a
210 209

95 ing the Goddess in union with Bhairava as the radiant nave of a cakra of yogevars, i.e. s yogins, upon whom one meditates to attain identity with Siva.214 Also noteworthy is the fact that yogins, in contrast to the Seven Mothers, have not been incorporated into the rich Saiva mythology of this work to any signicant degree: they gure only in the above reference to esoteric ritual.215 In marked contrast, the thirteenth-century Haracaritacint mani of R j naka Jayadratha, also a Kashmiri, prominently integrates a . aa yogins into Saiva mythology.216
the brhatkatha and its retellings .

The various retellings of the lost Prakrit Brhatkath of Gunadhya, The Grand Tale, a .. . offer a wealth of material pertinent to Tantric Saivism and the Yogin cult. The Kath sarits gara, an early eleventh-century Sanskrit version by the Kashmiri author a a Somadevabhatta, is permeated with references to tantric practices and practitioners .. and contains vivid accounts of yogins. Closely parallel to this is the mid eleventh century Brhatkath majar of Ksemendra, this too of Kashmiri provenance. An earlier a . . Sanskrit version, the Brhatkath slokasamgraha, also contains much that is relevant to a . . the study of Tantra; but there are telling contrasts in these texts, probably separated by a century or more in time. Reaching further back to the Vasudevahindi of Sa.. nghad saganin, an early Jaina retelling in Prakrit, evidence for Tantra and the yogin a . cult recedes from view.217 In the Vasudevahindi, The Odysseys of Vasudeva, we do nonetheless glimpse .. proto-tantric ideas and practices of some relevance to the formation of the Yogin cult, especially its tales of the vijj haras (i.e. vidy dhara) and their lores (vijj , i.e. a a a
to the Sanskrit Court Epic, 26366. 214 Haravijaya 4.28; this has been quoted and translated by Smith, Ratn karas Haravijaya, 262. a 215 I base my knowledge of the deities and myths in the Haravijaya largely on the studies by Smith (Ibid., 22576), and Santosh Kumari Sharma, Haravijaya of Ratn kara: a Criticism, 268316. a 216 An episode prominently featuring yogins is described in Sanderson, Religion and the State, 28586. 217 For an overview of the Brhatkath and a thesis on the relations between its various retellings, see a . Donald A. Nelson, The Brhatkath : a Reconstruction from the Brhatkath slokasamgrha, Perunkatai and a a . . . Vasudevahindi; and Brhatkath Studies: the Problem of an Ur-text, Journal of Asian Studies 37:4 (1978), a .. . 663676.

96 vidy ).218 While of uncertain dating, the text is written in a Prakrit Ludwig Alsa dorf argues belongs to the early centuries of the common era.219 The Vasudevahindi .. therefore comprises the earliest surviving retelling of the Brhatkath , though owing a . to its adaptation to Jaina universal history, not the most faithful.220 While yogins have no role in this tale, another category of divine woman does have a signicant position: the vidy dhar, female counterpart of the vidy dhara or lore-holder. It is a a these lores (vidy ) and their semi-divine masters, both male and female, that have a signicance for the present study. In Donald Nelsons assessment, the core narrative of the original Brhatkath had a . three concerns: the hero Narav hanadattas gradual acquisition of wives, vidy s, and a a overlordship of the vidy dharas. These also form an integral sequence, for vidy s must a a be mastered in order to conquer vidy dharas, and Narav hanadatta learns many of a a these from his vidy dhar brides.221 On the treatment of vidy s in the Vasudevahia a ndi, I take as example an episode from book fourteen, which tells the tale of the .. heros marriage to the vidy dhar princess Mayanaveg .222 In an aerial scufe with a a . his vidy dhara abductor, the hero, Vasudeva, falls from the sky and nds himself in a a the River Gang . He encounters a person endeavoring to master a vidy , standing a in the water in the garb of a mendicant ascetic, who says, on seeing you, my vidy a has become perfected (siddh ) . . . tell me, what token of friendship might I offer you? a I am a vidy dhara.223 Vasudeva requests to learn a vidy that bestows the power of a a
Sanghad saganins text has been published as Vasudevahindiprathamakhandam, 2 vols., edited by a . .. .. Caturavijaya and Punyavijaya[-muni?] (Kathiawar: Shri Jain Atmanand Sabha, 1930). On vidy s in the a . Vasudevahindi, see A. P. Jamkhedkars useful appendix, Vasudevahimd: A Cultural Study, 22535. Note .. . . also Jagdish Chandra Jain, Vidy dharas in the Vasudevahindi, Journal of the Oriental Institute of Baroda a .. 24 (1974?): 120127. 219 Alsdorf argues that the text is written in an archaic form of Jaina Mah r str dating to a period a a. . centuries prior to the sixth century c.e., possessing linguistic features that bear comparison with the Ardham gadh of the early Jaina canon. Alsdorf, The Vasudevahindi, a Specimen of Archaic Jainaa .. Mah r str, Bulletin of the School of Oriental Studies 8:23 (193537): 31933 (quote on p. 320). a a . 220 Nelson, Brhatkath : a Reconstruction, 256. a . 221 Nelson, Brhatkath : A Reconstruction, 28283, 294, 30506. a . 222 This chapter was brought to my attention by Nelson, 22223, and Jamkhekdar, Vasudevahimd: A . . Cultural Study, 228. Here and elsewhere, I have consulted the comparative summary/translation of the Vasudevahindi and Brhatkath slokasamgraha by Jagdish Chandra Jain, The Vasudevahindi: An Authentic Jain a .. . . .. Version of the Brhatkath . a . 223 Vasudevahindi 14 (vol. 2), p. 229: siddh me vijj tumha damsanena . . . samdisaha, kim v payacch mi a a a .. . . . . . a

97 ight. After initiating him, the vidy dhara instructs Vasudeva in this lore. He follows a the prescribed ritual for a day; in the evening a divine woman manifests before him, and with his assent, carries him off into the sky.224 The method and terminology for vidy -practice in the Vasudevahindi bears rea .. markable continuity with later tantric ritual. First and foremost, vidy s are not only a lores, in the sense of spells to be mastered that accomplish specic aims, but also deities to be propitiated. Thus when the divine woman manifests before Vasudeva, he wonders whether she is the vidy -goddess herself, pleased with his worship.225 a This shows clear similarity with the tantric conception of vidy s as goddesses, female a mantra-deities. Furthermore, access to the vidy requires initiation, for which the tera minology is identical to that of later Saivism: Vasudeva becomes dikkhio, i.e. dksitah, . . initiated; the rite is unfortunately not described. Just like a tantric mantra-deity, a vidy bears fruit when it becomes perfected or mastered (siddh ).226 This initial a a process of making the mantra effective is, in our example, described as puraccarana . (Sanskrit puracarana), prelimary service/propitiation, a term having similar means . ing in tantric ritual. As in tantric puracarana, this involves primarily incantation, s . alongside other rites and observances.227 For example, in order to master the vidy s a called sumbha and nisumbha, Vasudeva must make food offerings to deities (bali) on . . a mountain peak on the fourteenth night of the waning moon, and repeat the vijj a one-thousand and eight times. Here, signicantly, no distinction is made between the
ptid nam? aham vijj haro. a. . a . 224 Ibid., 22930; cf. Jagdish Chandra Jain, Vasudevahindi, 43537. .. 225 Vasudeva at rst wonders whether the divine woman is an obstruction personied (viggha, i.e. vighna) taking the form of a beautiful temptress, but then decides she is the goddess of the vidy a (vijj bhagavat), pleased by his preliminary worship (puraccaranatosiy , i.e. puracaranatosit ). It bea a s a . . . comes clear in the course of the story that she is rather Mayanaveg , a vidy dhar princessthe sister of a a . Dandavega, the vidy dhara Vasudeva encountered upon falling from the sky. Vasudevahindi 14 (vol. 2), a .. .. p. 230. 226 Hence, the vidy dhara who teaches Vasudeva the lore for magical ight states, ahor te ate aham te a a . samvam essam ti, punnapuraccaranassa ya vijj siddh bhavissa tti na samdeho so gato (He said, When a . a a . .. . . day and a night have passed, I will come to you, and you will master the vidy when the preliminary a worship is complete; of this, there is no doubt, then left). Vasudevahindi 14 (vol. 2), p. 22930. .. 227 On the subject of puracarana or purvasev , the preliminary ritual for making a mantra effective s a . (siddha), see Gudrun Bhnemann, On Pura carana: Kul rnavatantra, Chapter 15, in Teun Goudriaan s a . . (ed.), Ritual and Speculation in Early Tantrism. Studies in Honor of Andr Padoux, 6164 (quote on p. 61); and Sanderson, Religion and the State: Initiating the Monarch (forthcoming), n. 355.

98 vidy and its mantra: the vidy is the mantra to be perfected, upon which it grants its a a powers. In this case, however, the two vidy s are not explicitly deiedalthough they a might conceivably embody their namesake demons: the Sumbha and Ni umbha of s Devm h tmya fame.228 In another story, a vidy dhara who had lost his powers is said a a a to practice the s hana (Sanskrit s dhana) of a vidy , this being in Tantric Saivism the a . a a term for the observances one undertakes to achieve power (siddhi) with a particular mantra or vidy .229 It is unclear whether in the Vasudevahindi this is distinct from a .. puracarana. s . Further investigation is required into early Indian conceptions of vidy s, for they a seem highly relevant to the development of Tantra and the Yogin cult. In the Vasudevahindis vidy s, it would seem that we nd, already, a conation of mantras, a .. female deities, and specic magical powers, access to which requires initiation. In other words, by the term vidy /vijj is connoted a specic mantra, the power rea a sulting from its mastery, and the goddess who embodies it, who must be won over through puracarana and/or s dhana. This nexus of ideas has clear continuity with s a . later, self-consciously tantric conceptions of mantra and ritual. The idea of vidy s a as goddesses takes on extraordinary signicance in the later tradition, for the earli est literature of the Saiva Yogin cult identies itself as belonging to the Vidy ptha a . division of the Saiva canon: The Seat of Female Mantra-deities. This level of the tradition is distinguished by its emphasis on practices for attaining power, siddhi, and upon vidy s or female mantra deitiesmyriad manifestations of Sivas primordial ena ergy or power, sakti. Through the conception of this sakti as feminine, goddesses and
Vasudevahindi 6 (vol. 1), p. 195: aham vijj haro, atthi me duve vijj o suhas hanao sumbhanisumbh o a a a . .. . . . a uppayanippayano, t o tava demi. tumam si t sim bh yanam. jam puna balivih nam tam aham savvam uvanemi. a a . a . . a. . . . . . . . . . tumam k lacaddase eg g mamam milasu. at. asahass vattiy ya te vijj sijjhihiti tti [=sijjhahi tti?]: I am a a a a a . a . .t vidy dhara, and I have two vidy s easily mastered, Sumbha and Ni umbha, for ying up and coming a a s down; Ill give them to youyou are a t recipient for these. I will on my part take care of all the bali food offerings. Meet with me alone on the fourteenth of the waning moon, and when the vidy has been a repeated one-thousand and eight times, you will master it. My attention was directed to this passage by Jamkhedkar, Vasudevahimd: A Cultural Study, 228. . . 229 a Vasudevahindi 4 (vol. 1), p. 176: esa amg rao vijj bhat. o s hanam kunati vijj e (this is Ang raka, a .t a . . a .. . a . who has lost his lore and [hence again] does s dhana of the vidy ). Cf. the description of the vidy s a a a sumbha and nisumbha as suhas hana, the s dhana of which is easy, i.e. easily mastered, quoted in the a . a . . previous note.

99 power become consubstantiala connection absent in proto-tantric conceptions of vidy s. a It would appear that siddhi-oriented forms of tantric practice, which culminate in the cult of yogins, represent the continuity of proto-tantric vidy practices integrated a within Saiva and Buddhist soteriologial systems. One of the most distinctive charac teristics of Tantric Saivism, as well as Buddhist Tantra, is indeed the claim of efcacy for both soteriological ends and the pursuit of power. This pursuit of power was, in the broadest sense, a quest for superhuman agency: to embody the powers of deities. One important locus for such conceptions is the gure of the vidy dhara, the india vidual whose mastery of vidy s affords transcendence of the human condition. This a transcendence is expressed through abilities such as ight, and entry into paradisal realms neither of this world nor, strictly speaking, beyond it; one may enter into the community of vidy dharas. a Attainment of vidy dharapada, the status of a vidy dhara, was in fact the aim of a a much of the non-soteriologically oriented ritual outlined in early tantric literature, from the Saiva Niv sa corpus to early Tantric Buddhist texts.230 Recall that this s a was, for instance, the aim of the zombie rite (vet las dhana) in which the king a a Puspabhuti assisted Bhairav c rya, as portrayed in B nas early seventh-century a a a. . Harsacarita. Writing in 673 c.e., a Chinese monk in India named Yi-jing refers to . the emergent corpus of Buddhist tantric literature as the Vidy dharapitaka, the canon a . of the vidy dharas.231 This concern with vidy s and becoming a vidy dhara must a a a also be seen as a broader theme in ancient Indian folklore, apparently predating the early medieval formation of tantric sects. Nowhere is this more evident than in the retellings of the Brhatkath the narrative of an individuals journey from human a . prince to emperor of the vidy dharas. With the development of the cult of yogins, the a
230 Becoming a vidy dhara appears as the goal of many of the siddhi-oriented practices outlined in a the Guhyasutra of the Niv sa, this being mentioned a dozen times. Note, for instance, 11.85: om s a . paramevarapar ya namah | anena mantrena parvataikharam aruhya bhiksah ro daalaksani japet | vidy dharo s a s s a . . . a .. bhavati (om parame varaparaya namahafter ascending the peak of a mountain, one should s . . incant one million times with this mantra. One becomes a vidy dhara). a 231 Stephen Hodge, ed., The Mah -Vairocana-Abhisambodhi Tantra: with Buddhaguhyas Commentary, 10. a

100 gure of the vidy dhara/ recedes somewhat into the background in Tantric Saivism, a while the quest to embody divine powers takes on new forms. The gure of the yogin is itself among the foremost of these expressions, for as is the case with the vidy dhar, a woman may aspire to become one. a Though written perhaps half a millenium after the Vasudevahindi,232 the Sanskrit .. Brhatkath slokasamgraha still contains scant references to yogins. A scene in chapter a . . twenty describes a cremation ground, where at night a variety of tantric practitioners congregate: In one place I saw a circle of witches [dakinmandala] with upraised arms . .. and owing hair dancing naked around a corpse. In another place I saw a man brandishing a sword and holding a skull that he was using as a bowl. He was saying, Mighty ones, buy some human esh! Then, guarded in each of the cardinal directions by a troop of armed men, there was an aspiring magician [s dhaka] with a magical sword who was oating up a 233 into the sky. Here we see depictions of tantric cremation-ground ritual familiar from other literary sources: the sale of human esh (mah m msavikraya), as depicted in the M latm dhava, a a. a a and a rite parallel to the vet las dhana Bhairav c rya undertakes under armed guard a a a a in the Harsacarita. However, there is also described a coven exclusively of women, . dakins, engaged in a wild rite with a corpse. Most interesting of all is the subsequent . depiction of Dhanamat, a vidy dhar queen who has taken the guise of an elderly a tribal woman (m tangavrddh ). There in the cremation ground at the root of a banyan a a .
The dating of the Brhatkath slokasamgraha of Budhasv min is problematic. Its editor, Flix Lacte, a a . . places the text in the eight or ninth century c.e. Lacte, Essai sur Gunadhya et la Brhatkath . Suivi du Texte a . . . Indit des Chapitres xxvii xxx du Nep la-M h tmya, 147; cited by Nelson, Brhatkath : a Reconstruction, a a a a . 5. Nelson, without explaining his reasons, sees the text as a composition of the late Gupta period; ibid. The texts most recent editor and translator, James Mallinson, remarks more cautiously that we can only be sure of a pre eleventh-century dating. The Emperor of Sorcerors, vol. 1, 13. 233 Brhatkath slokasamgraha 20.9496: a . . ujjhit mbaram udb hu prakrnakacasamcayam | a a . . .. paritah kunapam nrtyad dakinmandalam kvacit 94 . . . . . . . kvacit purusam utkhadgam up ttaghatakarparam | a . . . . mah m msam mah sattv h kryat m iti v dinam 95 a a. . a a. a a saastrapurusavr taraksit sacatus. ayam | s . a . a .t s dhakam siddhinistrimsam utpatantam nabhah kvacit 96 a . . . . Text and translation as given by Mallinson, Emperor of Sorcerors, vol. 2, 18687.

101 tree she performs re sacrice, offering human blood into a charnel re with a ladle held in her left hand, incanting a mantra ending with the syllable ham.234 Animat. ing a corpse to do her bidding, she uses this rite of subjugation (vakarana) to bring s . the hero under her power and marry him to her beautiful granddaughter. He offers scant protest. In the parallel episode of the Vasudevahindi, chapter four, the disguised vidy dhar a .. is said merely to recite a mantra in the cremation ground, by which she causes an animated corpse (vey la, i.e. vet la) to kidnap the hero.235 It thus appears that a a a reference in the original to a vidy dhar practicing her art in the cremation ground a has in the Brhatkath slokasamgraha been eshed out with reference to later tantric a . . ritual. That is, vidy s and their vidy dhara/ masters are to some extent recast in the a a image of contemporaneous tantric ritual and ritualists. Besides the reference above to dakins, the gure of the yogin does not, however, feature with any prominence. . In contrast, the Kath sarits gara affords a fascinating view of yogins from a region a a and period in which their cult was highly prominenteleventh-century Kashmir. In a this text, yogins take on much of the imagery and roles of the vidy dhar from earlier retellings of the Brhatkath .236 In the Kath sarits gara, we encounter yogins as both a a a . divine and corporeal women: powerful and nameless groups of ying goddesses, as well as remarkably human characters, between which, perhaps, little distinction was intended.237 Those depicted as actual women range from impetuous, even vile

Brhatkath slokasamgraha 20.9798: a . . ity dibahuvrtt ntam payat pretaketakam | a a s a . . y tr m y gacchat drs. a s drs. a sthavir may 97 a a. a a . . t a . . t a a vatamule cit vahnau v mahast rpitasruv | a a a a . hamk r ntena mantrena juhvat naraonitam 98 s . . aa .

235 Vasudevahindi 4 (vol. 1), p. 17879, especially p. 179: . . . samp vio piuvanam | dit. h ya may a .. . a . .t a m yamgavuddh kim pi jampamt | (I reached a cremation ground, and saw the elderly tribal woman a . .. a . . . incanting some powerful [mantra]). 236 Though omitted from the present discussion, the same may be said of the Brhatkath majar, a text a . closely parallel to the Kath sarits gara written not long after it in Kashmir. a a 237 I am grateful to Isaacson for rst directing my attention to accounts of yogins in the Kath sarits gara, beginning with a Sanskrit reading seminar in the autumn of 2002. My awareness a a of the material has also benetted from a stimulating article by Adelheid Herrmann-Pfandt, which draws on stories of yogins from the Kath sarits gara: The Good Womans Shadow: Some Aspects of a a a the Dark Nature of Dakins and S kins in Hinduism, in Cornelia Vogelsanger and Anette Wilke, eds., .

102 witches to virtuous and accomplished tantric adepts. This spectrum of characters accords with taxonomies of yogins from tantric literature: in the Kath sarits gara, a a yogins are classied, variously, as dakins or sakinslowly and cruel varieties . while yogins without such qualiers are frequently benevolent. The yogins of the Kath sarits gara form a colorful assortment. In book six we a a meet one named Citralekh , whose prowess in ight facilitates the union of her a friend, the princess Usa, with Aniruddha of Dv ravat.238 Another well-meaning a . yogin instructs her friend in mantras for turning her illicit lover into a monkey, and for restoring her pet to human form on demand.239 In contrast, note for instance the dakin K lar tri, the grotesque and lusty wife of an orthodox brahmin . a a teacher (up dhy ya) who possesses the power of ight through mantra-practice and a a consumption of human esh. She secretly acts as guru to a coven of dakins, which . a future queen joins with disastrous consequences.240 Another story tells of a weary traveller who unknowingly accepts the hospitality of a sakin. He thwarts her at tempt to use enchanted barley to turn him into a goat, but ends up being turned into a peacock by the butchers wife, a wicked (dus. a) yogin.241 In another episode of .t book six, we nd not yogins, but false accusations: a jealous queen, a greedy female renunciant, and clever barber conspire to make the king think his newest bride is secretly a dakin, who sucks out his vitals whilst he sleeps.242 A different sort of yo. gin is represented by Sarabh nan , whose name and description are suggestive of a a a theriomorphic goddess rather than human sorceress.243 A number of tales in the Kath sarits gara pit benevolent and malec yogins a a against each other, much as does the Brhatkath -inspired M latm dhava. Book seven a a a . includes the story of a certain Bhava arman of V r nas, who had an affair with a s a a.
Wild Goddess in India and Nepal, 3970. 238 Kath sarits gara vi.5.136. a a 239 Ibid., vii.10718. 240 Ibid., iii.6.102218. 241 Ibid., xii.4.26377. 242 Ibid., vi.6, especially 15380. 243 Ibid., viii.5.123134. This name means Sarabha-face, the sarabha being an exotic, mythical beast.

103 ckle brahmin woman, Somad , a secret yogin (guptayogin, 150d) of the worst a sorta petty sakin (ksudraakin, 168b). One day, in a t of jealous anger he beats s . her. Hiding her rage, the next day she slips his sacred thread around his neck and turns him into an ox. Sold as a beast of burden, another yogin, Bandhamocin, spots him and restores him to human form. Transforming themselves into horses, the two yogins duel, and Bhava arman and Bandhamocan fell Somad .244 In an s a s other, parallel episode, a certain V madatta discovers that his wife, Sa iprabh , is a a secretly both an adultress and a sakin. Caught in the act with a herdsman, she turns her enraged husband into a buffalo, beats him, and sells him off. A perfected yogin, however, recognizes him in animal form and restores his humanness. She gives him her own daughter in marriage, supplies him with enchanted mustard seeds to enact revenge, and in the end, initiates him and his wife into the vidy of the goda dess K lasamkarsan. Upon the sacred mountain Srparvata, their mantra-practice a . . . achieves fruition, and K lasamkarsan herself manifests and bestows a magic sword. a . . . They become vidy dharas.245 a In the story of Kandarpa of Ratnapura, in Kath sarits gara xviii.4, we encounter a a yogins as groups of powerful, ying beings. One day, when at dusk the brahmin Kandarpa goes to fetch water from the river Vena, the current sweeps him away. He . nearly drowns, awakening on the riverbank near a deserted temple of the Mother goddesses. Entering the temple, he prays to the Mothers for succor. At night, from among the Mothers emerge a group of ying yogins who take Kandarpa under their protection. Later, while in ight, another group of yogins accost them, desiring to harm Kandarpa, and in the scufe he falls to the ground, lost.246 This has some
This episode occurs as vii.3.14769. Ibid., xii.1.3172. 246 There are considerable inaccuracies in David Whites summary of this episode (Kath sarits gara a a 18.4.20422). In his words,
245 244

A brahmin named Kandarpa from Ratnapura comes upon a deserted Mother goddess temple (unya m trdevagrha) in the night. Entering, he sees a brilliant light. He prays to s a. . the Mothers to protect him. When the daylight comes, he nds garlands of bones and the skulls of children. He realizes they are from a host of Mothers [i.e., witches]. He later hears the group of Yogins speaking amongst themselves: Today we must go to the gathering of the circle (cakramelaka) that is taking place in Cakrapura. The Yogins

104 parallels in a story from book fourteen.247 After failing in his studies, a certain N gasv min sets off from P talputra to see the goddess Vindhyav sin. In a town a a a. a along the way, a housewife gives him a red lotus along with the alms he seeks. The lady of another house sees him, and exclaims, O alas! You have been selected by a yogin. Under the pretext of a red ower, she has given you a human hand look!248 And so it was. The housewife sends him to a brown cow (kapil ) in the a home of a certain Devaraksita; nestled between her hooves, the divine cow affords . him protection through the night when the yogin and her coven descend upon him, seeking esh and blood. Next he seeks the aid of a great P supata, but he too a can keep the yogins at bay only temporarily, and sends him off to someone else. Along the way, the yogins catch N gasv min and whisk him off into the sky. Like a a Kandarpa, he falls to the ground when there ensues an aerial battle with a rival group. He lands in a deserted place near a temple, where dwells a beautiful yaksin, . . Sumitr , cursed to dwell among mortals. She takes him as lover for the duration of a her curse. Looking at the development of the Brhatkath corpus, it appears that the yogin a . becomes an increasingly important locus for tales of magical women, encroaching
nd him hiding there, and carry him off with them. . . One of their number, named Sumanas, marries him. Leaving the circular gathering of the Yogins (yoginyacakramelaka), s she carries him away with her up into the sky. Kiss of the Yogin, 213. Compare with the sound translation of Charles Tawney, The Ocean of Story, vol. ix, 5761. In fact, having been swept away by a river, Kandarpa comes upon the Mother temple at dawn (din game), and overcome by exhaustion, sleeps there into the night (20510). No reference is made to a garlands of bones and the skulls of children; the night is described as a female ascetic (rajanit pas) a who is white with the ashes of moonlight, who wears a necklace of bones that are the stars, and whose gleaming skull-bowl is the moon ( g t t r sthim l dhy jyotsn bhutisit tatah | saiubhrakap l ca a a aa a a. a a a aa . ss raudr rajanit pas, 211cd; perhaps White nds siu, child, in saiubhra?). In order not to leave him a s ss unprotected in the wild, the assembled yogins decide to take Kandarpa along for their airborne night journey, and drop him off at the home of a br hmana. It happens that the gentlemans daughters a . auspicious period for marriage (lagna) had arrived, but no suitable groom. Kandarpa is promptly married to Sumanas, and we are given no indication she was but an ordinary young woman (21320). Late in the night (pacime y me), the yogins return from their ritual gathering (yoginya cakramelak t s a s a agatya; White appears to understand yoginyacakramelak t as a compound, or else the nominative plural s a yoginya as genitive plural). They carry Kandarpa back up with them into the sky, from which he s is dropped when a battle ensues with a group of yogins desiring to do him violence ( jihrsubhih) . . (22122). 247 Kath sarits gara xiv.4.2061. a a 248 Ibid., xiv.4.25d26a: tatraty gehin h dhig yoginy svkrto bhav n paya datto nrhastas te rakt bjaa a a . a s a . vy jato nay . a a

105 upon the territory of the vidy dhar especially. A number of conceptual continuities a link the vidy dhar and yogin, above all their command of mantras, powers of ight a and shapeshifting,249 and their variegated transactions with mortals, for whom they present both danger and access to powers. We nd, for instance, the vidy dhar Hirana . namat of the Vasudevahindi recast in the Brhatkath slokasamgraha as a tantric sorceress, a . .. . . Dhanamat, while in the Kath sarits gara, the vidy dhar Bhadr is also referred to as a a a a a yogevar.250 Note also how in the Kath sarits gara story of V madatta, a perfected s a a a yogin (siddhayogin) takes on roles played by the vidy dh r in earlier retellings: the a a yogin bestows her own daughter in marriage to the hero, as did for instance Hiran . namat/Dhanamat, and she facilitates the heros mastery of vidy s, just as do Hira a . nnamats daughter and other vidy dhars. In this case the vidy is the mantra of a a .. K lasamkarsan, a form of K l important in the Krama cult of contemporaneous a a . . . Kashmir.251 Its fruit, however, remains that which the original Brhatkath would have a . attributed to mastery of vidy s: V madatta and his wife become vidy dharas.252 a a a While the Kath sarits garas yogins have direct continuity with the earlier gure a a of the vidy dhar, we also nd yaksins, a class of demi-goddess with strong conneca . . tions to trees and the natural world, recast as yogins. Note that the stories of Kan darpa and V madatta, summarized above, both involve the theme of yogins dropa ping a man to the ground when they are accosted in the sky by rival groups. This appears to recast tales of yaksins from earlier Brhatkath versions: in the Vasudevahia . . . ndi, in the beginnings of chapters ve and fourteen, the hero fabricates stories of .. having been dropped from the sky when there ensued a scufe over him between two yaksins.253 This is of particular interest, for it illustrates the breadth of the cate. .
249 While both share the ability to change forms, yogins more typically take on animal form, as discussed in the introduction to this dissertation. Cf. Kaulaj nanirnaya 23, quoted later in this chapter a . (n. 309). Vidy dhars, on the other hand, more often transform themselves by taking on human guise; a note e.g. the case of Dhanamat/Hirannamat, mentioned above, who along with her daughter takes on .. the form of a tribal woman (m tang). a 250 Ibid., iii.4.378. 251 On K l as K lasamkarsan, see Sanderson, Mandala and Agamic Identity in the Trika of Kasha a . .. . . mir, 188204. 252 Kath sarits gara xii.1.6468. a a 253 For example, in chapter fourteen, after being dropped by the wicked vidy dhara M nasavega, the a a

106 gory yogin, which could encompass earthly women of power as well as goddesses. This also reinforces the degree to which multiplicity was integral to conceptions of yogins: pairs of rival yaksins become cast as feuding groups of yogins. . .
the ya astilaka of somadevasuri s

Beyond the ninth century, references to yogins become common in literary sources, and a comprehensive review is beyond the present study. One work meriting par ticular mention is the Yaastilaka, composed by a Jaina author, Somadevasuri, in the s Saka year 881 (i.e. 959 c.e.).254 The Yaastilakas frame story concerns the battle-lusty s and passionate M radatta, a king who, advised by a Kaula guru (kul c rya) named a a a Vrabhairava, plans a massive sacrice to the goddess Candam r. Through this he a .. seeks to attain a magical sword and thereby become lord of the vidy dharas.255 On a the appointed day, a veritable zoo of sacricial animals is assembled, while the king himself is to slay two human victims. Unbeknownst to the monarch, those brought before him in the temple are his own nephew and niece, who had left home as children to become Jaina ascetics. Chastened by their tranquil presence, the sovereign sets down his sword and listens to his nephew narrate the life of king Ya odhara s the core narrative of the Yaastilaka. In the end, M radatta abandons his violent ways s a and becomes a renunciant. With its theme of a narrowly-averted human sacrice to
a hero says, aham jakkhinhim dohim maham tumam ti bhamdamthim padimukko ag se iham padio (I fell here, . . . . . . .. . . . . . released into the sky by two yaksins threatening youre mine! ). Vol. 2, p. 229. . . 254 I read primarily from the editio princeps: Yaastilaka of Somadev Suri, with the commentary Candrik of s a a Srutadeva Suri, ed. Pandit Sivadatta and V sudeva Laxman S str Pana ikar (2 vols.). The more recent a .. . s edition is that of Sundaral la S str, Srmatsomadevasuri-viracitam Yaastilakacampu Mah k vyam. This a a s a a . work has been the subject of a rich and useful study by Krishna Kanta Handiqui, Yaastilaka and Indian s Culture, or Somadevas Yaastilaka and Aspects of Jainism and Indian Thought and Culture in the Tenth Century. s On the dating of the text, see ibid., 2. 255 sa punar ekad nrpatir atmar jadh ny m eva candam ridevat y h puratah sakalasattvopasamh r t svayam a . a a a a a. .. a . . aa . ca sakalalaksanopapannamanusyamithunavadh d vidy dharalokavijayinah karav lasya siddhir bhavatti vraa a a . . . . bhairavan mak t kul c ryak d uparutya khecarlokalocan valokanakutuhalitacet s tathaiva pratipannatada a a a a s a a a ar dhanavidhih . . . (Once, the king heard from a kula-master named Vrabhairava that by sacricing . all [sorts of] creatures before the goddess Candam r in ones own capital city, and oneself slaying a a .. pair of humans endowed with all the auspicious marks, one gains the power of a magical sword that gives victory over the vidy dharas. His heart made eager to gaze into the eyes of the ying [vidy dhara-] a a women, and having learned the procedure for worshipping her [Candam r], . . . ). Yaastilaka i, vol. 1, a s .. p. 44.

107 the Goddess, this work hence appears to echo the M latm dhava or the latters own a a sources. References to Saivism in the Yaastilaka attest to Kaula goddess cults in the tenth s century, a period from which in any case there survives abundant Kaula literature.256 Unsurprisingly, the texts Jaina author provides a polemical characterization of the Kaulawhat he refers to as the teachings of the clan masters (kul c rya) a a connecting it intrinsically with blood sacrice and the consumption of forbidden substances. Somadevasuri also mentions a specic Kaula sect, referring twice to the Trika system (trikamata)the important cult of the goddesses Par , Par par , and a a a Apar .257 The text also alludes to practices connected with yogins: a spy of the a king Ya odhara disguises himself as a colorful and seedy Saiva, an antinomian siddha s who mimics Siva in his form as the tribal huntsman (kir ta), and who, on account a of communion with yogins, is advertized as possessing astonishing knowledge and powers.258 His communion with yogins (mah yoginsangati) undoubtedly signies a
256 Handiqui identies and discusses numerous references to Saivism in the Yaastilaka, which concern s P supatas, the Saivasiddh nta, and the Kaula. Yaastilaka and Indian Culture, 199219, 22429, and 354 a a s 60. 257 One reference to the trikamata occurs in a brief doxography of views on liberation: sarvesu pey peyaa . bhaksy bhakysadisu nihsankacitt d vrtt d iti kul c ryak h | tath ca trikamatoktih madir modameduravadanaa . a a a a. a a . a . . . . starasarasaprasannahrdayah savyap rvaviniveitaaktih saktimudr sanadharah svayam um mahevar yam as s s a a s a a . . . . nah krsnay sarv nsvaram [em. (silent) Handiqui, Yaastilaka and Indian Culture, 204 (n. 3); sarv n Ed.] s a. . . . . . a a . a ar dhayed iti (According to the kula-gurus, [liberation transpires] due to acting with a mind free of inhibition with regard to all that is considered t or unt for drinking, and t or unt for eating, etc. Hence the saying of the Trika system, having ones heart () . . . impassioned and gladdened on account of (?) a mouth dense with the fragrance of wine, having ones female ritual partner seated on the left side, holding the posture of the saktimudr , themselves imitating Um and Siva, one should worship a a the Lord of Sarv n [i.e. Bhairava] together with the Dark Goddess [krsnay ; or, with dark spirits?]. a. a .. . Yaastilaka vi, vol. 2, p. 269. The text appears suspect at the point vadanastarasarasa ; perhaps emend s vadanas taralasarasa ? The possibility that krsna could refer to K l/Candik was suggested to me a .. a .. . by Isaacson (personal communication, March 2007). The variant reading reported for krsnay in the .. . a edition of Sundaral la S str is madiray (with wine), which could suggest understanding krsna as a a a .. . a dark liquor. S str (ed.), Srmatsomadevasuri-viracitam Yaastilakacampu Mah k vyam, 184. Another s a a . passage in Yaastilaka i refers to initiation into the Trika system: sakalajanas dh rane pi svadehe trikas a a . matadksitasyeva devabhuyen bhiniviam nasya ([the king M radatta] who, as though initated into the . a s a a Trika system, was obsessed with [the idea] that his body is divine, even though it is like everybodys). Vol. 1, p. 43. Sanderson points out that the Yaastilaka thus provides evidence for the presence of the s Kaula Trika in the Deccan in the mid-tenth centuryone of several indications that the Trika was by no means a specically Kashmirian tradition. A Commentary on the Opening Verses of the Tantras ra of a Abhinavagupta, 13233. 258 Yaastilaka iii, vol. 1, p. 399400: . . . pracurapratikarmavikrtag traih sattriputrair dandajinikai ca paris s . a . .. vr jakaih esa khalu bhagav n samj tamah yoginsamgatir atndriyaj nodgatih siddhah s medhikah samvananaa a a . a . . . a . . a . . karmana karina kesarinam api samgamayati vidvesabhesajena jananm apy atmajesu vairinm vidadh tty a . . . . . . . ..

108 yoginmel pa, the union or encounter with deities texts of the Yogin cult describe a as among the foremost attainments of ritual, effecting access to the highest powers. Candam r (Grim Destroyer), the goddess to whom M radatta intended to ofa a .. fer sacrice, presides over a temple called Mah bhairava in the royal capital.259 The a goddess association with this temple and her vivid k p lika iconography link her a a to Bhairav, the goddess consort of Bhairava and primary locus of the concept of the Mah dev and supreme Sakti (par sakti) in the bhairavatantras. Candam rs a a a .. name and description also place her within the broader fearsome goddess typology epitomized by Candik /Candthe primary name with which B na addresses a. .. a . . the Great Goddess in his panegyric, the Candsataka.260 While the name Candam r a .. . . seems unattested in tantric sources, a goddess M r does appear to have had ima portance. The BraY mentions texts by the names M r and Mah m r, presumably a a a a centered upon this goddess cult, while the Tantrasadbh va contains some material a concerning this deity.261 Pur nic sources too refer to Mah m r; in particular, the a. a a Agnipur na has a chapter devoted to her vidy -mantra,262 while the Devm h tmya a. a a a provides mah m r as an epithet of the Goddess as universal destroyer.263 a a
avedyam naj namantratantraprabh vah (He has his powers of [occult] knowledge, spells, and mantras a a a . announced by () sattriputras, whose bodies are disgured by plentiful apotropaic rites (?), and by ascetics with staves and antelope skins, who say, this lord is veritably a perfected wizard (s medhika); a he has attained communion with powerful yogins, and reveals extrasensory knowledge, and possesses supernatural power. Through magical rapprochement, he can make even a lion accord with an elephant, and through drugs for causing enmity, he can make even a mother inimical to her children.). Cf. Handiqui, Yaastilaka and Indian Culture, 5758. Handiqui renders the highly uncertain sattriputrair s with acting as informers, a description of the parivr jakas. The commentary of Srutadevasuri reads a a the no less opaque sanniputraih, as does the edition of S str, who reports satriputraih as a variant. . . Srmatsomadevasuri-viracitam Yaastilakacampu Mah k vyam, vol. i, p. 253. s a a . 259 That the temple of Candam ri stands in the royal capital is stated in Yaastilaka i, vol. 1, p. 44 a s .. ( tmar jadh ny m eva candam ridevat y h puratah), while its name is given in the prose on p. 148 (line a a a a a a. .. a . 45, mah bhairavam n ma tad dev yatanam). a a . a 260 Descriptions of Candam r include Yaastilaka i, vol. 1, pp. 15051 (in part quoted and translated by a s .. Handiqui, Yaastilaka and Indian Culture, 56), and verses i.13637. Her iconography is entirely k p lika, s a a and her descriptions include images of sacricial violence; cf. B nas descriptions of Candik in the a. .. a K dambar (pp. 22428 of the edition). a 261 BraY xxxviii.41c lists texts called M r and Mah m r in its account of the Vidy ptha division a a a a a . of the bhairavatantras. A goddess called Mah m r is mentioned in Tantrasadbh va 20.11113, with her a a a mantra and its application given in 20.124cd139ab. This mantra-deity is, ttingly, connected with magical slaying (m rana). a . 262 Chapter one-hundred and thirty seven, of which the colophon reads ity agneye mah pur ne yuddhaa a. jay rnave mah m r n ma saptatrimsadadhikaatatamo dhy yah. a . a a a s a . . 263 M rkandeyapur na 92.3536 (Devm h tmya 12). a .. a. a a

109 Somadevasuri connects Candam r and her temple with mah yogins, bellicose a a .. goddesses whose ight through the heavens he depicts vividly. In contrast to the M latm dhava, in which yogins were accomplished human adepts, here they are pora a trayed as a horde of powerful deities. Brandishing skull-staves (khatv nga), adorned . a with snakes and skulls, and with third-eyes blazing, the yogins assemble in the Ca ndam r temple from the skies, the earth, the depths of the netherworlds, and the a .. intervening regions (digantar la).264 The ight of the tempestuous goddesses shakes a the heavens as they descend like the nights of universal destruction.265 While not specically described as enshrined within its precincts, their association with Can. dam rs temple is noteworthy, for, as Vidya Dehejia observes, the Yaastilaka belongs a s . to the period in which major Yogin temples were constructed.266 The temple en virons being lled by alighting yogins evokes the language of embodiment used for describing deities as living presences in a temple, and suggests the possibility of their presence as sculpted images. This possibility receives some support in the reference made to worship of the circle/group of Mothers (m trmandala) within the a. .. temple precincts; at least one period source appears to uses the term m tr to refer a. to the sixty-four goddesses of a yogin temple.267 Irrespective of this possibility, the Yaastilaka appears to point toward a contemporaneous association between yogins s and temples.
a Note in particular, sasamrambham ambaratal d il y h p t lamul d digantar lebhya ca vibh vary m a a a. a a a s a a. . tamahsamtatibhir iv virbhavantbhih . . . (describing mah yoginbhih; Yaastilaka i, vol. 1, pp. 4445). a . a . s . . 265 Cf., especially, Yaastilaka i, vol. 1, p. 47: sakalasya jagatah ksayaksap bhir iv tid runadrghadeh bhir s a a . a . . . a mah yoginbhir apuryam naparisaram . . . ([the temple of Candam r,] the precincts of which were lling a a. a .. up with powerful yogins, whose bodies were extremely frightful and long, like the nights of the whole worlds destruction; describing dev yatanam [p. 49]). Cf. Handiqui, Yaastilaka and Indian Culture, 5657. a s 266 Yogin Cult and Temples. A Tantric Tradition, 26. Dehejia provides an English translation of this passage from the Yaastilaka, ibid., 2627. s 267 Yaastilaka i, vol. 1, p. 49, kvacit tksnapurusapakrs. asvaky ntrayantradolanatosyam nam trmandalam, s .. a a. a . . . . ..t . describing dev yatanam (in one part [of the temple], fanatic people appease the group/circle of Mothers a by swinging by the contraption of their own extracted entrails). Cf. Handiqui, Yaastilaka and Indian s Culture, 22. The eleventh-century Siy n inscription of Bengal refers to the installation of Bhairava a surrounded by sixty-four Mothers, i.e. yogins; this is discussed in the subsequent section.



temples of the yogins i

By the middle of the tenth century, if not somewhat earlier, yogins became the focus of a temple cult of wide geographic distribution and evident prominence, with the construction of major yogin temples continuing through perhaps the thirteenth cen tury. Erected in stone from Orissa to Hinglajgadh, on the Madhya PradeshRajasthan border, and as far south as Tamilnadu, the circular, hypaethral (open to the sky) temples of the yogins are architecturally unique in medieval India, and remain enigmatic in terms of religious history and ritual function. Numerous temples and much loose yogin statuary have been documented by Vidya Dehejia, beyond whose important monograph on yogin temples only a few additional contributions have been made in the past two decades.268 Despite her efforts to relate the art historical record to literary accounts of yogins, the state of the study of Saiva literature limited Dehe jias access to the textual corpus of greatest relevance to the yogin temple tradition. Bringing these two bodies of evidence together still presents considerable challenges, and in the present section I offer only preliminary, general reections upon the yogin temples in light of Saiva textual sourcesan area to which I expect to devote further study. Of the yogin temple sites and image sets no longer in situ identied by Dehe
Dehejia, Yogin Cult and Temples. A Tantric Tradition. From the earlier literature, I should mention the unpublished doctoral thesis of Margrit Thomsen: Kult und Ikonographie der 64 Yogins (Free University of Berlin, 1976). Though less comprehensively than Dehejia, Thomsen was the rst to bring together references to yogins from a broad range of pur nas, primarily, and analyse the major extant a. yogin temples. Note also her brief article, Numerical Symbolism and Orientation in Some Temples of the 64 Yogins, Art and Archaeology Research Papers (March 1980): 5356. Of less signicance is H. C. Das, Tantricism: A Study of the Yogin Cult (of 1981); Das puts together summaries of a number of episodes concerning yogins from the pur nas, and provides image-by-image descriptions of yogins a. from several templesbut only in some cases based upon rst-hand observations. His subsequent a discussion of yogins, Iconography of S kta Divinities, vol. ii, 40867, is drawn largely from the former publication. The only monograph concerning a single yogin temple is R. K. Sharmas The Temple of Chaunsatha-yogin at Bheraghatironically titled, as Dehejia points out, since this temple probably . enshrined eighty-one rather than sixty-four (chaunsatha) goddesses. The bulk of this book consists of an . image-by-image description of the murtis. As Dehejia remarks (p. 126), Sharma contributes little new knowledge concerning the Bher gh t temple, which had been more ably surveyed by R. D. Banerji, The a a Haiyayas of Tripuri and their Monuments. A revision of a doctoral thesis of 1984, the somewhat supercial monograph of Rajendra Prasad Simha also concerns yogins: Caumsatha Yoginiy m evam unake Mandira. a. . . More recently, Thomas Donaldson has brought his considerable knowledge of regional art to bear upon a the two yogin temples of Orissa. Tantra and S kta Art of Orissa, vol. 2, 66174.

111 jia, which number around fteen, at least ve were concerned with sets of sixty-four yogins. The Bher gh t temple contains eighty-one goddesses, while two ruined tem a a ples of central India might have housed sets of forty-two.269 In the remaining cases, the original number of images cannot be determined. With exceptions, such as the rectangular Khajuraho yogin temple, the temples have circular structures, and often feature a central shrine in the courtyard with a cult image of Siva or Bhairava. All are hypaethral. Although the sites of the extant yogin temples and those associated with known loose statuary are concentrated in modern Madhya Pradesh and Orissa, the medieval temple cult was certainly not conned to these regions: one set of yogin images has been found in northern Tamilnadu,270 while inscriptional evidence points towards construction of yogin temples in Bengal as well.271 The extant images do not enshrine a group of deities with xed individual identities: any given set of yogins is unlikely to tally with another, with the exception of the Seven Mothers, who appear with particular frequencya fact entirely consonant with textual accounts of yogins. One of two extant temples located in Orissa, the smallest of the yogin temples is situated in a secluded clearing near the village Hr pur, not far from Bhuvane var a s (figure 2.15).272 According to one regional source, this would place the temple within the boundaries of Ek mra, an important religious center that is included among the a p. has or sacred sites enumerated in early tantric literature, such as the BraY .273 t a Thomas Donaldson, whose comprehensive surveys of Orissan art have considerable value for the study of early tantric traditions, opines on stylistic grounds that the
Near Dud h survive the remains of a temple of forty-two yogins, while Dehejia argues that the a superstructure of the present Gadarmal Mother temple at Badoh was built atop a rectangular temple of forty-two yogins; she nds evidence for forty-two niches, as well as eighteen fragmentary goddess images. Yogin Cult and Temples, 14243. 270 Ibid., 18183. 271 On the Siy n inscription of Bengal, see below. a 272 The following description of the temple is based upon analysis of photographs in the aiis Photo a Archive, and the images published by Thomas Donaldson, Tantric and S kta Art, vol. 3. 273 a Donaldson, Tantra and S kta Art, vol. II, 661; he refers to the Ek mracandrik , as quoted by R. L. a a Mitra, Antiquities of Orissa, vol. II (Kolkata: Wyman & Co., 1880), 103. The sacred geography of the BraY is discussed in the subsequent chapter. a

112 temple belongs to the middle of the tenth century.274 Comparatively well preserved, this small-scale hypaethral temple has an entrance passage protruding outward from its circular structure, lending to the whole the shape of the yoni-pedestal of a Siva linga, as Margrit Thomsen suggests.275 A (rebuilt) rectangular shrine stands in the center.

Figure 2.15: Frontal view of the Hr pur yogin temple. AIIS Photo Archive. a

Nine grim goddesses of relatively little iconographic distinction appear around the exterior of the circular structure, which is unusual in yogin temples. Local tra dition points toward identication of these as the nine k ty yans (figure 2.18).276 a a Two male door guardians (dv rap la) ank the entranceway, while a skeletal male a a deity is present on either side of the entrance passages interior. The iconic program of the interior of the temple features sixty-four yogins, sixty of whom are arrayed in shallow cells along the interior perimeter of the circular wall. The goddesses therefore face towards the central shrine, a rectangular structure with four entrances
This date is posited based upon similarity to the alasakany images on the Mukte var temple of a s nearby Bhuvane var, which Donaldson assigns to the mid tenth-century. Dehejia had also noted these s similarities, but considered the Mukte var temple to be somewhat older; she places the Hr pur temple s a in the second half of the ninth century. Yogin Cult and Temples, 99100. 275 Thomsen, Numerical Symbolism, 53. 276 a See the discussion of Donaldson, Tantra and S kta Art, vol. II, 662.

113 that might once have housed an image of Siva.277 Flanking each door of the central shrine is a deity pair: two pairs of yogins, who complete the sixty-four, and two pairs of bhairavas. The latter include Ekap dabhairava (figure 2.20)a deity of regional a a signicance who also features in the BraY under the name Jhank rsa or Jhamk raa . a bhairava.278 One of the sixty goddesses along the perimeter (no. 31) appears to be leader of the group, for her image is positioned opposite the entrance, possesses ten arms rather than two or four, and is signicantly larger. Perhaps intended as female counterpart to the central Siva, this unidentied image is unfortunately in a poor state of preservation.279 Viewing the temple as a mandala, an analogy likely to .. have substance in this case, Siva stands at the center with an inner circuit of deities comprised of four yogins and four bhairavas, surrounded in the second circuit by sixty yogins. Nine k ty yans form the outer layer of divinities, while additional a a deities guard the mandala entrance (dv ra). a .. The yogins of the Hr pur temple compose an expressive and variegated set, a for detailed descriptions of which I refer the reader to K. N. Mahapatra, Donaldson, and Dehejia.280 Their iconography ranges from macabre (figure 2.21) to martial, playful (figure 2.16), and sensuous (figure 2.17). A number of the goddesses are theriomorphic (figure 2.19) and still more have animal v hanas (figures 1617, 21), a encompasssing species of considerable variety. Particularly common are images of dance, music, and war. Several, but still a small minority of the yogins have k p lika a a iconography; some, for instance, carry or even drink from skull-bowls (figure 2.17) or stand upon severed heads or corpses. Excepting C munda (figure 2.20), the most a . .
Dehejia reports that the image of Siva had been in situ when the temple was rst discovered. Yogin Cult and Temples, 95. However, the basis for her claim is unclear, for nothing of the sort is reported by the rst scholar to document the temple: K. N. Mahapatra, A Note on the Hypaethral Temple of Sixtyfour Yogins at Hirapur, Orissa Historical Research Journal II (1953?): 2340; reprinted in H. K. Mahtah, ed., Orissa Historical Research Journal, Special Volume, 1982. 278 a On images of Ekap dabhairava in Orissa, see Donaldson, Tantra and S kta Art, vol. 2, 46464. I a have noted sculptures of Ekap da only in Orissa and neighboring Andhra Pradesh. a 279 Published in Dehejia, Yogin Cult and Temples, 63. A full-bodied goddess, this image probably does not represent C munda, who appears elsewhere in the circle (figure 2.21). None of the iconic emblems a .. associated with her damaged ten arms remain discernable. Perhaps Bhairav? 280 a Mahapatra, Note on the Hypaethral Temple; Donaldson, Tantra and S kta Art, vol. 2, 66167; and Dehejia, Yogin Cult and Temples, 95102.


Figure 2.16: Hr pur yogin with damaru a . drum, standing upon wheel and rat (?). AIIS Photo Archive (detail).

Figure 2.17: Hr pur yogin on lotus and elea phant, drinking from a skull-bowl. Photograph by Don Stadtner, from the AIIS Photo Archive.

macabre deities of all are not the yogins themselves, but the goddesses upon the tem ples exterior facegrim, well-armed deities perched upon heads (figure 2.18)and the two skeletal male deities in the entrance passage. While few other sculpted sets approach the expressiveness of the Hr pur deities, the breadth of characterization a of the latter is typical of both textual and iconic representations of yogins. Also perhaps of the mid-tenth century, the larger yogin temple at R npur-Jharial a in Orissa houses sixty-four goddesses, with its simple, four-pillared central shrine containing a dancing Siva (figure 1.1). Donaldson suggests that a damaged and relocated image of similar scale, perhaps of C munda, might have originally have been a . . situated in the company of the central Siva.281 Probably of the same period, the yogin temple of Khajuraho in central India has a rectangular rather than circular plan,

a Donaldson, Tantra and S kta Art, 670.


Figure 2.18: One of nine k ty yans (?); Hr pur a a a yogin temple, to the right of the entrance, exterior facade, standing upon severed head. AIIS Photo Archive.

Figure 2.19: Lion-faced yogin, Hr pur temple. Photograph by Don a Stadtner; included in the AIIS Photo Archive.

although it too is open to the sky. Though none of its images appears to be extant, the niches number sixty-ve, with no indication of there having existed a central shrine; a comparatively large niche opposite to and facing the entrance presumably housed the central cult image, whether of Siva or the Goddess. A ninth-century dating had been proposed for this temple, but Dehejia considers the mid-tenth century more plausible on stylistic grounds.282 To the south, on an isolated hilltop near Jabalpur, stands the most imposing and perhaps best known of the yogin templesthat of Bher gh t. A circular structure, this temple appears to have enshrined yogins in a a each of eighty-one cells along the covered walkway of its inner perimeter. The shrine in the courtyard belongs to a later period, while, as Dehejia points out, the images of eight Mother goddesses now included among the yogins belong to an earlier

Dehejia, Yogin Cult and Temples, 11517.


Figure 2.20: Ekap dabhairava, Hr pur yoa a gin temple. AIIS Photo Archive.

Figure 2.21: C munda, Hr pur yogin tema a . . ple. AIIS Photo Archive (detail).

period.283 As for the temples dating, Dehejia suggests, with a degree of plausibility, that the labels inscribed on the yogin images bases date to the late tenth or early eleventh century, based upon paleographic comparison.284 This might make the Bher gh t temple roughly contemporaneous with the badly damaged temple of a a forty-two yogins at Dud h, near Lalitpur, M.P., but earlier than the Mitauli temple a of sixty-four yogins near Gwalior; a damaged inscription seems to place construction of the latter in the third quarter of the eleventh century.285 There are clear indications of royal patronage of several of the above temples.286
Ibid., 12527. While Dehejia describes this as a temple of eighty-one yoginsand quite plausibly sothe iconic program is not entirely clear; many images are missing and few remain in their original positions, while three images of Vin yaka and two of male Saiva deities presently occupy niches. Some a of these appear similar in style to the yogins, and might have had places in the circle or in the original central shrine, asssuming one had been present at all. 284 Ibid., 13839. 285 Dehejias proposed dating of the Dud h temple is based upon inscriptional evidence from a a nearby Brahm temple. Dehejia, Yogin Cult and Temples, 141. On the date of the Mitauli temple, see a ibid., 88, 12324. 286 The Mitauli temple appears to have been errected by the Kacchapagh ta ruler Devap la (105575 a a c.e.), and an inscription shows that this temple still received major patronage as late as 1503 c.e. Ibid.

117 Among the loose sets of yogin images identied by Dehejia, one that was prob ably installed on a hill-top temple in Lokhari, Uttar Pradesh, merits particular mention. Although only twenty images have been identied, these are almost all theriomorphic,287 and hence suggest a unique iconic program (figure 2.6). If Dehejia is correct that these images belong to the rst half of the tenth century, this could also represent one of the earliest of the temple sites so far identied. Indeed, no compelling evidence I am aware of points toward the existence of yogin temples prior to the tenth century, although it would warrant little surprise if earlier examples surface. Most of the extant temples appear to have been constructed in the tenth through twelfth centuries, and the case appears similar with regard to the loose statuary identied as once belonging to yogin temples.288 It is of course probable that stone structures were preceded by yogin shrines and temples made from perish able materials; worship of yogins has indeed continued into modern times in more humble temple contexts, sometimes using aniconic images.289 Davidson mentions the cases of two temples at Mandi, Himachal Pradesh, where yogins are represented .. by sixty-four sets of footprints on slabs of stone.290 That this could represent an old tradition is suggested by the depiction, below one of the Bher gh t yogins pedestal, a a of worshippers paying reverence to a set of sandals (figure 2.22). The hypaethral temples of the yogins could perhaps have continuity with ancient conventions of shrines open to the elements.291 However, the template that appears most directly applicable to the yogin temple is the tantric yogincakra, as Dehejia recognized:292 the mandala of mantra-goddesses surrounding Siva/Bhairava, instal..
The R npur-Jharial and Khajuraho temples, as part of large complexes of stone temples, also were a surely constructed under royal patronage, the later under the auspices of the Candella rulers. 287 Among the images from Lokhari published by Dehejia, only C munda is not theriomorphic. a . . 288 On the loose yogin statuary, see Dehejia, Yogin Cult and Temples, 14584. 289 David White, for instance, publishes a photograph of painted stones embodying yogins outside of the Ll d temple of Ghatiyali, Rajasthan. Kiss of the Yogin, 268. a. 290 Indian Esoteric Buddhism, 18183. 291 Note, for instance, depictions in early Indian art of simple linga shrines, with the phallic cult object set up upon a platform under a tree. Two are published by Doris M. Srinivasan, Many Heads, Arms and Eyes. Origin, Meaning and Form of Multiplicity in Indian Art, plates 17.2 (rst century), 17.67 (a mukhalinga of Mathur , early centuries c.e.). a 292 Dehejia, Yogin Cult and Temples, 2, 18586.

118 lation of which in various inections and upon a variety of substrates was central to the ritual of the esoteric Saiva cult of yogins. One might suspect other inuences, for instance tribal or folk traditions of worshipping goddesses in a circle; yet tantric worship of circles of yogins appears to predate the temples by at least two centuries, and the remarkable congruity in Saiva textual representations of yogins and their depiction in sculpture suggest direct continuity. It does appear possible that yogin temples incorporated local deities into their iconic programs;293 but in no case has this been demonstrated to be a process of central signicance. Some yogins who have been singled out as local deities in fact present ambiguous cases.294 It is thus difcult to concur with Dehejia and Donaldson, without more evidence, in the view that the yogins enshrined in temples represent localised cult traditions of village deities that eventually were transformed into potent groups of sixty-four yogins.295 The extraordinary diversity of names and identities attested in yogin scultpures appears wholly consistent with the tantric yogin cult, with almost each scriptural tradition offering its own pantheon of this malleable category of divinity. Non-elite and tribal traditions might have been highly signicant to the yogin cult, however difcult to document; but if these were formative inuences, they should be looked for in the early Saiva and Buddhist esoteric traditions, given the apparent chronology of evidence.
Dehejia, for instance, points out that sixteen yogins listed in a late medieval Orissan pur na are a. worshipped as independent goddesses in Orissa. Ibid., 93. 294 The non-Sanskrit names of some goddesses might suggest locally meaningful identities. Dehejia draws attention to the yogin labelled Teramv at the Bher gh t temple, identical to Durg a a a a Mahisasuramardin in iconographic type. One might mention the cow-faced Erudi of the same site. . . See ibid., 13336. However, the name Teramv , in the Sanskrit form tryambak , is an attested name of the a a spouse-goddess of Siva; cf. Kaulaj nanirnaya 14.28. In both vernacular and Sanskrit forms, this occurs a . in pan-Indian sources as a personal name (in the masculine) and as the name of a Saiva monastic order (mathik ). See Sanderson, A Commentary on the Opening Verses of the Tantras ra of Abhinavagupta, a . a 121 (n. 79). It is hence problematic to ascribe local meaning to this name in the absence of evidence from regional sources. Similarly, a Saiva p. ha called Erud is listed in the Tantrasadbh va (13.74b and 15.70b), t a . and is also mentioned by Abhinavagupta in Tantr loka 15.91a (as Erudik ). The Bher gh t yogin is a a a a . perhaps connected to the p. ha by this name, but there are no grounds for linking the Erud-p. ha to the t . t Bher gh t region. a a 295 a Donaldson, Tantra and S kta Art, 658 (quotation); Dehehjia, Yogin Cult and Temples, 9394.

119 No textual material intimately connected with the temples of yogins appears as yet to have come to light, such as a manual concerned with their construction, consecration of the images, and worship within the temple precincts. This is hardly surprising, for given the fragmentary record and present state of research, this level of correlation between text and temple remains rare in the study of early meFigure 2.22: Veneration of sandals representing a yogin, depicted below a yogin images pedestal. Bher gh t yogin temple. AIIS Photo Archive. a a

dieval India. Perhaps more surprising is the apparent paucity of references

to yogin temples in period literary sources. As discussed previously, the association between yogins and the temple of Candam r in the Yaastilaka is highly suggestive, a s .. given that this mid tenth-century work dates to the period of the construction of major yogin temples. Dehejia claims that a tale in the Akhy nakamanikoa, a Jaina a . s work of the late eleventh century, makes reference to a yogin temple.296 We also nd fascinating references to yogins in an eclectic Persian work presenting itself as a translation of the K mrubj ksa, evidently a tantric Saiva text connected with the a a . cult of the goddess K m khy . The Persian redaction, dating to as early as the thira a a teenth century, declares the sixty-four yogins the most revered deities of the Hindus, who worship their idols.297 The R jatarangin of Kalhana attributes several gures a . of various Kashmiri courts with endowments to m trcakras (circles of the Mother a. goddesses), but despite suggestions to the contrary, none of these are likely to refer to yogin temples.298 As for the epigraphic record, this also appears minimal; an
Dehejia, Yogin Cult and Temples, 55. I have not yet had access to this text. Carl Ernst, The Islamization of Yoga in the Amrtakunda Translations, 2038, 21920. I make reference to this text in the introduction (n. 65). 298 Four references to m trcakras or devcakras in the R jatarangin were brought to my attention by a. a
297 296

120 eleventh-century inscription from Siy n in Bengal does however state that a yogin a temple was constructed in this region.299 A number of textual sources on temple architecture, such as the Brhatsamhit and Agnipur na, mention the possibility of cira. . . a cular structures, but the yogin temples are the primary surviving representatives of this rare type.300 Tantric Saiva literature itself appears silent on the construction of yogin tem ples, although a large quantity of potentially relevant material awaits study with this question in mind.301 There is nonetheless much that can be said, preliminarily, concerning the yogin temple cult and Saiva literature. Although congurations of eighty-one and forty-two are attested, the extant temples suggest that in the period of their construction, yogins were normally considered a set of sixty-four goddesses, a numerical association pervasive in pur nic accounts of yogins that persists through a. the late medieval and modern periods. Much like the popular notion that tantras number sixty-four, this numerical association belies the fact that few lists agree in their particulars.302
White, Kiss of the Yogin, 137, who cites Marie-Thrse de Mallmann, Les Eseignements iconographiques de lAgni-Purana (Paris: Presses Universitaires de France, 1967), 173. White implies that these might have been yogin temples. Three cases do clearly refer to templesprobably of the Mother goddesseswhile that of v.55 refers to a temple of Bhairava associated with a m trcakra. If M. A. Steins identication of a. the latter site is correct, it seems improbable that the ruins of the Bhairava temple at Buthiser resembled extant yogin temples. See Stein (trans.), Kalhanas R jatarangin. A Chronicle of the Kings of Kamr, vol. a s . i, 194. In any case, this section of the text pertains to events of the ninth century, before yogin temples are attested; the other references (i.122, i.333 [not 350, as gives White], and iii.99) pertain to even earlier periods. 299 The eleventh-century Siy n inscription mentions, among various pious works undertaken by the a Bengali monarch Nayap la and his ancestors, installation of Bhairava surrounded by sixty-four Mothers, a i.e. yogins, possibly in the capital city of the ruler (ghan. sam yah svanagare nyadh t ksem ya dehin m | a . a a. . t . . catuhsas. y ca m tnam partan [em. Sircar; partat] tatra bhairavam ). D. C. Sircar, Epigraphia Indica a r. . . . .t a . . xxxix, 3956. This inscription was brought to my attention by Davidson, who also cites the Momin b d a a inscription of Uday ditya as recording the construction of a yogin temple. Indian Esoteric Buddhism, a 183. However, the 1144 c.e. Western C lukya inscription of Momin b d appears to pertain only to a a a a cave temple of the goddess Jog or Ambejog (jog yogin); I see no indication in the inscription for a a a the presence of multiple yogins. P. B. Desai, Selected Stone Inscriptions of Andhra Pradesh, 9495. 300 See Dehejia, Yogin Cult and Temples, 42. 301 The Pratis. h tantras in particular, works concerned with the construction and consecration of tem.t a ples and images, are likely to be pertinent; I have unfortunately studied little even of the Pratis. h tantra .t a a associated with the BraY : the unpublished Pingal mata, transmitted in Nepalese manuscripts. a 302 Several scholars have given accounts of references to yogins in the pur nas, perhaps the most a. complete being that of Olga Serbaeva, Yogins in Saiva Pur nas and Tantras. Their role in transformative a. experiences in a historical and comparative perspective, especially 4360. Incidentally, the earliest reference to tantras as sixty-four appears to be Siddhayogevarmata 29.19. s

121 The earliest textual sources of the Yogin cult do not in fact associate yogins with the number sixty-four. In the BraY , groups of six dominate, with the primary set a of six comprising the ancillary (anga) mantras of the supreme Goddess; the largest congurations of yogins involve multiple groups of six, such as twenty-four.303 The primary group of sixty-four is a set of rudras, deities whose mandala forms the subject .. of BraY xxx. Sets of sixty-four goddesses are also absent from the Siddhayogevara s mata, another early scripture of the Vidy ptha. A Vidy ptha text postdating the a . a . BraY and Siddhayogevarmata,304 the Tantrasadbh va presents what might be the eara s a liest reference to sixty-four yogins. In chapter thirteen, after delineating a mandala .. of sixty-four bhairavas distributed in eight lotuses around a central lotus, a parallel conguration of sixty-four goddesses is introduced, called the m try ga or pana. a theon of Mothers. Their names are feminine-gender mirrors of the bhairavas, suggesting a secondary status. Around the central Bhairava, the sixty-four yogins form octads in lotuses grouped according to clans of the Eight Mothers, from Brahm n a. to Aghor.305 This explicit link to the Mothers is highly signicant, for the same association is evident in most of the extant sculpted sets of yogins, who include among them identiable images of the Mother goddesses. This association persists in the profuse yogin name lists of the pur nas as well, which most frequently concern a. sixty-four yogins and often include among them the Mothers.306 In tantric Saiva literature, the association between sixty-four and yogins appears particularly common in Kaula sources, which, as discussed in the next chapter, ap303 On the Six Yogins of the BraY , see the section nav ksaravidh na in chapter 5 of the present a a . a dissertation. BraY xxix in particular features large sets of goddesses. This chapter teaches a pantheon a and worship specic to the Four Devs, the extensive mandala based upon which involves numerous .. sets of six yogins. This mandala is based nine cakras: the root cakra of Bhairava and Aghore, in s .. the center, with cakras in the eight directions upon which are installed one of the Four Devs or Four Duts/Kinkars, each associated with Six Yogins and servants of the servants (kinkaryanucar h). The a. latter groups of eight are associated with each of the Eight Mothers, forming, in a fashion, a set of sixty-four goddesses. 304 See the discussion in chapter 4. 305 The m try ga begins with Tantrasadbh va 13.56, and continues until the end of the chapter (verse a. a a 88). See the collation/draft edition of Dyczkowski. 306 For pur nic and other yogin name lists, see Dehejia, Yogin Cult and Temples, 187218 (appendices a. iiii). As the case of the sixty-four yogins of Tantrasadbh va 13 suggests, the absence of Mothers from a a list of yogins should not be take as evidence of dissociation between the two goddess types.

122 pear in general to postdate the early Vidy ptha literature. Hence in the Kaulaj naa . a nirnaya, a scripture describing itself as the Yoginkaula (Kaula Scripture of the Yo . gins), the most important cult deities are the sequence of the sixty-four yogins (catuhsas. iyoginkrama), whose mandala and ritual worship (the as. as. akavidhi) are cen .. . . .t . t . t tral to this text.307 Preserved in a manuscript of the mid eleventh century, this text contains no indications of particular antiquity,308 and might date to the early period of the production of yogin temples. Its twenty-third chapter describes the manifes tation of the sixty-four yogins as female beings of every variety in particularly vivid terms. Yogins are said to sport on the earth as female animals of any type, from pigeons and vultures to cows and cats. When they assault non-devotees, they manifest as snakes, rats, tigers, and so forth, and as dangers such as disease, lightning, thieves, and royals. One is admonished never to insult women; we are not informed whether this is because yogins afford them special protection, or because any given woman might secretly be one.309 While not specically concerned with temples,
307 The titular epithet yoginkaula occurs in most of the colophons of the Kaulaj nanirnaya, and also a . in 16.49c. The sequence or cakra of the sixty-four yogins forms the subject if Kaulaj nanirnaya 8 in a . particular, as well as chapter ten, while a number of other chapters refer to the yogins as sixty-four in number. 308 On the Kaulaj nanirnaya, see the discussion in chapter 3, section 3. a . 309 Kaulaj nanirnaya 23.112ab (provisional edition, reporting the codex nak 3-362 or ngmpp reel a . a48-13 [Kjncod ] and the editio princeps of P. C. Bagchi [Kjned ]):

devy uv ca a kaulave yogin deva samcaranti katham bhuvi | . . tan mam caksva devea bhakt j nanti bhutale 1 a . s a a bhairava uv ca a martye smin devat n m tu samc ram srnu bh mini | a a. a . a . . . kayotik tath grdhr hams caiva nakh tath 2 a a . . a khajar. akabh s tu kok bh s tu sundari | t a. a a. ::::: uluk pecak v tu sarar v gdul tath 3 a a a srg l aj mahis us. r m rj rarupin | a a .t a a . . . vy ghr hast mayur ca kukut nakul tath 4 a a . a. any ni y ni rup ni samsthit ni mahtale | a a a . a. t ni rup ni samgrhya yoginyah krdante bhuvi 5 a . . . . nipatanti yad bhadre abhaktesu kul dhipe | a a . . tadrupam kathyate bhadre sraddh yukt vadh rayet 6 a a a haya ca nakharah sarpa citriko ghonasas tath | s s a . . vrsciko dhyantara sv no musako dardurah priye 7 a . . . grahabhutasvarupena jv l gniastrasamkataih | aa s . . . vedan ca jvaravy dhi r j na caiva taskar h 8 a a aa s a. vidyut tungo tath ganda vy ghra simho gajas tath | a .. a a . anek k rarupena bhayam n n vidham viduh 9 a a . . a a . .

123 this text, among numerous other period sources one might cite, gives indications of contemporaneous conceptions of the enshrined yogins. Although the Kubjik mata does not itself refer to yogins as sixty-four in numa ber,310 later literature of the Kaula cult of the goddess Kubjik suggests a strong link. a The Cicinmatas rasamuccaya, the essence text (s ra) of the Kubjik mata,311 appears a a a to refer to airborne yogins (gaganag min) as sixty-four, while other types, such a as terrestrial (bhuvanag min), are considerably more numerous.312 A work called a the Kubjik puj paddhati enumerates sixty-four yogins by name, while the Srmatottara a a (i.e. Kubjik mata Uttaratantra) has extensive material on yogins, both sixty-four and a eighty-one.313 More signicantly, the unpublished fteenth chapter of the Sats hasra. . a a samhit provides a detailed account of the iconography and worship (puj ) of sixty. a
catuhsas. hi ca yoginyo yath kupyanti s dhake | a a . . .t s evam rupam sam sritya ksipram grhnanti tam paum a s . . . . . . . kopam tu naiva kartavyam n pam nam sur dhipe | a . a . . a kum rik h striyo v pi n pam net kad cana 11 a a. a a a a yath sakty sad k lam str caiva vratam asthitaih | a a a a . . a pujany prayatnena kum rya ca kul sritaih 12 a s a . 10

1b bhuvi ] Kjncod ; avi Kjned 3a khajartaka ] conj.; kamja - - - Kjncod Kjned 3c uluk ] Kjncod ; . . .. uluk Kjned 3d v gdul ] Kjncod ; v gul Kjned a a 4b ustr ] Kjncod ; ustr Kjned m rj rarupia a .. n ] Kjncod ; m rj ranakul tath Kjned a a a 4d nakul tath ] Kjncodpc ; na - - - - Kjncodac Kjned a 6b . abhaktesu ] Kjncod ; abhaksyesu Kjned 6c kathyate bhadre ] Kjncodpc ; kathya - - - Kjncodac ; kathyante . . . . . . Kjned 6d sraddh yukt ] conj.; s ( - ?) yukt Kjncodpc ; - - yukt Kjncodac Kjned a a a a 7b citriko ghonasas ] Kjncod ; citrikotam nasas Kjned a 7c dhyantara sv no ] Kjncod ; dhyantara c no a s a Kjned 8b sastra ] Kjncod sastra Kjned 8c vedan ca jvaravy dhi ] Kjncodpc ; veda - - - - vy dhi a a a Kjncodac ; veda . . . vy dhi Kjned a 8d r j na ] Kjncod ; r j na Kjned aa s aa taskar h ] Kjncodpc Kjned ; a. taskak r hKjncodac a a. 9a vidyut tungo ] Kjned ; vidyu tungo Kjncod 10a catuhsasthi ] Kjned ; . . .. s cod ed ; yoginya Kjncod catuhsasthi Kjn yoginyo ] Kjn 10b yath kupyanti ] Kjncodpc ; - - - pyanti a . . .. Kjncodac Kjned 10c sam sritya ] corr.; sam srtya Kjncod Kjned 11b n pam nam ] Kjncod ; bh sam a a . a a . a. a nam Kjned 11c kum rik h ] em.; kum rik Kjncod Kjned 11d n pam net ] em. Isaacson; n pam ne a a. a a a a a a . . Kjncod ; bh sam ne Kjned 12b asthitaih ] em. Isaacson; asthitam Kjncod Kjned a. a . .

I defer discussion of the interpretation of this passage until publication of a new edition of the Kaulaj nanirnaya, currently under preparation. a . 310 Kubjik mata 6.8791, for instance, speaks of the Mothers as seven groups of seven, with additional a unspecied subdivisions; the yogins are yet more numerous. 311 Cf. Cicinmatas rasamuccaya 8.4cd5ab: a kathitam devadevei samksep n na tu vistaram 4 s . . . a srkubjik matas ro yam vistaram kathitam may | a a a . . . From the draft edition of Dyczkwoski. 312 See the ve mantras given in prose following 9.121 in Dyczkowskis draft edition; the rst appears a a to be addressed to the sixty-four ak sag min yogins. 313 I have not personally studied either of these unpublished works. The sixty-four yogins of the Kub jik puj paddhati are tabulated in Dehejia, Yogin Cult and Temples, 214; she draws the list from Gopinath a a Kaviraj, T ntrika S hitya, 135. Dehejia makes extensive reference to the Matottara throughout her study, a a especially 4551.

124 four yogins.314 The Sats hasrasamhit survives in manuscripts as early as the twelfth . . a . a century, apparently, but postdates the Kubjik mata (10th century?), therefore belonga ing to the period of the major yogin temples.315 Olga Serbaeva tabulates three ad ditional sources for the iconography of the same set of sixty-four yogins: the Agni pur na, Pratis. h laksanasamuccaya, and the Mayadpik , the latter as cited by Hem dri a. a a .t a . . in the Caturvargacint mani.316 Although not explicitly connected with temples, these a . pre fourteenth-century sources, both Kaula and otherwise, attest a tradition of iconic representation of sixty-four yogins. The yogin temple pantheons hence have parallels in contemporaneous textual accounts of the sixty-four yogins. However, the representation and worship of yo gins in images have earlier precedents. The BraY itself has extensive material on a iconography and iconometry in its fourth chapter, which has as its subject the characteristics of images (pratim laksana). This chapter provides detailed instructions a . . on constructing and empowering images of the cult deities of the BraY , whose ima ages it classies as supradivine, divine, and semidivine (divy dhika, divya, and a divy divya). The goddesses of the mandala of Kap lsabhairava are considered dia a .. vine; these include the Six Yogins, the four guhyak s/devs, four kinkars/duts, and a Mothers, as well as their male counterparts (pati, vra), the sixty-eight rudras, yogins of the sacred elds, and guardians lokap las.317 As an iconic type, the BraY s a a
314 This was brought to my attention by Olga Serbaeva, who quotes Sats hasrasamhit 15 from a draft . . a . a edition of Sanderson. Yogins in Saiva Pur nas and Tantras, 75 (in Cited Sanskrit Passages) and 56 a. 61 (appendix 7.6). This set of sixty-four yogins is linked to the Eight Mothersthe standard seven of Br hm to C munda, plus Mah laksm. a a a . . . 315 On the date of the manuscripts of the Sats hasrasamhit , a text incorporating within itself the . . a . a Kubjik mata, see J.A. Schoterman, The Sats hasra Samhit . Chapters 15: Edited, Translated and Annotated, a . . a . a 1213. 316 Serbaeva, Yogins in Saiva Pur nas and Tantras, 5661 (appendix 7.6). a. 317 BraY iv.47ab: a

saktih sad siva caiva saktitrayam tu eva ca | a s . . srkan. ha ca tath dev um patis tath parah 4 . . t s a a a . ete divy dhik h prokt s tath divyam srnu priye | a a. a a . . . a. yoginyo guhyak dy s ca m try s t s ca sakinkar h 5 a a a a a guhyak pataya caiva vr s caiva prati prati | a s a as. asas. hi tath rudr yoginyah ksetrasamsthit h 6 a a a. .t . .t . . . lokap l s tath caiva ete divy prakrtit h | aa a a a.
4d um patis ] em.; um pati Bya a a 5a dhik h ] corr.; dhik Bya a. a 5d m try s ] em.; m try m a a a a. Bya sakinkar h ] em.; sakinkar m Bya 6a guhyak pataya ] conj.; guhyak m pataya Bya 6d a. a. a s a. s

125 yogins bear general comparison to the extant yogin statuary. The names of the pri mary Six Yogins aloneKrosthuk (Jackal Woman), Vijay (Victoria), Gajakar a .. na (Elephant-ears), Mah mukh (Big-mouth), Cakraveg (Wheel-speed), and a a . Mah n s (Big-snout)suggest much of the iconographic range of extant images, a a a including theriomorphism, auspiciousness, power, and the grotesque. As with the other divine images, all are four-armed k p lika deities bearing the Five Insignia, a a skull-staves, skull-bowls, and rosaries, raising a right hand in the gesture of municience. Garlands of skulls extend to their feet, and they stand upon human corpses. In contrast to the other mandala goddesses, yogins have only a single facethe norm .. in sculpted images.318 One contrast with the extant statuary lies in the disproportionate body parts suggested by some yogin names in the BraY , such as Mah n s a a a a
yoginyah ] em.; yoginyo Bya .

BraY iv.170cd78ab, 254cd55ab, 263cd64: a divyam caiva pravaksy mi yath vad vaktrasamkhyay 170 a a . . a . . a. guhyak nucar ye tu kinkarnan tu kinkar h | a a yoginpataya caiva m tnam patayas tath 171 s a r. . a . pacavaktr h sam khy t vrabhakty s tathaiva ca | a. a aa a guhyak s tu caturvaktr h kinkaryas trimukh h smrt h 172 a a. a. . a. yoginyas tv ekavaktr s tu caturvaktr s tu m tar h | a a a a. divy dhik s tath divy h pacamudr samanvit h 173 a a a a. a a. khatv ngamundadh rinyo mah pretakrt san h | a . a . a .. . a a. varadodyatahast s tu caturhast h prakrtit h 174 a a. a. aksasutradhar h sarv h kap lakarak s tath | a. a. a a a . daksine tu kap lam sy t khatv ngam v matah sthitam 175 a . a . . . a . a . daksinam varadam jeyam v mahaste ksasutrakam | . . . . . a . . a. kap lam l bharana ap d t kan. ham asrt h 176 a aa . a a .t evamvidh s tu kartavy divy divy dhik s tath | a a a a a a . ekavravidh ne tu pacavaktr s tu guhyak h 177 a a a. m tara ca tath caiva kartavy s dhakottamaih | a s a a a . ... kros. huk vijay caiva gajakarna mah mukh 254 a a .t . cakraveg mah n s sad yoginyah prakrtit h | a a aa. . a. . ... yoginyah svetavarnas tu sakhakundendusaprabh h 263 a. . . pn vayavasampurnah stanapnapayodhar h | a a. . . . kum rvratadh rinyo aklinnaratayas tath 264 a a . a
171b kinkar h ] Byapc ; kinkar Byaac 171c yogin ] corr.; yogin Bya 172a vaktr h ] corr.; vaktr a. a a. a . Bya 172c vaktr h ] corr.; vaktr Bya a. a 173c divy h ] corr.; divy Bya a. a 173d samanvit h ] em.; samanvit Bya a. a 174b asan h ] em.; asan Bya a. a 175c tu kap lam a . a (tops cut off ) sy t ] conj.; tukap la(sy ?) By a a a 175d khatv ngam ] em.; khatv nga Bya (tops cut off ) . a . . a 176c haste ksa ] em.; hastemksa Bya 177b tath ] Byaac ; tath h Byapc a a. 255b sad ] corr.; sat . . . . . . . Bya 263c sveta ] corr.; sveta Bya 264a sampurnah ] corr.; sampurna Bya . .. . .

126 (Big-nose); animal features are depicted in sculpture, but not, that I am aware of, disproportionate human mouths, ears, and so forth. Otherwise rich in detail, BraY iv nonetheless leaves many questions unanswered a concerning the architectural and ritual contexts of religious images. Their scale suggests grand structures; the iconometric description of divine images, which include those of the Six Yogins, approaches the human: goddesses possess feet twelve angulas (nger-breadths) long,319 hips thirty-four nger-breadths wide,320 and throats ve nger-breadths in width,321 for instanceonly marginally smaller than the ve-faced supradivine deities. Yet this chapter is silent on temples, and the material elsewhere in the text claries little. The instructions for image making are decidedly oriented toward the BraY s own esoteric pantheon, hardly a range of a deities found in large public temples; yet at the same time, among the supra-divine .. images are deities of no cultic consequence in the BraY Um pati and Srkantha. a a Religious images are clearly a signicant concern, but their roles and context require further inquiry. Certainly the presence of the mantra-deities in a substrate is central to ritual: thus the ubiquitous instruction to perform ritual action before/in the presence of the goddesses (devn m agratah). Yet this need not, and usually clearly does a . not, refer to religious images.
the problems of ritual and sectarian affiliation

Much as the architectural and ritual contexts for the religious images the BraY dea scribes remain unclear, so too the cultic context of the yogin temples. Attempts to reconstruct the ways in which yogin temples served as cultic spaces have, given the paucity of textual descriptions, involved reading ritual from iconography. Thus Dehejia, remarking upon the presence of corpses and severed heads in the iconography of some yogins, wonders, is this all an indication of human sacrice, or is it in
319 BraY iv.157abc: p dau drghena c khy tau kal h [em.; kal m Bya ] tu satpram natah | p rsny s [conj.; a a a. a. a. . a . . a . a a . . p sny Bya ] c ngus. akam y vat. a. . a a .t . a 320 BraY iv.163ab: nitambas tu tath proktah [em.; prokt m Bya ] kal saptadaas tath . a a a. a s a . 321 BraY iv.168cd: kan. has tu protthatah proktah [em.; protthato prokt h Bya ] s rdham caiva kal dvayam. a a. a a .t . . .

127 stead a pointer to sava-s dhana or corpse ritual?322 Her general conclusion is that a the yogin temples served primarily as sites for Kaula worship involving sacrice, consumption of wine, blood, and esh, and ritual copulation, suggesting that the temples would have invited lay devotional worship as well.323 In contrast, Ronald Davidson questions the Kaula connection. The grounds for this skepticism include the curious claim that most Kaula works appear composed after the sites were constructed; he argues that the primary activity depicted at these sites . . . is the display of severed heads, indicating that the sanguinary rites were probably the principal activity practiced.324 It should be pointed out, however, that the image he captions with Yogin from Hirapur displaying human sacrice does not depict the act of sacrice, but a well-armed goddessnot in fact one of the yoginsstanding upon a smiling, severed head.325 Sacrice might well have taken place in connection with the Hr pur temple; but to argue this on the basis of iconography of a variety ema blematic of extreme tantric deities, both Buddhist and Saiva, seems uncompelling, all the more so given that such iconography is not particularly prominent in the Hr pur a yogins. While Kaula texts offer insight into the nature of the divinities enshrined, and attest a tradition of iconic representation of yogins, it is unclear to what extent the temples should, in a meaningful cultic sense, be described as Kaula. Despite their novel architectural forms, it remains entirely possible that the liturgy of the yogin temples differed little in basic character from that of contemporaneous temples of other goddesses, or Bhairava. For although possessing historical links with esoteric Saivism, period literary sources such as the Kath sarits gara point toward the growa a
322 Dehejia, Yogin Cult and Temples, 59. By ava-s dhana, a term I have encountered only in late s a medieval sources, Dehejia refers to rites of the variety sometimes referred to as vet las dhana; on the a a latter, see the discussion of the Harsacarita earlier in this chapter. . 323 Ibid., 186. 324 Indian Esoteric Buddhism, 1801. 325 Ibid., figure 7. This is the very image I include as figure 2.18: one of the nine goddesses upon the exterior facade of the temple, which as mentioned earlier have been identied as nine k ty yans. a a The depiction of jackalsemblematic of the cremation groundon either side of the severed head reinforces the k p lika rather than sacricial nature of the iconography. a a

128 ing prominence of the yogin in the religious landscape of medieval India, while bur geoning material on yogins in pur nas suggests worship beyond the narrow connes a. of the tantric traditions. Dedicated to divinities widely regarded as potent agencies, the yogin temples need not have been centers for the activity of initiated Kaula spe cialists, although it is entirely possible that worship of the goddesses was presided over by ofciants with links to esoteric Saivism. Despite their hilltop locations, the yogin temples appear by and large to have been prominent, visible monuments, as evidenced by their proximity to other major state-sponsored templesespecially those of Khajuraho and R npur-Jharial.326 Along with the increasing signicance a of yogins in the pur na corpus, the yogin temples in fact appear to mark the entry a. of these deities into a wider religious domain, beyond the connes of the esoteric traditionto the point that their ritual mandalas are translated into monumental .. circular temples.



The foregoing chapter has focused upon the background and early evidence of the Yogin cult in non-tantric literature, inscriptions, and the art-historical record. In the course of presenting this material, I have attempted to establish a chronological framework and to place these varieties of evidence in relation to tantric literature, especially the BraY . In the subsequent chapter the focus shifts to yogins in tantric a literature proper, prior to undertaking more detailed examination of the BraY in a chapters four and ve. It was shown that the cult of yogins, as described in tantric literature, pre supposes the Saiva cult of the Seven Mothers (saptam tr), a development that arta. historical and epigraphic sources situate in the fth century, with possible fourthcentury precedents. The roots of this heptad of Brahmanical goddesses lie in ancient
Hilltop temple sites include those of Mitauli, R npur-Jharial, Bher gh t, Khujaraho, Dud h, and a a a a probably Lokhari; on the latter two see Dehejia, Yogin Cult and Temples, 141, 156. The Hr pur temple a appears to be neither on a hill nor near other major temples; but as mentioned, it is in the vicinity of the major pilgrimage site of Bhuvane var/Ek mra. s a

129 traditions of maternal deities possessing close ties to the natural world, fertility, and death. Although the Mothers have a close connection with Skanda in early sources such as the Mah bh rata, Siva largely displaces this deity as companion of the Seven a a Mothers by the sixth century. Yet as the early fth-century Gangdh r inscription ila lustrates, a tantric or proto-tantric cult of Mother goddesses and female spirits might already have been in existence, possibly centered upon the mysterious, k p lika goda a dess C munda. It is in the context of describing the greatness of Kotivarsaa place a . . . . sacred to the Mothers and presided over by C munda, as Bahum msa, and Siva, a a. . . as Hetuke varathat the old Skandapur na (circa 6th7th century) provides momens a. tous references to the Saiva yogin cult and its early scriptural sources, including the BraY . a The Saiva temple cult of the Mothers becomes prominent from the sixth century, and the inclusion of Mah k la in m tr-shrine iconic programs from the late sixth cena a a. tury might reect developments in Tantric Saivism, for tantric cults of Mah k la and a a the erce Goddessidentied with C mundand unambiguous attestation in the a . . early seventh-century works of B na and V kpati. In these works, evidence for the a. a ritual of the bhairavatantras is abundant, yet there are only vague suggestions of the cult of yogins. The most signicant textual account of yogins from this period re mains therefore the Skandapur na, which attests the Saiva cult of yogins and Mothers a. and provides a list of Tantras of the Mother-goddesses (m trtantra). That a Saiva a. Yogin cult of the variety attested in Vidy ptha literature existed by the eighth cen a . tury receives conrmation in Bhavabhutis M latm dhava, a work displaying detailed a a knowledge of Vidy ptha ritual systems, if not specic texts. On the other hand, the a . Haravijaya of the early ninth century and Yaastilaka of the mid-tenth show knowledge s of a particular tradition: the Trika. The sources reviewed attest to the growing prominence of yogins and their cult in the religious landscape of medieval India, especially from the tenth centurythe period in which temples of these goddesses were constructed throughout India. Ex-

130 amination of the Brhatkath corpus illustrates this historical trajectory, for only in a . the eleventh-century redactions of Kashmir does the gure of the yogin become a signicant locus for tales of magical women and demi-goddesses. This period is largely beyond the scope of the present dissertation, and the subsequent chapters are concerned primarily with pre tenth-century tantric literature. I have nonetheless attempted to show, cursorily, that the yogin temple cult appears to draw upon Kaula traditions of sixty-four yogins, although iconic representation of yogins has precendent even in early Vidy ptha literature. By the period of the temples and lata . ter Kaula literature, yogins appear to have become signicant sacred gures beyond the connes of the esoteric traditions, as reected in the material devoted to them in pur nas of the second millennium. a.

Chapter 3
The Cult of Yogins and its Background in Early Tantric i Literature



Beyond the literary, epigraphic, and sculptural evidence for yogins in early me dieval India reviewed in the previous chapter, there survives a substantial body of pre eleventh-century tantric Saiva literature devoted in various degrees to their cult. The bulk of this material remains unpublished; and while fortunately extensive, the extant texts represent only a fraction of what might once have existed. Complement ing the textual corpus of the Saiva yogin cult, there survives another large body of yogin-cult literature as well: the Buddhist scriptural sources that came to be classi ed as yoganiruttaratantras (The Ultimate Tantras of the Yoga Class) or yogintantras ( Tantras of the Yogins), upon which there also survives a considerable quantity of exegetical literature. The present chapter comprises a discussion, far from exhaus tive, of the background of the Yogin cult in early Saiva and Buddhist textual sources, and a (rather cursory) review of actual tantras of the Yogin cult. As with the previous chapter, the focus remains upon the early evidenceprimarily pre tenth-century with the BraY remaining a constant point of reference. Chapters four and ve will a then focus upon detailed examination of the BraY itself. a The nondualist Saiva exegetical and philosophical literature that ourished from the tenth century, particularly in Kashmir, draws on an enormous and diverse corpus of tantric scripture, a canon that must have developed over the course of mul131

132 tiple centuries. In his denitive review of the available early evidence, Sanderson concludes, it is quite possible that by the seventh century most of the literature available to Saiva scholars in the tenth century was already in existence. But it is not until the beginning of the ninth that we have rm evidence of specic texts.1 Many of the sources that come into evidence in this period are siddh ntatantras of the cult a of Sad siva, and have little direct relevance to yogins. Exceptional in this regard is a the ancient Niv satattvasamhit , a text which, although not directly concerned with s a . a yogins, provides clues concerning their roots in earlier forms of Tantric Saivism. The present chapter begins with review of the Niv sa. Subsequently, the discussion turns s a to varieties of tantric literature highly signicant to the development of yogin tradi tions, but poorly represented in the surviving literatureparticularly the v matantras a of the cult of the Sisters (bhagin) of Tumburu. The earliest extant tantras of the Yogin cult belong to the corpus of bhairava tantras, scriptures centered upon Siva as the archetypal skull-bearing ascetic (kap lin), a Bhairava, as well as allied goddesses. One of the earliest of the bhairavatantras appears to be the Svacchandatantra, which, as will be discussed, attests the cult of yogins only in its nal chapterprobably a late addition to the scripture. Following discussion of the Svacchandatantra, section three addresses the Saiva tantras of the Yogin cult. These belong to two primary categories: those identifying themselves as scriptures of the Vidy ptha, The Seat of Wisdom-[goddess] Mantrasa division of the bhairavaa . tantrasand the diverse corpus of Kaula scripture: tantras Of the [Goddess] Clans. The earliest attested yogintantras, including the BraY , belong to the Vidy ptha, a a . while on the other hand, the bulk of surviving Saiva literature concerned with yogins belongs to various Kaula systems.
History through Textual Criticism, 18. Sanderson has compiled a list of the sources cited by circa tenth to eleventh-century Saiva authors, and also has identied the extant tantric scriptures we can infer, on the basis of datable references or manuscripts, were in circulation in the ninth century. For the texts of which early Kashmirian authors show that they have direct knowledge, see ibid., 34 (n. 1). This list supercedes those of Navjivan Rastogi, who enumerates the works Abhinavagupta cites in the Tantr loka, and those Jayaratha refers to in his commentary thereon. Introduction to the Tantr loka: a a A Study in Structure, 25383; cf. 28485. For Sandersons list of the texts probably in circulation in the ninth century, see History through Textual Criticism, 48 (fns. 25).

133 In the fourth section of this chapter, I attempt to trace the emergence of the Buddhist cult of yogins. Through analysis of tantras of the Buddhist Path of Mantras (mantranaya), it is shown that the Yogin cults development can be linked to the in creasing prominence of Mother goddesses and a variety of other female deities and spirits, primarily as reected in the literature classied as yoga- or mah yogatantras. In a particular, the emergence of a tantric Buddhist cult of yogins appears closely tied to the Buddhist conversion and transformation of the dakin, a female being guring . as a pernicious and lowly variety of yogin in Saiva typologies of goddesses. In Bud dhist yogintantras, the term dakin becomes entirely synonymous with yogin. . The nal section of the chapter takes up an aspect of the complex problem of the relation between Saiva and Buddhist yogin traditions. An analysis of paral lel passages in the BraY and the Buddhist Laghuamvaratantra, rst identied by a s . Sanderson, conrms the latters ndings concerning the Saiva sources of this Buddhist yogintantra.


the background of the yogin cult in tantric saiva literature i

the ni vasatattvasamhita s .

Preserved in a Nepalese codex copied most probably in the ninth century,2 the Niv satattvasamhit has been recognized on strong grounds as being among the ears a . a liest surviving texts of Tantric Saivism, and perhaps the most ancient of all. Although a scripture of the cult of Sad siva and subsequently classied as a siddh ntatantra, the a a
2 nak 1-227 (ngmpp reel a41/14). Two apographs of this codex also survive: nak 5-2406 (ngmpp reel a159/18), and Wellcome Institute Sanskrit ms i.33. See Dominic Goodall and Harunaga Isaac son, Workshop on the Niv satattvasamhit : the Earliest Surviving Saiva Tantra?, Newsletter of the s a . a NGMCP 3 (JanFeb 2007), 4. Sanderson assigns the Ni v sa manuscript to approximately 850900 c.e. s a The L kulas: New Evidence of a System Intermediate Between P c rthika P supatism and Agamic a a a a Saivism, Indian Philosophical Annual 24 (2006), 152. Cf. Teun Goudriaan and Sanjukta Gupta, Hindu a Tantric and S kta Literature, 34. Unless otherwise noted, I cite the text of the Niv sa from transcriptions s a of the aforementioned manuscripts, as circulated among the participants of the Workshop on Early Saivism: the Testimony of the Niv satattvasamhit , Pondicherry, Ecole francaise dExtrme-Orient, s a . a January 2007. Those involved in preparing the transcriptions are mentioned in Goodall and Isaacson, Workshop on the Niv satattvasamhit , 5 (n. 4). Note that in speaking of the Niv sa, I exclude from s a s a . a consideration the Niv sak rik , which appears to be a late supplement to the Niv sa-corpus. s a a a s a

134 Niv sa appears in fact to predate Tantric Saivisms bifurcation into the Sad siva cult s a a of the siddh ntatantras, and the non-Saiddh ntika traditions, of which the cults of a a Bhairava and related goddesses form the primary division.3 The religious world of the Niv sa is undoubtedly far removed from the Vidy ptha cults of Bhairava and yos a a . gins, deities of powerful mortuary iconography whose siddhi-oriented worship was often radically antinomian in character. Yet in comparison to the later Saiddh ntika a tradition, the gulf between the Niv sa and early Vidy ptha sources appears less pros a a . nounced. In particular, the Niv sas fth and largest book, the Guhyasutra (hereafter s a Niv saguhya) contains a wealth of siddhi-oriented ritual presaging themes central to s a non-Saiddh ntika magical traditions. This material appears to afford a window a into the formation of characteristic ritual forms of the bhairavatantras, and is hence relevant for inquiry into the roots of the Yogin cult. The following pages examine the Niv sa from this perspective, primarily with reference to the BraY , as representive s a a of the early Vidy ptha. a . The Niv sa places little cultic importance upon goddesses. It makes reference s a to worship of the Great Goddess (Mah dev), yet only in the context of describing a non-tantric, lay religious practices (laukikadharma).4 The goddesses K l and Vijay a a also nd passing mention, but in a context of no cultic consequence.5 In the domain
One of the most compelling arguments for the antiquity of the Niv sa concerns its considerable s a continuity with the early, non-Tantric Saivism of the Atim rga, on which subject see Sandersons pioa neering study, The L kulas: New Evidence. See also Sanderson, History through Textual Criticism, a 2931. Another, related indication of the early date of the Niv sa concerns its unique position within the s a corpus of Saiddh ntika scriptures: although later included in canonical lists of siddh ntatantras, the text a a shows no awareness of this classicatory category nor of any other division within Tantric Saivism, and is in numerous ways highly anomalous. Sanderson, ibid., and Goodall and Isaacson, Workshop on the Niv satattvasamhit , 6. Although the dating of the Niv sa corpus remains an open question, Goodall s a s a . a and Isaacsons preliminary assessment would place the earlier parts of the text between 450550 ad. Ibid. 4 Mukh gama (hereafter Niv samukha) 3.103cd107 describes worship of the Mah dev (mah devy s a s a a a a tu pujanam, 103d). Present here is an unmistakable and perhaps comparatively early articulation of the idea of the Great Goddess, with Um , the spouse of Siva, heading a list of the following names a and epithets: Um , K ty yan, Durg , Rudr , Subhadrik , K lar tri, Mah gaur, Revat, Bhutan yik , a a a a a a a a a a a a Ary , prakrtirup (She Who Takes the Form of Prakrti), and ganan m n yik (Leader of Sivas Troops) . . a . a. a a (104cd105). Noteworthy is the identication of the Mah dev with the prakrti of S mkhya, and absence a a. . of explicit identication with sakti and m y . a a 5 In Uttarasutra (hereafter Niv sottara) 1.34d, K l and Vijay are listed among the interlocutors s a a a of tantric literature, alongside various gods, Sivas ganas, and sagesas well as Mother goddesses, . guhyakas, and other divinities, a number of which seem unlikely to gure in the revelation of a

135 of ritual, a triad of Sivas female powers (akti) almost ubiquitous in later Saivism s does gure: V m , Jyesth , and Raudr.6 Another common set of saktis appears as a a .. a well: the series of nine beginning with V m and ending with Manonman, who here a a as well as in the later tradition have place in the inner circuit of deities in the mandala .. of Sad siva.7 Present in the Niv sa, furthermore, are key elements of the theology of a s a sakti familiar from later sources, such as the cosmogonic role of the supreme Sakti and the function of the descent of Sivas power (aktip ta) in grace and initiation.8 Bes a sides the aforementioned saktis, goddesses appearing in the Niv sa as tantric deities s a are to a large extent apotheoses of feminine principles, such as V gsvar, goddess of a speech; Susumn Dev, the central channel of the yogic body; and tattvas with femia .
siddh ntatantra. This, and a vague reference to a derivative compilation genre of scripture (yo nyah a . parataro bhedo sivatantresu pathyate | sangrahah sa tu vijeyo, Niv sottara 1.32abc), might suggest the ex s a . . . istence of diverse tantric literature in the period of the Niv sas composition. The list of interlocutors s a does not suggest the existence of bhairavatantras or v matantras, however. a 6 As described in Niv samula (hereafter Niv samula) 5.34, the initiation of the liberation seeker s a s a (muktidksa) involves linking/fusing (yojana) the initiates soul to tattvas via the saktis V m , Jyesth , . a a .. a and Raudr, although the mantras of the three are not there explicated. Elsewhere in the text, cf., e.g., Niv saguhya 7.260ab (v m jyes. h ca raudr ca saktitrayam atah param). Remarkably, the well-known triad s a a a .t a . of j naakti, kriy sakti, and icch sakti seems absent as such. There does occur reference to icch sakti a s a a a (Niv saguhya 8.79a), and j naakti and kriy sakti nd mention as a pair (Niv saguhya 7.260cd, and s a a s a s a probably 8.65b), but these are not, it seems, linked as a set of three. This might lend plausibility to Hlne Brunners suggestion that the pair j naakti and kriy sakti underlies the later groupings of a s a a three or more. See J na and Kriy : Relation between Theory and Practice in the Saiv gamas, in a a Ritual and Speculation in Early Tantrism. Studies in Honor of Andr Padoux, edited by Teun Goudriaan, 17. 7 Describing the basic mandala for worship of Siva (iv rcanavidhi, 2.1b), Niv samula 2 enjoins s a s a .. installing the Nine Saktis upon a white lotus (tasyopari sitam padmam navaaktisamanvitam, 2.2cd). This s . . appears in continuity with the later tradition, in which the Nine are installed on a lotus of eight petals, forming the inner layer ( varana) of the mandala of Sad siva; see Goodall, et al., The Pac varanastavah of a a a .. . . . Aghoraiv c rya: A Twelfth-century South Indian prescription for the visualization of Sad siva and his retinue, s a a a gure 4. Note also that Niv saguhya 8.65a refers to the ninefold sakti pantheon/worship (navadh s a a saktiy gam). a . 8 On the cosmogonic role of the supreme Sakti, see especially Niv sottara 1.56, where is attested the s a notion that Sivas sakti, possessing his power, gives rise to bindu, the material cause of the upper levels of the universe (tasya saktih siv nity [conj. (Diwakar Acharya); siv niry Cod.] sivatejopabrmhit | sivatejena a a a . a .. a samyukt sakter j yati bindukah, 1.5cd6ab). Nayasutra (hereafter Niv sanaya) 1 connects saktip ta with a a s a a . . Sivas grace (anugraha) and with initiation, graces quintessential expression in Tantric Saivism: sivaaktinip tena dksa[j ]nam prayacchati 88 s a . a . so nugrahah smrto [hy eva d ]t caiva sad sivah | a a a . . . Through the descent of Sivas power, he bestows initiation and the scriptural wisdom. It is this which is known as grace, and its giver is Sad siva. a (The codex, nak 1-227 (hereafter referred to as A41/14 ), is not fully legible here; the bracketed syllables are transmitted by its two apographs.) Note also that in Niv sanaya 4, the guru (deika) whose initiation s a s is efcacious is described as sivaaktyopabrmhitah, brimming with Sivas sakti (4.41b). s .. .

136 nine gender names, such as vidy and m y .9 Despite the presence of these goddesses, a a a their roles appear largely ancillary in the Niv sas ritual systems, and much of the s a material concerning saktis belongs to the Guhyasutra, probably a comparatively late stratum of the text.10 There alone do we nd allusion to a different sort of goddess, the erce Candik : the text gives in passing an otherwise anomalous candimantra .. a .. and its application, alluding to a tradition of ritual centered upon this deitythe goddess to whom the Dravidian t ntrika of B nas K dambar professed devotion.11 a a. a As for the Mother goddesses so central to the development of the Yogin cult, the Niv sa describes them as cult deities only within the sphere of public, lay religion s a (laukikadharma)not as tantric mantra-deities.12 Like numerous other Saiva sources, the Nivasa also lists temples of the Mother goddesses among the places appropriate s for performing solitary tantric ritual, along with crossroads, jungles, mountain peaks, Siva temples, and so forth.13 It appears to be the liminality of the Mother temple
The Niv samukha, for instance, twice describes V ge var/V ge (i.e. V gs) as the source (yoni) s a a s a s a of the pranava, viz. 4.94ab and 4.125cd. An intriguing prescription for meditation upon the goddess . Susumn is given in Niv saguhya 7.29398; she is said to be white, with the shape of a lotus stalk, and a s a . a . . a to emerge from the body of Siva (padmasutr krtih sukl , 294c; sivak y d vinihsrt , 297b). As for tattva a a . . a goddesses, Niv sanaya 3 describes meditation upon the series of tattvas as male and female deities, s a beginning with the goddess Prakrti in 3.20. See also Niv saguhya 7 (especially 219252) and 8 (52 s a . 57); in the latter section on the subject of tattvayojana, note worship of M y (8.52c) and the goddess a a Vidy (vidy dev, 8.53c). a a 10 Here I follow the working hypothesis on the stratication of the text put forward by Goodall in a presentation entitled, The Structure of the Niv sa-corpus, at the Workshop on Early Saivism: the s a Testimony of the Niv satattvasamhit held in Pondicherry, January 2007. s a . a 11 The passage giving the mantra is unfortunately lacunose; A41/14 reads, om candike krama . . . can. .. . dimantro yam (3.25 ab). . 12 Niv samukha 2.28 lists the Mothers among cult deities of temples (pr s da), while 3.3334ab refers s a aa to devotees of and places sacred to the Mothers and other divinities. Both of these passages occur within the description of laukikadharma (Niv samukha 23). s a 13 Viz. Niv sottara 2.4c and Niv saguhya 6.32. These bear comparison with the lists of suitable s a s a locations for s dhana provided in later sources, e.g. Siddhayogevarmata 6.24, or BraY xviii.1819, the a s a latter of which, however, omits reference to Mother temples: smaane ekadee v nadtre viesatah | s s a s . . parvat gre samudre v ekalinge catuspathe 18 a a . nagare gr madee v rathy y m gopure tath | a s a a a. a ekavrkse visesena kalpoktam tu sam caret 19 a . . . . .
sma ane ] corr.; smas ne Bya s a

In a cremation ground, () a solitary place (?), the bank of a river, particularly, the top of a mountain, the ocean, a solitary linga, a crossroads, a town or village, a road or a town gate, or a solitary tree, in particular, one should practice what is stated in the kalpa-text [for propitiation of the mantra]. Cf. BraY xiv.11cd13ab, which adds empty houses, old wells, ant hills, and Mother temples to the list a

137 presumably a secluded shrine rather than public templethat makes it suitable for the rituals envisioned, for these do not directly involve worship of Mother goddesses. In the Niv sa, there is in fact only limited evidence for the Saiva appropriation of the s a Mothers, and this occurs in the context of cosmology rather than ritual. Chapter ve of the Niv saguhya lists the Mother goddesses among other lords (patayah) of a series s a . of seven p t las, netherworlds, together with deities such as rudras, ganas, n gas, aa a . r ksasas, bhagins, and yogakany s.14 None of these are presented as tantric deities a . a proper, that is to say, mantra-deities. Several of the Niv saguhyas netherworld divinities warrant discussion. By and s a large, the cosmological spheres described are populated with male mantra-lords (mantrevara, etc.) and manifold rudras. While in later Vidy ptha accounts of the s a . cosmos, goddesses would largely eclipse male deities, in the Niv saguhya the gods a desses of the netherworlds are exceptional. In the lowest p t la are present groups aa of Mothers (m trganah) and Sisters (bhaginyah).15 It seems unlikely that these rea . . . . fer to their most famous representatives, the Brahmanical Seven Mothers and the Four Sisters of Tumburu; most probably are intended amorphous bands of minor female deities. Two other sets of Mother goddess nd mention: tawny (pingala) Mothers who bear blue lotuses in the third netherworld, while in the fourth preside kap lam trs, Skull Mothers. The latter, who have parallel in the kap larudras of the a a. a next higher cosmological sphere, appear to represent a transformation of the Mothers into Saiva, tantric goddesses, whose k p lika iconography presages the image of the a a yogin.16 Positioned even higher in the series of netherworlds are yogakany s, yoga maida ens or daughters of Yoga, deities of the sixth and seventh p t las. Here described aa merely as possessing great power (mah vry h),17 goddesses by this designation a a.
of places suitable for worship of the kulavidy mantra. a 14 Niv saguhya 5.121. s a 15 anye m trgana rudr bhaginya ca ganas tath (5.6cd). a . . a s a . 16 a See Niv saguhya 5.9ab (m tar [h] pingal yatra vasanty utpalahastik h), 5.11cd (kap lam taro yatra s a a a . a. a a . . . [lacuna]), and 5.14ab (tath kap larudr s ca asamkhyey h puna[h] sthit h). a a a a. a. . . 17 yogakanyak s are mentioned in Niv saguhya 5.15d, in a description of the city Ratnavat of the sixth a s a

138 are little attested in Saiva textual sources familiar to me. Yet as powerful, youthful goddesses connected with yoga who transcend the Skull Mothers, these yoga maidens might have continuity with the deities later referred to as yogevars, fes male masters of yoga, or yogins. This connection is in fact drawn much later by a Kashmirian, Ksemar ja, in commenting upon a parallel passage in the Svacchandaa . tantra.18 Unfortunately, the Niv sa provides little material for further exploration s a of this possibility. Nonetheless, this cosmological system describes a hierarchy of divinities with goddesses suggestive of those later brought within the rubric of the yogin: multiple categories of Mother goddess, the Sisters, and maiden goddesses born from or possessed of yogic powers.19 Although the cosmology of the Niv saguhya only faintly suggests developments s a relevant to the formation of the Yogin cult, its ritual offers more substantive material for comparison. It is rich in siddhi-oriented practices that foreshadow varieties of ritual elaborated upon in the bhairavatantras. Particularly noteworthy are its mortuary (k p lika) practices, the use of magical substances (siddhadravyas), and instances of a a sexual ritual. The k p lika rites of the Niv saguhya appear largely magical in oriena a s a tation, which places them in much closer relation to those of the bhairavatantras than the liberationist k p lika ritual of the P supatas. This is particularly evident in the a a a Niv saguhyas prescriptions for preparing magical substances in a skull; in one case s a the end-product is an ash that would turn the practitioner into a vidy dhara, and in a another an eye-ointment that induces invisibility. Similar procedures are common
p t la; other denizens include atharvarudras (atharvarudr s tatraiva vasante yogakanyak h). Yoga maidens aa a a. possessed of great power (yogakany mah vry s, 5.19a) are also met with in a description of the seventh a a a p t la, along with sons of Agni and V yu (agnikum rak h, 18d; v yukum rak h, 19b). Note that the aa a a a. a a a. fth p t la contains agnikum rik s, female counterparts of the sons of Agni. aa a a 18 See below, n. 59. 19 Cf. BraY lv and lxxiii; both chapters contain detailed typologies of yogins, and are edited in a the present dissertation. Compare also the list of potentially malevolent divinities in Netratantra 2, the female ones among these being hordes/groups of sakins and yogins; the bhagins, rudram trs, a. a a etc.; davs, damarik s, and rupik s ( sakinyoginganaih | bhaginrudram tr didavdamarik dibhih rupik bhir a . . a a . . a . . . . . . , 13b14a). This list hence includes both Mother goddesses and Sisters, the latter, according to Ksemar jas commentary, originating from partial incarnations of [the Seven Mothers,] Br hm, etc. a a . (brahmy dyamsakotth bhaginyah). a a . .

139 in the BraY , where however more marked use of the impure is made.20 Perhaps a the most extreme of the k p lika siddhi-practices taught in the Niv saguhya is chapter a a s a threes rite of re sacrice, performed in the mouth of a corpse, which in structure and aim parallels vet las dhana as described in the BraY and Harsacarita. As in the a a a . BraY , the rite culminates with the corpses tongue emerging, which when severed a becomes a magical sword.21 Elsewhere, the Niv saguhya describes rituals for mags a
20 A process for producing invisibility-inducing kohl in a human skull is given in Niv saguhya 3.81 s a 82. The recipe involves nothing more offensive than ghee. The process in 11.110 utilizes ash:

om yog dhipa namah | anena mantrena bhasma grhya kap lasampute sth pya t vaj japed y vad a a a a a . . . . . vartitamatih tatas tu tenoddhulane vidy dharo bhavati | . . . a a . om yogadhipa namahtaking hold of ashes using this mantra and placing them in the . . hollow of a skull, one should repeat the mantra until () . . . (?). Then, when one is dusted with these ashes, he becomes a vidy dhara. . . . a Compare with BraY xlix, which taps the powers of considerably less innocuous substances: a kros. hukasya tu piitam harit lamanacchil m | s . a a .t rocan ca mah m msam ekkrtv tu psayet 8 a a a. . . a . kap lasamputam krtv atmaraktena miritam | a s . . . . a sahasr s. adhikam japtv trskrtv tilakam kuru 9 a. t a .. . a . . bhavate bhutale siddho adreyah k lav sinah | s . a a .
8b manacchil m ] em.; manacchil Bya 8c rocan ] em.; rocan Bya a a a a Isaacson; sahasr nt dhikam Bya 10b adre yah ] em.; adre yoh Bya a a s . s . . 9c sahasr stadhikam ] em. a. . .

One should mix together and make a paste of the esh of a jackal, the harit la and a manahsil minerals, yellow pigment, and human esh. After placing this in the hollow . a of a skull mixed with ones own blood, and reciting the mantra one-thousand and eight times, make a bindi [with this] thrice. He becomes perfected on this [very] earth, invisible, having power over death. This brief chapter, the kros. ukakalpa, has as its theme magic using jackal (kros. uka) esh. .t .t 21 Guhyasutra 3.60cd64ab: ekalinge ekavrkse smaane samgame vane 60 . . s . tatra mandalam alikhya tattv ngabhuvanam subham | a .. . aksatam mrtakam grhya sth payitv viankitah 61 a a s . . . . . . . susn pitam ca lipt ngam puspasragd mabhusitam | a a . a . . hrdi tasyopavis. am tu tattvena sravitam carum 62 . .t . . tasya vaktre tu hotavyam satas hasra-m- yutam | a a . tato jihv viniskr met t m tu mantrena cchedayet 63 a a. . a . s jihv bhavate khadgo guhyavi | a a . an hatagatih so hi jved acandrat rakam 64 a a .

61a tatra mandalam ] conj.; - - - - - m A41/14 61d vi ankitah ] conj. (D. Acharya?); sankitah s .. . . A41/14 62a lipt ngam ] em.; lipt nge A41/14 62b puspasragd ma ] Kathmandu apograph; pu - a a a . . - ma A41/14 64a khadgo ] em.; khadg A41/14 64c an hata ] Kathmandu apograph; - n hata a a . . a 41/14 41/14 A 64d jved ] Kathmandu apograph; - ved A

At a solitary linga, a solitary tree, or in a cremation ground, conuence [of rivers], or forest, one should draw the auspicious mandala which houses the tattva-mantra and its .. ancillaries. One should take an unmutilated corpse and place it there, without hesitation, one well-washed and with oiled limbs, decorated with ower garlands. Seated upon its chest and using the tattva-mantra, the melted (? sravita) food offering (caru) should be given in re sacrice [into a re] in its mouth, one billion times. Then [its] tongue would

140 ically enlivening a corpse that does ones bidding.22 Such magic was undoubtedly ancient, described for instance in the Vasudevahindi, as mentioned previously. .. As for sexual ritual, the Niv saguhya teaches an asidh r vrata (the razors edge s a aa observance) that involves the participation of a beautiful young woman. She is to be skilled in the erotic arts, although one who succumbs to lust in her embrace falls into hell. The description is brief; but it is worth noting that a full chapter of the BraY , its thirty-ninth, is devoted to a rite of the same name almost identical in a nature, though spelled out with attention to much more intimate details.23 Absent
emerge. One should sever this with the mantra. That tongue becomes a sword . . . With unimpeded movement, he lives as long as the moon and stars. mss as reported in the transcription of Goodall, et al. (see note 2 above). satas hasram yutam in 63b a a appears to mean satas hasr yutam, one-hundred thousand ten thousandsa case of sandhi-breaking a a -m- as well as metrical lengthening of ayuta to ayuta, it seems. However, ayuta might be a variant orthography; this is what the mss of the Siddhayogevarmata read in 13.10a ([ /a]yut s. akahomam ca), as s a a. t . reported by Trzsk. 22 Niv saguhya 11.86 describes magically enlivening an unmutilated corpse, which if male becomes s a a slave whom one may ride as a mount and go anywhere; if female, the corpse becomes like a celestial maiden, with whom one may live ten-thousand years, invisible. The more elaborate rites described in 14.12729 are performed in a cremation ground, and the enlivened corpse (vet la) may be dispatched to a do a particular task or fetch magical substances or treasure: tata cottis. hati bruvate ca | bho vrapurusa kim s .t . . karomti | sa vaktavyah | psitam k mam dadasveti | tatah sarvam samp dayati | atha v janarocanamanahsil hy a . a a . . . . . a ausadhiratnanidh nam v anayasveti | tatas tat karma krtv tatraiva gatv nipatati (And then [the vet la] a . a a a . . a arises and says, O heroic man, what shall I do? He is to be told, provide the [following] desired wish. The he accomplishes everything. Or [one should say,] fetch ointment, yellow pigment, the manahsil . a mineral, or herbs, gems, or hidden treasure. Then, after doing that work, [the corpse] goes there and falls [back] down). 14.127, excerpt. 23 Guhyasutra 3.38cd43ab: ratisambhogakual m rupayauvanaalinm 38 s a. s drsm striyam as dya niruddhendriyagocarah | a . . . cumban linganam kury l lingam sth pya bhagopari 39 a a . a . a japadhy naparo bhutv asidh r vratam caret | a aa . yadi k mavaam gacchet patate narake dhruvam 40 a s . nav tmakam japel la dvaye | a . abdam sanm sam tram v ya cared vratam uttamam 41 a . a s . . . a tasya siddhih praj yeta adham madhyamottam | a a a . . vratasthah pacalaksani punar japtv tu siddhyate 42 a . . sarve mantr s ca siddhyante psitam ca phalam bhavet | a . .

40cd va am gacchet patate ] conj. (Goodall); va am cche patate A41/14 42b ottam ] em. (Goodall); s . s . a ottam h A41/14 a. 42c vratasthah ] corr. (Goodall); vratastha A41/14 .

Compare with BraY xxxix, the asidh ravratapatala (excerpts): a a . atah param pravaksy mi asidh ravratam mah n | a a . . . a . sarvasiddhipradam proktam sarvayoginipujanam 1 . . purvalaksanasamyukt m yosit m suratocchuk m | a. a . . . . a. atvarupasampann m navayauvanadarpit m 2 a. a . ... cumban linganam krtv lingam tatra viniksipet | a . . a . . nityanaimittikam k myam japam kury d avagrahe 11 a . a . .

141 from the Niv saguhya, however, is focus upon what in the BraY are the most potent s a a of potent substances: the guhy mrtas or secret nectars, i.e. sexual uids. However, a . besides metal ores, neem oil, and the like, the Niv saguhyas ritual occasionally taps s a the power of conventionally impure substances, such as blood and beef.24 In the Niv saguhya, some glimpses may be had of the cult of spirits prominent s a in later bhairavatantras. In an example from chapter eleven, one who fasts, smears the body with crematory ash, and performs twelve-lakh repetitions of the mantra om anathaya namah has the daran of spirits (bhuta), who offer him magical subs . . stances (siddhadravya) that induce invisibility.25 Such magical, transactional experiences are greatly elaborated upon in the Yogin cult, wherein encounter (mel pa) a with goddesses becomes one of the central aims of ritual. Also noteworthy is the Niv saguhyas prescription for gaining the aid of a yaksin, a subject to which the s a . . BraY devotes a chapter.26 Further reference is made to achieving power over such a
nir c rapad vastho vyom rnavanisevakah | a a a a . . pram d d yadi ksobhah sy t svayam eva ak ritah 12 a a a . . a . japed daasahasr ni tattvayuktah tu s dhakah | s a. a . . cumban linganai caiva stk raih savil sakaih 13 a s a . a . paratattv valok ca ksobho naiva sam caret | a a . ... anena kramayogena avrataghnas tu s dhakah 27 a . m sadvayam trim sam v c turm syam ath pi v | a a a a a a a . pacasanm sikam y vac cared dhy napar yanah 28 a a . . . . a . a pratim s bhavet tasya uttam divyagocarah | aa a . sanm sena praj yeta anim digunanvitah 29 a . . a . a . . ...
12c ksobhah ] em.; ksobho Bya 13b yuktah ] em.; yuktim Bya 28c sanm sikam ] corr.; sa. . . . . . . . a . tm sikam Bya 28cd y vac cared ] em.; y va care Bya 29c sanm sena ] corr.; satm sena Bya 29d a a . a . . a . . . a gunanvitah ] em.; gunanvitau Bya . . .

Sanderson draws attention to a similar rite in another early Saiddh ntika sourcethe Matangaa p ramevara. Review of N. R. Bhatt, Matangap ramevar gama (Kriy p da, Yogap da et Cary p da), avec a s a s a a a a a a le commentaire de Bhat. a R makan. ha. Edition Critique, bsoas 48, 3 (1985): 565. .t a .t 24 Note for instance that Niv saguhya 10.87 mentions homa using beef (gom msa), while 14.66 des a a. scribes smearing an efgy (pratikrti) with blood as part of a rite of subjugation (vakarana). s . . 25 Guhyasutra 11.64: anena mantrena smaanabhasman sn tv nir h ro dv daalaksam japet bhutaganani s a a a a a a s . . . . payati [em.; payanti Cod.] | siddhadravy ni prayacchanti | taih siddhadravyair antarhito bhavati 64 (Havs s a. . ing bathed in ashes using this mantra, while fasting, one should repeat the mantra twelve-hundred thousand times. He sees groups of spirits; they bestow magical substances. Through those magical substances, he becomes invisible). 26 Niv saguhya 10.8184. This procedure, called yaksinvidhi (yaksiny esa vidhih), involves worship of s a . . . . a . . an image that comes to life when the rite is complete: siddh s kim karomti bh ry me bhavasveti | tay a a . a a a saha ramate y vad acandrat rakam (when accomplished, she [says] what shall I do? Be my wife. He a a enjoys himself with her for the duration of the moon and stars). The subsequent verse (10.84) provides

142 s female spirits as the bhut, piacin, and n gin, although generally, feminine-gender a spirits are little emphasized. Erotic magic is present, such as a rite in the Niv saguhya s a wherein one magically transforms a female goat or sheep (aj ) into a woman who a fullls all of ones desires.27 Also noteworthy are the numerous references to joining the ranks of the vidy dharas, suggesting that even at this level of the tradition, a magical perfection and the attainment of embodied divinity had emerged as welldened aims of ritual. In general character, the Niv saguhya thus suggests the extent s a to which the ritual of the bhairavatantras and Yogin cult has deep roots in earlier tradition, representing a shift in emphasis rather than something altogether novel.
tantras of dakins, bhutas, and the sisters of tumburu i .

While aspects of the Niv saguhya afford insight into the cultic background of the s a bhairavatantras, the gulf between the Niv sa corpus and Vidy ptha sources remains s a a . considerable. There appear, however, to have been early traditions within Tantric Saivism possessing closer links with the cult of yogins, possibly representing phases intermediary between the Niv saguhya and the Vidy ptha, although the documens a a . tation for these is fragmentary. Most historically signicant is the cult of the Sisters of Tumburu, the scriptures of which came to be classied as tantras of the v masrotas, a the Left Stream of scriptural revelation spoken by Sad sivas northern or leftward a face, the feminine V madeva. We also nd references to an early tantric literature a devoted to exorcism, the bhutatantras, and one concerned with a cult of female spirits called dakins. . In the rst half of the seventh century, the Buddhist philosopher Dharmakrti makes critical remarks concerning dakintantras and bhagintantras, the latter of which . his commentator Karnakagomin identies as tantras of the Four Sisters (catur. bhagintantras)in all probability, Sanderson argues, scriptures of the Saiva v masrotas. a
means for making a wife of a snake goddess (n gin). Cf. Niv saguhya 14.83, which describes rites for a s a subjugating female spiritsthe yaksin, piacin, and bhut. BraY lxiv, the yaksins dhanapatala, teaches a . . s . . a . in considerable detail processes for winning over various yaksins, and well as a female ear goblin . . (karnapiacin). . s 27 rupavat str bhavati s sarvak m [n] dad ti. The rite is described in Niv saguhya 14.153. a a a a s a

143 The dakintantras Dharmakrti refers to do not appear to have survived, yet the exis . tence of Saiva texts by this designation can be conrmed through several additional references.28 Authors mentioning these texts associate them with parasitic, violent magical practices of the sort ascribed to dakins, female beings characterized in Saiva . sources largely as malevolent. While this literature is no longer extant, descriptions of the activities of such beings and similar ritual practices do survive in Vidy ptha a . sources, and it is possible that the tradition represented by the dakintantras was in . part subsumed into the Yogin cult of the bhairavatantras.29 While not clearly documented until Dharmakrtis reference in the early seventh century, magical practices centered upon dakins could be considerably older than . this; as discussed earlier, such deities are attested in the fth-century Gangdh r ina scription, in association with the Mother goddesses. This association is certainly suggestive, for in Vidy ptha scriptures, dakins and m trs gure with prominence in a . a. . typologies of the clans (kula, gotra) of goddesses. Nonetheless, it is unclear to what extent Dharmakrtis reference to dakintantras should be taken as evidence for the . existence of a cult of yogins in the seventh century, for a tradition of ritual centered upon dakins does not necessarily presuppose a Yogin cult of the variety evidenced . by Vidy ptha and Kaula sources. In any case, the dakintantras undoubtedly have a . . signicance for the history of the cult of yogins, either representing one of its early forms or comprising one of the independent strands coming together in its formation.
Concerning the statements of Dharmakrti and Karnakagomin, see Sanderson, History through . Textual Criticism, 1112. Sanderson identies several other references to dakintantras, including . Ksemar jas Netroddyota, ad Netratantra 20.39. Ibid., 12 (n. 10). a . 29 On dakins and their characterization in Saivism, see chapter 2, n. 46. Cf. Sanderson, ibid., 12 . (n. 10). BraY xcix.1112 associates dakins with violent transactional encounters (hathamel pa), and ata a . . tainment of the state of being a dakin with inverted (viloma) ritual means. See the edition in the . present dissertation. Cf. Tantrasadbh va 16.181218, which describes the pernicious activities of several a varieties of yogin, such as the adhoniv sik and its sub-types; several verses from this passage are s a a quoted by Ksemar ja ad Netratantra 19.55. One would imagine that dakintantras taught practices such a . . as pac mrt karsana, extraction of the ve [bodily] nectars, said in the M latm dhava to be the source a . a . . a a of the wicked yogin Kap lakundal s ight. This ritual is described in e.g. BraY iii.198207. Regarding a a .. a such practices in the Jayadrathay mala, see Sanderson, Purity and Power among the Brahmans of Kasha mir, 213. The rst half of Netratantra 20 contains a description of the yogic means by which yogins liberate their victims. Cf. Kulas ra 12, which describes the manner by which ve varieties of yogin a invade the body; Trzsk discusses the relevant passage in the entry dikcar in T ntrik bhidh nakoa a a a s iii (forthcoming).

144 Perhaps a contemporary of Dharmakrti, the Brahmanical author Bh ruci in his a commentary on the Manusmrti mentions bhutatantras, a class of tantric literature . apparently concerned with exorcism and magic pertaining to spirits (bhuta). Although little early literature of this variety survives, it was once consequential enough to be classied as one of the ve major divisions of Tantric Saiva scripture, alongside the siddh ntatantras, bhairavatantras, v matantras, and g rudatantras.30 Unlike dakina a a . . tantras, the bhutatantras are unlikely to have concerned yogins directly; their ritual might however lie in the background of the Vidy pthas cremation-ground cult of a . female spirits. The Netratantra in particular, a text Sanderson shows to be of Kashmiri provenance (circa 700850 c.e.),31 provides a window into the exorcistic and apotropaic dimensions of the cult of yogins, which might have had roots in bhuta tantra material. In the Netratantra, yogins and their numerous varieties gure pri marily as potentially harmful deitiesa role that may be viewed in continuity with the early cult of Mother goddesses and the grahas of Skanda.32 A question of considerable importance to the early history of the Saiva Yogin cult concerns its relationship to the v matantras, scriptures of the cult of Tumburu and a four goddesses known as Sisters (bhagin). Dharmakrti in all probability refers to the v matantras when he speaks of bhagintantras, Tantras of the Sisters, a literature a that is in any case demonstrably old. Unfortunately, there appears to survive only one complete v matantra: the brief Vnasikh tantra, published by Teun Goudriaan a . a
Sanderson remarks, [Bh ruci,] who may also belong to the rst half of the seventh century, refers a to the Bhutatantras in his commentary on Manu as sources teaching rites for the mastering of Vet las. a Sanderson also provides canonical lists of bhutatantras from two sources, and identies the Kriy k lagua a nottara as an early extant source of this type, surviving in a twelfth-century Nepalese ms and quotations . in the Netroddyota of Ksemar ja. History through Textual Criticism, 1314. On the inclusion of bhutaa . tantras as one of the streams of Saiva revelation, see e.g. Jrgen Hanneder, Abhinavaguptas Philosophy of Revelation: M linslokav rttika I, 1399, 1719. a a 31 Religion and the State: Saiva Ofciants in the Territory of the Kings Brahmanical Chaplain, 27394. 32 Cf., e.g., Netratantra 20.5075; this begins with a list of harmful entities including bhutas, m trs, a. and yogins, and outlines means for their appeasement (praamana). Among many other skills, the s practitioner capable of averting the dangers they pose should be versed in the rites of the bhutatantras (bhutatantravidhau, 61a). On the cult of Skanda, the Mothers, and grahas, see chapter 2, section 2; see also David White, Kiss of the Yogin, 3563.

145 on the basis of its single extant Nepalese manuscript.33 Concerning dating, there are other indications that v matantras existed in the seventh century, if not earlier. a Most compelling is the discovery of loose folios from a Saiva text in the tradition of the v matantras in the Gilgit manuscript horde, copied presumably prior to the a eighth century.34 It seems that Sankara, the famous Ved ntin (. c. 800 c.e.?), refers a to worship of the Four Sisters as well, alongside the Mothers, in his commentary on the Bhagavadgt .35 Furthermore, the core pantheon of the v matantras features a a in several chapters of the Buddhist Majurmulakalpa, which includes a narrative of s their conversion to the Dharma; portions of this tantra are held to date from the eighth century.36 Another indication of their antiquity lies in the fact that the BraY , early as a it may itself be, shows clear knowledge of the Saiva v masrotas and mentions several a of its scriptures by name, including the extant Vna (i.e. Vnasikha).37 As Sanderson . .
The Vnasikhatantra: a Saiva Tantra of the Left Current, ed. Teun Goudriaan. This edition is based on . the manuscript nak 1-1076 (ngmpp reel a43/3). Sanderson has, in addition, suggested the possibility s that the Siracheda, a text mentioned in early lists of v matantras, has been redacted into the rst of the a Jayadrathay malas four satkas. History through Textual Criticism, 3132 (n. 33). a . . 34 An edition of this material is under preparation by Sanderson and Somadeva Vasudeva; I thank the latter for providing this information. 35 Sankaras comments ad Bhagavadgt 9.25 have been discussed by R. Nagaswamy, The Sixty-four a Yogins and Bhuta Worship as Mentioned by Sankara in his Commentary on the Bhagavadgt , Berliner a a Indologische Studien 910 (1996): 23746. Commenting on 9.25c, bhut ni y nti bhutejy (worshippers of a a spirits attain to the spirits), Sankara remarks, according to the Anand srama edition, bhutani vin yakaa a a a . a. m trganacaturbhaginy dni yanti bhutejya bhut n m pujak h; he hence glossess bhutas, with Gane a, a. . a . s the group of Mother goddesses, the Four Sisters, etc. Nagaswamy, puzzled about the identity of the Four Sisters, points out that the variant catuhsas. iyogin (the Sixty-four Yogins) is reported for . . .t caturbhagin in the editions apparatus, and is read by the commentator Dhanapatisuri; the mss he consulted in the Sarasvati Mahal Library read caturbhagin , however. Ibid., 24244. Although further examination of the manuscript evidence is warranted, I doubt Nagaswamy is correct in opining that the reference to sixty-four yogins is original. As discussed in the previous chapter, the notion of the yogins as sixty-four does not seem particularly early, and only becomes prevalent in Kaula sources from around the tenth century. 36 See the discussion of the Majurmulakalpa in section 4 of this chapter. s 37 In a list not necessarily of v matantras alone, BraY lxxvi.9193 mentions the Bhairava, Naya (or a a Bhairavanaya?), Saukra, Mah sammoha, and Vna: a . .

satyam etan mah devi na kathyam yasya kasya cit 91 a . bhairave tu naye caiva saukre caiva tu s dhakah | a . mah sanmohavne ca guhyatantre sudurlabham 92 a . brahmay malatantre tu vidy p. he tu bh sitam | a a t a. kathitam picutantre tu prayogam ida durlabham 93
92b saukre ] corr.; saukre Bya

This is the truth, O Mah dev; it must not be told to just anyone. [This] is very difcult a to obtain in the Bhairava, Naya, Saukra, () . . . (?) and the secret Mah sammoha- and Va . na- tantras. But it has been spoken in the Brahmay malatantra in the Vidy p. ha; this rare a a t . procedure has been taught in the Picutantra.

146 shows, there are also indications that texts of the v masrotas such as the Vnasikha a . were in circulation in Southeast Asia in the ninth century, along with Saiddh ntika a sources such as the Niv sa.38 s a It is possible that the v matantras represent the earliest signicant tantric Saiva a goddess cult, the pantheon of Tumburu and the Four Sisters presaging the goddessdominated mandalas of Bhairava in the Vidy ptha. There are in fact signicant a . .. iconographic parallels between Tumburu and the Four Sisters and certain forms of Bhairava and the Four Devs in the BraY .39 In general concerns and character, the rit a
Of these, the Saukra, Mah sammoha, and Vna (=Vnasikha) are certainly tantras of the v masrotas; the a . . . a former two are listed as such in BraY xxxviii.77, along with the Nayottara and bhav . See table 4.3. a a Note in 93d ida for idam, metri causa. In 93d, guhyatantra could, rather than be an adjective, refer to a specic textpotentially the Niv saguhya. s a 38 History through Textual Criticism, 78 (n. 5). See also The Saiva Religion among the Khmers (Part I), 35557. 39 That the cult of the bhagins lies in the background of the BraY is suggested by their position in a the latters cosmology. BraY lv describes a hierarchy of clans (kula) of the goddesses in which bhagins a occupy the penultimate position, at the level of svaratattva, just below the [Four] Devs of the mandala .. of the BraY s Kap lsabhairava. See table 4.7b, and verses 1114 in the critical edition. a a In the Buddhist Majurmulakalpa, which contains rare and important material on the cult of bhagins, s the iconography of the Four Sisters and Tumburu has unusual maritime elements. The Four Sisters are repeatedly described as mounted in a boat (nauy nasam rudh h) with Tumburu as the helmsman (kara a . a. nadh ra). Cf., e.g., 47.2324. The maritime association of the bhagins nds conrmation in Netratantra a . 11 as well, which envisions the deities on a boat in the Ocean of Milk (n vam ksr rnavam corvm saktim a . . a . . . a a. a adh rik m subh m asan rtham prayujta santyartham sitanrajam, 11.25cd26ab). While impossible to a . . determine with certainty, it seems probable that this representation of the Sisters and Tumburu is the source for the iconography of certain forms of Bhairava and the goddesses in the BraY . Chapter four, a a veritable treasury of tantric iconography, describes the eight goddesses of the retinue of Kap lsathe a Four Devs and Four Dutsin a row (pankti) in a boat, mounted upon human corpses, with Bhairava the helmsman, sporting in the Ocean of Milk. Compare with Mah mardakabhairava in BraY lxxvii a a a ardhanarsvara form of the deity possessing eight arms and four faces, standing upon a corpse in a boat. A similar four-faced, eight-armed Bhairava is mentioned in lxx.2829. While Mah m rdaka is a a worshipped as a solitary deity (ekavra), chapters lxxxii and lxxxvii describe Bhairava and the eight mandala goddesses with similar iconography. Cf. BraY lxxxii.16266: a .. rupakan tu pravaksy mi devn m s dhanasya tu | a. a . a purvoktarupakam jeya kap lamundasamyutam 162 a . .. . n v rudh s tu dhy tavy n v rudh s tu pujayet | a a . a a a a a . a a evam kramavibh gena nagnarup s caturmukh h 163 a a. . mah pretasam rudh devo devya ca krttit h | a a . a s a. dutyo vai padmahast s tu trimukh h sadbhuj h smrt h a a. . . a. . a. devyas tv as. abhuj jey h s dhakena tu dhmat | a a. a a .t anena vidhin jey asusiddhiprad yik h 165 a a a a. jy jalinamask raih siddhid muktid s tath | a a . a a a svarupadhy nayogena ekaikasy h prthak prthak 166 a a. . .


I shall teach the form of the goddesses, and of [their] s dhana. The form [of the deities] a should be known as that stated previously, endowed with skulls and severed heads.

a a 163c krama ] conj.; nagna Bya 163d rup s ] em.; rup Bya mukh h ] em.; mukh Bya a. a devo ] em.; devau Bya 164d sadbhuj h ] corr.; satbhuj Bya 165a jey h ] corr.; jey Bya a. a a. a . . . . prad yik h ] em.; prad yik Bya a a. a a 166a namask raih ] corr.; namask rai Bya a . a

164b 165d

147 ual world of the Vnasikha is largely consistent with the Vidy ptha, and the colophon . a . of the Vnasikhas Nepalese manuscript in fact refers to the text as a y malatantra.40 . a However, despite elements of congruity, it is unclear whether and to what extent a cult of yogins was present in the v matantras. The short Vnasikha contains only a a . single passing reference to yogins, describing them as deities who would punish those violating the scripture, the initiatory Pledges, or the gurus.41 And although predating the BraY , the Vnasikha cannot be regarded as one of the earliest of its a . class of scriptures, for it situates itself as revelation subsequent to fundamental v maa tantras such as the Nayottara and Saukra.42 Hence, while the v matantras undoubtedly a gure prominently in the background of the cult of yogins, it seems impossible to adequately assess the nature and extent of this role.
But [they] should be meditated upon as mounted in a boat, one should worship them mounted in a boat. Thus the deity [Bhairava] and the goddesses, () in their respective order (?), are said to be naked, with four faces, mounted upon human corpses. The Duts are said to have lotuses in hand, and to have three faces and six arms. But the Devs the wise s dhaka should know to have eight arms. Through this procedure, they are known to a bestow siddhi rapidly. Through worship, supplication, and salutations they grant siddhi, and through the yoga of meditation on their forms, of each one individually, they give liberation. In this schema, the Four Devs appear to supplant the Four Sisters, attended upon by four more goddessesthe Duts. 40 According to Goudriaans edition, the colophon reads, vnasikh s rdhaatatrayam y malatantram . a a s . a . sam ptam iti (thus ends the Vnasikha, a y malatantra of three-hundred and fty verses). The middle a . a section, in particular, concerns rituals for pacication (antika), nourishing (paus. ika), magical subjus .t gation (vakarana), attraction ( karsana), driving away enemies (ucc . ana), causing enmity (vidvesana), s a . . at . . . and slaying (m rana), similar in character to those of the BraY . Some, for instance, involve cremationa . a ground re sacrice using human esh (mah m msa); cf. 162, 18990ab. Sanderson reports that the a a. Jayadrathay mala classies v matantras as belonging to the Vidy ptha. History through Textual Critia a a . cism, 31 (n. 33). 41 Vnasikha 329cd21ab: . svayamgrhtamantr s ca n stik vedanindak h 329 a a a a. . . . a. samayebhyah paribhras. as tath tantravidusak h | a . . t .. gurunam vihethanapar s tantras ravilopak h 320 a a a. . yoginbhih sad bhras. ah kathyante dharmalopak h | . a t . a. . Those who take up mantras on their own, atheists, critics of the vedas, breakers of the Pledges, desecrators of the tantras, those intent on harming the gurus, and those who violate the essence of the tantrasthose who violate Dharma are said ever to be ruined by the yogins.

Vnasikha 410. The Nayottara and Saukra are both mentioned in the BraY ; see above (n. 37). . a

the mantraptha and svacchandalalitabhairavatantra i.

In a model of the Saiva canon expounded in BraY xxxviii and a number of other a sources, scriptures of the cult of Bhairava and associated goddessesthose designated bhairavatantrasare classied according to four p. has or mounds: those t of mudr s, mandalas, mantras, and vidy s.43 However, this classication appears to a a .. mask what Sanderson identies as a more fundamental twofold division between the Mantraptha and Vidy ptha, the pantheons of which consist predominantly . a . of male mantra-deities and female vidy -mantras, respectively.44 Literature of a the Vidy ptha, The Division/Seat of Female Mantras, is thus intrinsically cona . cerned with goddesses, and the Vidy ptha/Mantraptha divide itself appears ina . . tended, primarily, for distinguishing bhairavatantras connected with the cult of yogins from those which are nota distinction bearing comparison with that between yoga/mah yoga- and yogintantras in the canon of tantric Buddhist scripture (discussed a subsequently). As appears true of the latter division, this might reect an historical development, with the Vidy ptha yogin traditions developing within a Mantraptha a . . cultic context. In some respects, the Vidy ptha/Mantraptha division appears contrived: note a . . that the BraY places the Svacchandatantra at the head of the Vidy ptha as the rst a a . of eight tantras named after particular forms of Bhairava. Only four scriptures are assigned to the Mantraptha, none of which appear extant.45 However, the Svaccha. nda or Svacchandalalitabhairavatantra is in some sources, including its own colophons in the Nepalese manuscripts, held to belong to the Mantrapthaa scriptural cate. gory otherwise poorly represented in the surviving literature, having this text alone as its major early exemplar.46 The paucity of surviving texts might suggest that
43 The BraY s vision of the Saiva canon is discussed in chapter 5 of this thesis. The notion of the a bhairavatantras being divided according to four p. has is not uncommon; note for instance, in the Svact chandatantra, the Goddesss initial question and 13.6cd. Jayadrathay mala i, chapter thirty-six, comprises a an exposition on this subject. 44 Sanderson, Saivism and the Tantric Traditions, 66871; and History through Textual Criticism, 1920. 45 See table 4.4 in the next chapter. 46 On the Svacchandatantra and the Mantraptha, see Sanderson, History through Textual Criticism, .

149 the yogin cult of the Vidy ptha corpus largely superceded Mantraptha forms of a . . the Bhairava cult, much as the v matantras appear to have lost relevance and gone a out of circulation. The k p lika Vidy ptha itself appears to have been eclipsed by a a a . Kaula cults, eventually, the scriptures of which comprise the bulk of surviving non Saiddh ntika Saiva literature. a The Svacchandatantra survives in two recensions, one in comparatively polished Sanskrit transmitted in Kashmir, commented upon by Ksemar ja in the eleventh cena . tury and published in the Kashmir Series of Texts and Studies; and another preserved in Nepalese manuscripts.47 Though these cover most of the same content, the rustic language of the latter in all likelihood reects the text in an earlier form.48 Sanderson highlights a number of respects in which this scripture and the cult of Svacchandabhairava have distinctive historical signicance, representing an early and widely inuential tradition in Tantric Saivism.49 Although it has not been rmly dated, the Svacchandatantra heads several lists of the non-Saiddh ntika tantras, including the a BraY s account of the bhairavatantras and the sixty-four tantras of the s d sivacakra a a a .t in the Srkan. hyasamhit .50 It bears a close relationship to the ancient Niv satattvas a . a samhit , from which it redacts substantial material, and predates the Tantrasadbh va a . a
1921. While Svacchandatantra 14 refers to itself as the mudr p. ha (14.26ab), the Svacchandatantra does a t not place itself as a whole in a single p. ha; the Ur-Svacchanda, like the BraY , is said to contain all t a four within itself (Svacchandatantra 1.5cd; see the discussion in chapter 5 of this thesis). As mentioned, chapter colophons of the Nepalese manuscripts nonetheless assign the Svacchandatantra to the mantrap. ha. t 47 I am aware of four Nepalese palm-leaf manuscripts of the Svacchandatantra: nak 1-224 (ngmpp reel b28/18), nak 9-68 (ngmpp reel c6/5), R stry bhilekha ms 5-691 (ngmpp reel a988/4), and the a. . a incomplete Bodleian Library codex (Svacchandalalitabhairava Mah tantra, ms Sansk. d. 38 [R]). Eight a paper manuscripts that appear complete or nearly so have been lmed by the ngmpp as well: reels a201/4 (nak 5-4974; lmed a second time as a1176/16), e137/3, e2188/11, a204/3&5 (nak 1-43), a203/4 (nak 5/6165), and a201/9a202/1 (nak 1-224), all in Newari writing; and a201/7 (or a210/7? nak 54974) and a203/9 (nak 1-11), both in the Devan gar script. a 48 The case of the Svacchandatantra bears comparison with that of the Netratantra, studied in detail by Sanderson, Religion and the State, passim. The Netratantra too exists in a comparatively polished Kashmirian recension and more rustic version preserved in Nepalese mss; Sanderson proposes, compellingly, that the latter is comparatively archaic. Ibid., 243. 49 Sanderson, History through Textual Criticism, 2021. 50 Srkan. hyasamhit 273b; numbering as given in Jrgen Hanneder, Abhinavaguptas Philosophy of .t . a .t Revelation. M linslokav rttika i, 1399, 263. This section of the Srkan. hya is quoted by Jayaratha, coma a menting on Tantr loka 1.17. a

150 of the (Trika) Vidy ptha.51 No evidence I am yet aware of rmly establishes its a . chronology relative to the BraY an issue discussed in the subsequent chapter. a In cultic orientation, the mildly k p lika Svacchandatantra shares much with the a a siddh ntatantras, and in fact dees or potentially predates distinct divisions between a the Saiddh ntika and non-Saiddh ntika traditions; indeed, the scripture had suba a stantial authority in Saiddh ntika circles, evidently.52 In the Svacchandatantra, the a cultic status of the spouse-goddess of Bhairava, Aghore or Bhairav, is secondary,53 s while goddesses in general have little prominence in its pantheons. The Mother goddesses who feature in the background of the early Yogin cult have only a marginal presence in the Svacchandatantra.54 However, Svacchandatantra 10, in describing the cosmological sphere (bhuvana) called Suc ru, describes Siva (Um pati) in a mandala a a .. of the Brahmanical Mothers. This appears to be an elaboration upon a brief reference in Niv saguhya 5, upon which this section of the Svacchandatantra is based, to s a a divine city called m trnand , dear to the Mothers. One of thirty-four cities (pura) a. a on mount Meru, the Niv sa describes this as the abode of Um pati, where sport s a a inebriated Mother goddesses of unspecied number.55 In the Svacchandatantra verSanderson, History through Textual Criticism, 2232. The evidence for this includes the existence of a Saiddh ntika tradition of Svacchandatantra exa egesis, referred to by Ksemar ja. Sanderson, Purity and Power among the Brahmans of Kashmir, a . 204. Furthermore, Sanderson shows that the inuential Saiddh ntika ritual manual of Bhojadeva, the a Siddh ntas rapaddhati, drew upon the Svacchandatantra extensively and deeply. History through Texa a tual Criticism, 2122 (n. 26; quote on p. 22); and The Saiva Religion among the Khmers, 35960, 403 (n. 197). 53 Sanderson, Saivism and the Tantric Traditions, 670. 54 Although referred to as prominent attendants of Siva in the texts opening (stuyam nam maheanam a . s . ganam trnisevitam, 1.2cd), the Mothers are otherwise mentioned primarily in passing, in lists of deities. a. . . Cf., e.g., Svacchandatantra 10.214cd15ab:
52 51

devagandharvasiddh s ca rsayo tha vin yak h a . . a a. ganam trbhaginya ca vet l r ksas dayah | a. s aa a . a . .


Niv saguhya 5.67cd68: s a [e]k daatm lok tm vrarudra um patih 67 a s a a a a . matrnand pur ramy sarvaratnavicitrit | a a a . . a. krdante m taras tatra madhup navighurnit h 68 . a a

Cf. Svacchandatantra 10.140cd41ab: pacime dharmar jasya m trnand pur smrt s a a. a . a . a. krdanti m taras tatra madhup navighurnit h . a a 140

151 sion, these are recast as a mandala of the Brahmanical Seven Mothers, described in .. full iconographic detail and apparently joined by an eighth goddess, Mah laksm.56 a . The Kashmiri recension of the Svacchandatantra provides additional evidence for the incorporation of Mother goddesses: in a description of the M trk , the alphabetical a. a Matrix, the eight vargas of the Sanskrit alphabet are correlated with the Eight Mothers, a connection absent from the text as transmitted in Nepalese manuscripts;57 no link of this sort is made concerning the vargas of the M trk in the Niv sa either,58 a. a s a although this becomes commonplace in latter times. The cult of yogins goddesses registers a clear presence only in the Svacchanda tantras fteenth and nal chapter, although these goddesses are mentioned elsewhere in the text.59 This chapter concerns subject matters characteristic of the Yogin cult: chomm (chumm /chummak in the Kashmiri recension), the verbal and nona a a linguistic secret codes used for communication between initiates and with yogins;
Svacchandatantra 10.1017cd30 describes the Seven Mothers in some detail, closing with a remark that suggests the presence of Mah laksm as their eighth and highest member: a . evam sa bhagav n devo m trbhih pariv ritah | a a. . a . . aste paramay laksmy tatrastho dyotaya jagat a . a 1030

57 This has been pointed out to me by Sanderson (personal communcation, January, 2007). See Svacchandatantra 1.31cd37ab in the ksts edition. Mah laksms prominence is illustrated in the fact that she a . presides over the a-varga or the vowels, while C munda merely presides over the sibilants. a . . 58 The rst two chapters of the Niv sottara in particular concern the M trk and its y ga. See also s a a. a a Niv sanaya 1 and Niv saguhya 12. s a s a 59 In particular, note Svacchandatantra 10.116cd19ab, describing the temple of Siva as H take vara: a. s yadurdhve caiva sauvarnam p t lam parikrtitam | . . aa . tatra vasaty asau devo h . akah paramevarah 116 at . s . purakotisahasrais tu samant t pariv ritah | a a . . siddhai rudraganair divyair bhaginm trbhir vrtah 117 a. . . . yoginyogakany bh rudrai caiva sakanyakaih | a s . siddhadravyasamair mantrai cint maniras yanaih 118 s a . a . siddhavidy samrddham vai h . akeasya mandiram | a at s . . Here yogins are mentioned alongside siddhas, rudras, bhagins, m trs, yogakany s, and perhaps a. a rudrakany s, in the entourage of Siva-H take vara. The commentator Ksemar ja interprets the yoga a a. s a . maidens (yogakany s) as a high grade of yogin (yoginyo yogena siddh h, yogakany s tu j tam tr eva a a. a a a a samsm ritayog h, yogins are perfected through [practice of] yoga; but yogakany s are caused to recall a. a . a their yoga upon merely being born). This passage in the Svacchandatantra is an elaboration upon Niv saguhya 5.16cd17ab, where the deities mentioned are rudras, vidy s, and vidyevaras: s a a s

sauvarnam saptamam jeyam p t lam n gasevitam | . . . . aa . a yatra citravat n ma pur rudrasam kul 16 a a a tatr sau h . hako devo vidy vidyevarair vrtah | a at a s . .

152 and yoginmel pa, transactional encounters with the goddesses. The section on mel pa a a describes a visionary encounter in which the yogin, it would appear, indicates by gesture the reality level (tattva) corresponding to the supernatural attainment the s dhaka shall by her blessing obtain.60 Otherwise, the yogin would bestow a food ofa fering (caru), the mere consumption of which transforms the s dhaka into Bhairava.61 a Chapter fteen of the Svacchandatantra is present in both the published Kashmiri recension and in all the old (palm-leaf) manuscripts of the Nepalese-transmitted recension. Nonetheless, one might suspect, as William Arraj suggests, that it belongs to a late stratum of the text.62 The Goddesss initial questions in chapter 1 do not intimate its subject matter, which appears out of character with the Svacchandatantra as a whole. At least one of the verbal code words does occur elsewhere in the text (giri, mountain, for s dhaka63 ); yet by and large, these imply a cultic context disa tinct from earlier chaptersone that includes ritual coitus, sacrice, and engagement with a level of impurity otherwise uncharacteristic.64 By all appearances, this chapter would seem to have been appended somewhat awkwardly. If so, the growth of the
Svacchandatantra 15.2432ab provides a concordance between a series of points along the body that the yogin might indicate, from the crest-tuft (ikh ) to the feet, and a series of tattvas or reality levels. s a The Nepalese-transmitted recension contains several verse-halves absent from the Kashmiri recension, which would occur after 24ab, 27cd, and 31cd of the ksts edition. These add the tattvas niyati and prthv to the concordance; the absence of the latter in particular would otherwise be inexplicable. The . chomm section ends with the statement, darayanti mah dhv nam n n bhogasamanvitam: [the yogins] a s a a . a a indicate the Great Course [i.e. the hierarchy of tattvas that comprise the universe], together with its various supernatural attainments (34ab). 61 Svacchandatantra 15.36: satat bhy sayogena dadate carukam svakam | a a . yasya sampr san d devi vreasadrso bhavet 37 a a s . Due to [his] engaging in constant practice, she bestows her own caru, by the mere consumption of which, O Goddess, he would become equal to [Bhairava,] Lord of the Heroes. Cf. Kaulaj nanirnaya 23, quoted in section 4 of the previous chapter. a . 62 Arraj, The Svacchandatantram: History and Structure of a Saiva Scripture, 36669. 63 The verbal code is given in 15.2c (s dhakas tu girir jeyah). Giri occurs in the sense of s dhaka in a a . 5.46c, in the context of enumeration of the Pledges. Arraj also suggests that the compound mrtasutra . (sacred thread from a corpse?) in 13.21b is used in the code sense of ligament (sn yu). This is a uncertain, however; the verbal code-word given for sn yu is sutra alone (15.5d). Since the thread of a a corpse would itself satisfy the contextk p lika re sacriceit seems unnecessary to posit a different a a referent. 64 An exception is Svacchandatantra 13 (referred to in the previous note), on the subject of re sacrice. The esh of a man slain with weapons and mixed with the three honeys is among the various impure offering substances listed (ranaastragh tapatitam narapiitam trimadhusamyutam juhuy t, 13.24cd). a s . a . s . . .

153 Svacchandatantra might support the hypothesis that the yogin cult of the Vidy ptha a . evolved within a Mantraptha contexta cult of Bhairava and male mantra-deities, . primarily, its k p lika dimension and ritual engagement with impurity presaging maa a jor concerns of the Vidy ptha. a .


i. scriptures of the saiva yogin cult: the vidyaptha and kaula i

Pre eleventh-century Saiva scriptures in which the cult of yogins is prominent appear to be of two basic categories: those of the Vidy ptha (Seat of Female Mantraa . deities) and Kaula ([Tradition] of the [Goddess] Clans). The distinctions between these are at once signicant and problematicproblematic because the Kaula tradition appears, most probably, to have developed within and had substantial continuity with the Vidy ptha, complicating a neat division between the two. Thorough a . investigation of this important issue is beyond the present study. Most relevant is the fact that the earliest attested literature of the Saiva Yogin cult, including the BraY , a belongs to the Vidy ptha, while in contrast, the greater portion of the extant Saiva a . literature concerned with yogins belongs to various Kaula traditions. Four Vidy ptha works of the k p lika yogin cult appear extant: the BraY , Siddhaa . a a a yogevarmata, Tantrasadbh va, and Jayadrathay mala.65 Among these, the BraY and s a a a Tantrasadbh va alone survive in comparatively early and complete forms. The Siddhaa yogevarmata is preserved only in a short, probably secondary redaction transmits ted in Nepalese manuscripts, missing some passages attributed to it in the exegetical literature,66 while in the form we have it, the Jayadrathay mala might not prea date Abhinavagupta.67 However, the third book of the Jayadrathay malathe Yoa ginsac raprakaranaappears to have been an early, independent work of the Vidy a a . ptha, for along with the BraY , Siddhayogevarmata, and Tantrasadbh va, it is one of . a s a the extant Saiva texts that Sanderson identies as sources for the Buddhist LaghuSee Sanderson, History through Textual Criticism, 3132 (n. 33). See Trzsk, Doctrine of Magic Spirits, ivv. 67 See Sanderson, Remarks on the Text of the Kubjik mata, 2. Sanderson points out that the Jayada rathay mala is cited by Ksemar ja, but not apparently by his preceptor Abhinavagupta. a a .
66 65

154 samvaratantra.68 Of these four, only the Siddhayogevarmata has been critically edited, . s in part, while the present dissertation contributes an edition of selections from the BraY . a Although focused upon distinct pantheons, these early Vidy ptha scriptures a . share much in ritual orientation, exemplifying the yogin cult in its most radically antinomian and possibly archaic form. The character of the ritual world of the Vidy a ptha has been memorably captured in the words of Sanderson: . Smeared with the ashes of funeral pyres, wearing ornaments of human bone, the initiate would carry in one hand a cranial begging-bowl and in the other a khatv nga, a trident-topped staff on which was xed beneath . a the prongs a human skull adorned with a banner of blood-stained cloth. Having thus taken on the appearance of the ferocious deities of his cult, he roamed about seeking to call forth these gods and their retinues in apocalyptic visions and thereby to assimilate their superhuman identities and powers. These invocations took place precisely where the uninitiated were in greatest danger of possession: on mountains, in caves, by rivers, in forests, at the feet of isolated trees, in deserted houses, at crossroads, in the jungle temples of the Mother-Goddesses, but above all in the cremation-grounds, the favorite haunts of Bhairava and K l and the focus a of their macabre and erotic cult. The initiate moved from the domain of male autonomy and responsibility idealised by the Mm msakas into a vi a. sionary world of permeable consciousness dominated by the female and the theriomorphic. Often transvestite in his rites he mapped out a world of ecstatic delirium in which the boundaries between actual women and the hordes of their celestial and protean counterparts, between the outer and the inner, was barely perceptible. Intoxicated with wine, itself the embodiment of these powers, he sought through the incantation of mantras and the offering of mingled menstrual blood and semen, the quintessential impurities, to induce these hordes to reveal themselves. Taming them with an offering of his own blood, he received from them the powers he desired. At the same time he was alert to perceive their incarnation in human women and was provided by the tradition with the criteria by which he might recognize their clan-afnities. For a divinatory rite at the time of his initiation had determined his occult link with one of these clans, in order that by the grace of his clan-sisters, who embodied the clan-goddesses and were his spiritual superiors, he might attain by the most direct route liberating possession by the ferocious cosmic deity who was the controller and emanator of all these forces.69
The matter of the Laghuamvaras sources, as identied by Sanderson in History through Textual s . Criticism (pp. 4147), is taken up in section 5. 69 Sanderson, Purity and Power, 2012. The annotation to this passage, rich with references to

155 As Sanderson portrays, yogins are central to the ritual world of the Vidy p a thaas goddesses met with in visionary ritual encounters, as the luminous matrix . (j la) of Sivas feminine powers or saktis, and as embodied in female practitioners. a Recognition of yogins and transactional encounters with them are among the most characteristic subjects treated in Vidy ptha scriptures. Indeed, the entire edice of a . tantric ritual appears oriented within the Vidy ptha toward the aim of yoginmel pa, a . a power-bestowing union or encounter with the goddesses. In the BraY , the cena trality of encounters with yogins nds expression in the abundance of references to such as the outcome of ritual. These vary from cursory statements, such as the goddesses manifest directly, he becomes dear to the yogins, or he attains melaka, to vivid accounts of transactions with the deities.70 The very material ingredients of ritual are selected for their ability to bring about the goddesses proximity.71 Several passages explicitly describe yoginmel pa as the ultimate fruition of tantric ritual in a the broadest sense: BraY lxxiii species ritual discipline (cary ), yoga, and rites a a (kriy )besides Sivas volitionas the causes of melaka, while longer lists of modes a of tantric ritual are enumerated in passages in BraY xcix and Svacchandatantra 15.72 a Through his communion with yogins, the s dhaka attains the powers of Bhairava a himself.73
Vidy ptha literature, is omitted here. a . 70 Note expressions such as devyah pratyaksat m yayau (the goddesses would become directly percep. . a. tible, BraY iv.358b; here the perfect tense, third-person singular, appears optative plural in meaning), a yoginvallabho bhavet (he would become beloved of the yogins, BraY lxxvi.120d), yoginmelako bhavet a (there would transpire melaka with the yogins, BraY lviii.111f), and sanm s r dhanenaiva siddh a a . . aaa d syanti melakam (BraY lxxiii.70ab; see the critical edition). Detailed accounts of the encounters ena a . visioned with goddesses are numerous, and will be discussed in a future publication on yoginmel pa. a 71 In particular, impure incenses (dhupa) and mixtures offered in re sacrice are described as dev s naidhyak raka or -k rana, cause of the goddesses proximity (s naidhya having the sense of the clasa a a . a sical Sanskrit s mnidhya, proximity). Cf., e.g., lxxvi.118 (kajikam c tmaukra ca harit lamanahcchil a. s a a . a . . a [i.e. manahsil ] | guggulum ghrtasamyuktam dhupam s naidhyak rakam). In lxxxvii.126cd, 159ab, and 232a, a a . . . . . particular mudr s are similarly described as s naidhyak rik . a a a a 72 See BraY xcix.24 in the critical edition in part ii. In Svacchandatantra 15, the description of yogina melaka concludes with remarks suggesting it is the fruition of total accomplishment in tantric ritual a from worship (puj ) and re sacrice (homa) to mantra incantation (japa) and meditation (dhy na). See a verses 32cd38. 73 Cf., e.g., BraY lviii.108: a sarv dhvani mah devi vatsaraikanisevan t | a a a . a . pr ptamel pako bhutv krdate bhairavo yath 108 a a a After one year of observances, O Mah dev, he, being one who has obtained mel pa [with a a

156 Subsequent chapters offer a more detailed view of the Vidy ptha through the a . lens of the BraY a demonstrably early and well-preserved scripture of this catea gory. In the critical edition of part ii are presented editions and translations of select chapters of the text, including material focused upon charateristic aspects of the cult of yogins. The distinction between Kaula and Vidy ptha levels of the tantric Saiva tradition a . was posited by Sanderson, whose exposition of two decades past remains the only signicant contribution on this issue.74 Seeing the roots of Kaulism in the Vidy a ptha or Kula cult of yogins,75 he identies multiple levels of distinction. In the .
the goddesses], sports through the entire universe like Bhairava. Sanderson, Saivism and the Tantric Traditions, 67990. For the term Kaulism, see ibid., 679. What Sanderson describes there as the Kaula-Vidy ptha a . dichotomy, he spoke of in an earlier article as one between the Kaula and Kula, rather than Vidy pa tha: .
74 75

The distinction between Kula and Kaula traditions . . . is best taken to refer to the clan-structured tradition of the cremation-grounds seen in the Brahmay mala-Picumata, a Jayadrathay mala, Tantrasadbh va, Siddhayogevarmata, etc. (with its K p lika kaulika a a s a a vidhayah) on the one hand and on the other its reformation and domestication through . a the banning of mortuary and all sect-identifying signs (vyaktalingat ), generally associated with Macchanda/Matsyendra. Sanderson, Purity and Power among the Brahmans of Kashmir, 214 (n. 110). The distinction between Kula and Kaula is found in primary sources; the Kaulaj nanirnaya, for instance, contrasts Kula a . and Kaula scriptures, though in some cases using the former in a broad sense that includes the latter. Cf., e.g., Kaulaj nanirnaya 9.9ab, ete purvamah siddh h kulakaul vat rak h (these are the great Perfected a a a. a a a. . ones of yore, revealers of the Kula and Kaula [scriptures]); for kul gama in its broader sense, note, e.g., a Kaulaj nanirnaya 17.5ab, kliyanti manuj []tyantam aj tv tu kul gamam (Having failed to learn the a s a a a a . Kula scripture, human beings suffer grievously). Signicantly, the nal chapter of the BraY provides a solitary reference to Kula scriptures (kulaa j n ni), which comprise or at least include Vidy ptha texts: a a a . sanm s bhyantar d devi kulas m nyat m vrajet 16 a a a a. . . aa aesayoginn thah kulavij nasampadah | s . a . a . bhavate s dhakendras tu bhairavea iv parah 17 a s a . kulaj n ni y vanti aesadhvasthit ni tu | a a a s . a vetti sarv ni devei dad ti ca tadarthin m 18 a. s a a
16c san ] corr.; sat Bya . . . . 17a n thah ] corr.; n tha Bya a a . 17c bhavate ] em.; bhavatet Bya

After an interval of six months, O goddess, one would attain equality with the Clans [of goddesses]. Master of all yogins, endowed with the scriptural wisdom of the Clans (kulavij na), he becomes an Indra among s dhakas, like another Lord Bhairava. He knows a a all the Clan scriptures (kulaj na)as many as exist in the entire universeand he gives a [these] to their seekers, O Queen of the gods. In the subsequent chapter I show that this section of the BraY is likely to belong to a late stratum of the a text. Nowhere else does this vast work mention or describe itself as a Kula scripturea designation which in fact seems alien to early Vidy ptha sources themselves. a .

157 domain of ritual, the Kaula tradition attenuated the mortuary or k p lika dimension of a a the Vidy ptha, shifting the primary locus of ritual from the cremation ground to the a . body and consiousness itself. This shift involved internalization and simplication of ritual processes, increasingly interiorized conceptions of divine agencies, disavowal of the outer trappings of the k p lika ascetic, emphasis on ecstatic experience in erotic a a ritual, and development of comparatively sophisticated systems of yoga. The Kaula tradition hence transforms the Vidy ptha emphasis on the potency of impure ritual a . substances, the cult of spirits, and visionary, transactional encounters with deities, although these dimensions persist to some degree. In addition, on the social level the Kaula opened up new possibilities for the engagement of practitioners maintaining conventional social identities and kinship relations. The gure and cult of the yogin retain considerable signicance in Kaula scrip tural sourcesthe corpus of texts which designate themselves, frequently, as Kaula agamas, sastras, or j nas. A notable shift lies in increasing emphasis upon internal a ized, yogic conceptions of the presence and manifestation of the saktis.76 Nonethe less, decidedly exoteric conceptions of yogins persist; indeed, Kaula conceptions of yogins appear to inform the temple cult of the sixty-four goddesses and pur nic a. accounts of yogins from the early second millenium, as discussed in chapter two. By way of illustrating Kaula conceptions of yogins, I shall focus presently upon their roles in the Kaulaj nanirnaya (The Denitive Judgment on Kaula Scriptural a . Wisdom). The choice is admitedly arbitrary, for no single text is typical or representative of this large corpus. Preserved in two Nepalese manuscripts, the oldest of which belongs to the mid eleventh-century,77 the Kaulaj nanirnaya describes ita .
76 Note, for instance, that the M linvijayottaratantra of the Kaula Trika describes yoginmelaka as either a an external encounter with goddesses, who assemble to bestow power upon the s dhaka, or as the a manifestation of yogins within the yogic body. See M linvijayottaratantra 19, especially 18cd27ab. a 77 The Kaulaj nanirnaya was published in 1934 by Prabodh Candra Bagchi, in Kaulaj nanirnaya and a a . . Some Minor Texts of the School of Matsyendran tha. His edition is based upon a single eleventh-century a manuscript from the collection of the erstwhile Darbar Library in Kathmandu: Mah kaulaj nanira a naya, nak 3-362 (ngmpp reel a48/13). This manuscript has been described by Harapras d S str in the a a . Catalogue of the Palf-leaf and Selected Paper Mss. Belonging to the Darbar LIbrary, Nepal, vol. II, 32; and by Bagchi, Kaulaj nanirnaya and Some Minor Texts, 1. Although undated, its writing bears to my eye a a . strong afnity to a Nepalese Svacchandatantra manuscript dated to 1068/9 c.e.: nak 1-224 (ngmpp reel

158 self as the Yoginkaula, Kaula scripture of the Yogins.78 Revealed by Matsyendra at the mythic Moon Island (candradvpa), the text also associates itself with the sa a a a a cred site (p. ha) of K marupa or K m khy , in modern Assam; indeed, the powers t of the yogins of K m khy are attributed to the Kaulaj nanirnaya.79 As with many a a a a . Kaula scriptures, its constellation of divinities features perfected ones (siddha) and lineages of past gurusgures of little cultic status in early Vidy ptha texts.80 Its a . primary pantheon of mantra-deities is however the krama (sequence) or cakra (circle) of sixty-four yogins explicated in chapter eight. Although the Kaulaj nanirnaya lacks an exposition on the typical Vidy ptha a a . . topic of chomm , secret signs used for communication with yogins, it contains a abundant yogin material, including a vivid exposition on their movement/manifes
b28/18). Bagchis paleographical assessment would also place the manuscript towards the mid-eleventh century. Ibid., 5. Moreover, a second, more recent palm-leaf manuscript of the Kaulaj nanirnaya has a . come to my attention that was unavailable to Bagchi: Mah kaulaj nanirnayas ra, ngmpp reel h333/14 a a a . (undated, private collection). This has particular value, for it contains the text of the Kaulaj nanira nayas rst chapter, which is missing from the older codex. I expect to publish a notice concerning this . manuscript in the near future. 78 See the introduction of this thesis, section 1. 79 The Kaulaj nanirnayas chapter colophons connect the scripture with Matsyendra and Candraa . dvpa, an association explicated in the revelation narrative of chapter sixteen. The connection with K marupa/K m khy is made in verses 22.10c and 12c, while 22.12 mentions both Candradvpa as a a a a well K m khy : a a a k marupe imam sastram yoginn m grhe grhe 10 a a. . . . . nigrah nugraha caiva siddhimel pakam tath | a a a . kurvanti satatam devi asya j napras datah 11 a a . . candradvpe mah sastram avatrnam sulocane | a . . . k m khye gyate n the mah matsyodarasthitam 12 a a a a
12a dvpe ] conj.; dvpam Kjncod ; dvpam Kjned . sthitam ] Kjncod ; sthitih Kjned . 12b avatrnam ] Kjned ; a[+va]tnnam Kjncod . . . . . 12d

In K marupa, this scripture is present in the home of every yogin. By its grace do a they ever [have power to] punish and favor, and [grant] siddhi [bestowing?] encounters (mel pa). The Great Scripture was brought down at Moon Island, O woman of fair eyes. a It is [then?] proclaimed in K m khy , O Mistress[the scripture that was previously] a a a located in the belly of the great sh.
80 Yogins, siddhas, and gurus form a trinity of sacred gures in the Kaulaj nanirnaya, being men a . tioned together numerous times. Cf., e.g., 18.4cd (p rve tu pujayet siddh [n] yoginm gurum eva ca; as a . singular for plural?). In some ways the siddhas, perfected ones, appear to function as male counterparts of the yogins. Note for instance that the secondary sacred mounds (upap. ha) are said to t be places of the goddesses and siddhas (devn m siddha- layam, 8.20d); see also 11.32, quoted below. a. a Chapter nine expounds the series (pankti) of siddhas, gurus, and yogins; this include an enumeration of several past great siddhas, who are described as revealers of the Kula and Kaula [scriptures] (kulakaul vat rak h, 9b). This role of revealing scriptural teachings is one shared with yogins; see, for a a a. instance, my discussion of the term samprad ya in the annotation on BraY lxxiii.74. a a

159 tation (yoginsac ra) on the earth in various guises.81 In this text, characteristic a Vidy ptha classications of yogins based on clans of the Mother goddesses lose siga . nicance; instead, chapter eight presents an expanded taxonomy of manifestations of divine women encompassing both ritual consorts (akti) and goddesses. Consorts s are threefoldsahaj , kulaj , and antyaj manifesting in both external and internal a a a forms (bahisth and dehasth ). Externally, the innate (sahaj ) consort is ones wife, a a a while the internal innate consort traverses the body, producing intoxication and bliss. The external clan-born (kulaj ) consort is a courtesan, her internal couna terpart being the Sanskrit alphabet. Finally, the outcaste woman (antyaj ) exists a internally as the great shining sakti called Vyomam lin.82 Absent from the BraY a a and Siddhayogevarmata, such internalized conceptions of ritual consorts do have s precedent in the Tantrasadbh va, a subsequent Vidy ptha scripture.83 The Kaulaa a .
81 82

Passages from Kaulaj nanirnaya 23 have been quoted and discussed in chapter 2, section 4. a . Kaulaj nanirnaya 8.612ab: a . saktiyukto mah tm nah sahaj kulaj pi v | a a . a a a antyaj v mah devi prthagbhedam vad my aham 6 a a a a . . viv ham tu krtam yasya sahaj s tu ucyate | a . a a . . kulaj veyam ity ahur antyaj varna antyaj 7 a s a . a bahisth kathit devi adhy tmy m srnu s mpratam | a a a a . . . a gam gamaprayogena madan nandalaksanam 8 a a . . . kurute dehamadhye tu s saktih sahaj priye | a a . kulaj kim na vij t varnar sikul tmik 9 a . aa a a . a dehasth trividh prokt bahisth trividh priye | a a a a a antyaj m sampravaksy mi srnu devi yath sthitam 10 a. a . . a . . a suddhasphatikasank sa mukt m l khagevar | a aa s . urdhvatryaksam suddh mah saktih sutejas 11 a a a a . esa saktir mah tm n antyaj vyomam lin | a a a a a .
6c mah devi ] Kjned ; mah dev Kjncod a a 7b s ] em.; sa Kjncod a 7d antyaj ] Kjned ; antaj a a cod ] Kjncod ; gamy gamya Kjned kul tmik ] Kjncod ; kul tmik m Kjn 8c gam gama a a 9d a a a a Kjned 10c antyaj m ] em.; antyaj Kjncod Kjned sampravaksy mi ] Kjncodpc Kjned ; sampravaksyami a. a . . . a . Kjncodac 10d yath sthitam ] Kjned ; yath sthitim ] Kjncodpc ; yath viyath sthitim Kjncodac 11c sam a a a a a . . suddh ] Kjncod ; samsuddh (unmetrical) Kjned 11d mah saktih ] corr.; mah sakti Kjncod Kjned 12a a a a a . . mah tm n ] conj.; mah tm na Kjncod Kjned a a a a a

Text provisionally edited from the oldest codex (nak 3/362; reported as Kjncod ), taking into account the editio princeps of P. C. Bagchi (Kjned ). Here and in the other passages quoted below from Kaulaj nanira naya 8, translation and discussion of the interpretation are deferred; a new edition and translation of . the text is under preparation by the present author. 83 Tantrasadbh va 15.125cd148 posits internal or spiritual ( dhy tmik ) homologues for a taxonomy a a a a of consorts, encompassing a variety of cosmic and internal manifestations of the sakti. The context of the chapter is observance of the vidy -mantra (vidy vrata). In this schema, consorts (dut) are ninefold, a a on the basis of kinship or caste: m t duhit bhagin sahaj ca tath ntyaj 127 aa a a a a rajak carmak r ca m tang c graj tmik | a a a a a

160 j nanirnayas schema features several varieties of yogin: ksetraj (born in sacred a . a . elds) and p. haj (born in sacred mounds), whose powers are rooted in the sat a cred places they arise from;84 yogaj (born from yoga) and mantraj (born from a a mantra-[propitiation]), attained to divinity by dint of mastery of yoga and mantra, respectively;85 sahaj (innate, natural) Mother goddesses, apparently born from the a
Characteristic of this material is its privileging of the internal, but not in manner that precludes actual performance of rites with a partner. Tantrasadbh va 15.25cd26ab, 144: a a duthn na siddhyanti tasm d dutm sam srayet 125 a . a sab hy bhyantar s tu j tavy kaulik nvaye | a a a a a a a ... aj tv dehaj m saktim bahihsth nagat m priye | a a a. a. . . a . a. s acaranti ca ye mudh h paavas te dur tmanah 144 a .
. 125 dutim ] em.; dut mss 144b bahih ] corr.; bahi mss .

Without consorts, they do not obtain siddhi; therefore one should resort to a consort. She should be known as both external and internal in the lineage (anvaya) of the Clans. . . . And without understanding the feminine power (akti) arising in the body, my dear, s those fools who resort to a feminine power in an external place are wicked-natured, bound souls.

Kaulaj nanirnaya 8.1617, 19cd22: a . pujitavy mah devyah ksetraj s tu vyavasthit h | a a a a. . . karavram mah k lam devikotyam var nane 16 . a a . a . . v r nasy m pray gam tu caritraik mrakam tath | a a. a. a . a a . at. hah sam jayant ca ebhih ksetrai ca ksetraj h | s a. .t a . . . . tesam madhye pradh n s tu ye j t h ksetraj h priye 17 a a a a. . a. . . ... ksetraj h kathit devi p. haj h kathay mi te 19 a. a t a . a . prathamam p. ham utpannam k m khy n ma suvrate | . t . a a a a upap. hasthit sapta devn m siddha- layam 20 t a a. a punah p. ham dvityam tu samj purnagirih priye | . . t . . a . . odiy nam mah p. ham upap. hasamanvitam 21 a t t . a . arbudam ardhap. ham tu upap. hasamanvitam | t . t p. hopap. hasandoham ksetropaksetram eva ca | t t . . . a p. h dy devat n m ca srnu puj vidhim priye 22 t a a a a. . . .

16a mah devyah ] em.; mah devy Kjncod Kjned a a a 16b ksetraj s ] em.; ksetraj Kjncod ; ksatraj a a a . . . . Kjned 16c karavram ] Kjned ; kanavram Kjncod 16d devikotyam ] Kjned ; deviko( - ?)m Kjncod 17d . . . . ksetrai ca ksetraj h ] Kjncod ; ksatrai ca ksatraj h Kjned s a. s a. 17e pradh n s ] Kjncodpc ; pradh nas a a a . . . . Kjncodac ; pradh nan Kjned 17f j t h ] Kjned ; j t Kjncod ksetraj h ] corr.; ksetraj Kjncod ; ksatraj a a a. aa a. a a . . . Kjned 19c ksetraj h ] corr.; ksetraj Kjncod ; ksatraj Kjned 19d pthaj h ] Kjned ; pthaj Kjncod 20b a. a a . a . . a . . . . . k m khy ] Kjned ; k m khyam Kjncod a a a a a 21b purnagirih ] Kjncodpc ; purnagiri Kjncodac Kjned 21c . . cod ; odiy na Kjned odiy nam ] Kjn 22a ardhaptham ] Kjncod ; arddham pthan Kjned . . 22d . a . . a . . ksetropaksetram ] Kjncod ; ksatropaksatram Kjned . . . .


Kaulaj nanirnaya 8.24: a . yog bhy sena ye siddh mantr nam ar dhanena tu | a a a a. a yogena yogaj m t mantrena mantraj h priye 24 a aa a. .
mantraj h ] corr.; mantraj Kjncod Kjned a. a

161 wombs of women who consume empowered caru in ritual;86 and the Sixty-Four goddesses comprising the core pantheon of the Kaulaj nanirnaya, whose mantras are a . inected forms of the eight syllables of its vidy -mantra.87 a Gaining visionary encounters with yogins and assuming their powers remains a signicant aim of ritual practice in the Kaulaj nanirnaya,88 alongside this texts more a . characteristic emphasis on attaining bodily immortality. Yogic practices take on far greater importance than in the BraY and Siddhayogevarmata, sources which place a s comparable emphasis on e.g. worship of the deities (y ga) and re sacrice (homa). a Thus while the BraY devotes a chapter to the secret nectars (guhy mrta)impure a a . liquids, especially sexual uids and menstrual dischargethe Kaulaj nanirnaya is a . more concerned with internal, yogic nectars.89 Nonetheless, the Vidy ptha concern a . with impure substances registers a continued presence, especially in Kaulaj nanira naya 11, a chapter devoted to nondual ritual cuisine (caru).90 Perhaps the most .

Kaulaj nanirnaya 8.2526: a . sahaj m tar devyo ruruyuddhe mah bal h | a a a a a. bhaksitam tu carum divyam saptajanm ntikam paum a s . . . . . . a a. tesam garbhe prasut n m niry sapr sitena ca | a a . garbhe j tena devei garbhe j nanti atmanah 26 a s a . 25

25a devyo ] em.; devy Kjncod Kjned a 25b ruruyuddhe ] Kjncod ; ruruyuddhair Kjned cod ; garbham Kjned garbhe ] conj.; garbha Kjn .



Kaulaj nanirnaya 8.3133ab: a . as. adh tu likhed vidy m pratham s. akabheditam | a. a. t .t a yath ekam tath sarve j tavy yoginkramam 31 a a a a . as. as. akavidh nena catuhsas. i yath kramam | a a . t . t . . .t yoginmelakam cakram anim digunas. akam 32 . . . a . . t a bhavaty eva na sandeho dhy napuj ratasya ca | a
31a vidy m ] em.; vidy Kjncod Kjned a. a 31c ekam ] Kjncod ; etat Kjned 32a astas. .. . takavidh nena ] Kjncodpc ; astastavidh nena Kjncodac ; astastakam vidh nena Kjned a a a 32d a. . .. .. .. .. nim di ] Kjned ; anim di Kjncod 33a eva ] Kjned ; evam Kjncod a . a .

References to attaining the state/status (pada) of or equality (s m nya) with yogins occur a a throughout the text. Additionally, several references are made to union/encounter (melaka) with the goddesses; see for instance Kaulaj nanirnaya 11.7cd10, quoted later in this chapter (n. 165). a . 89 Both BraY xxii and xxiv have the title guhy mrtapatala (Chapter on the Secret Nectars) provided a a . . in their colophons; the former, however, is concerned with such uids only in its nal section. The imagery of uids is prominent in the yogic visualizations of the Kaulaj nanirnaya. Note, for instance, a . the yoga of the short fth chapter, having conquest of death (mrtyujaya) as its aim, while chapter . fourteen speaks of churning nectar from a cakra of goddesses through yoga, effecting immortality (devyo a bhutv ca yoginyo m trcakr vaanug [h] | lyante khecarcakre ksobhayet param mrtam amrtena vin devi a. a s a . a . a . . amaratvam katham priye, 14.9394ab). . . 90 See especially Kaulaj nanirnaya 11.32: a .


162 dramatic shift is the occlusion of the cremation ground and mortuary rituala signicant departure, considering the pronounced mortuary and exorcistic dimensions of the Vidy ptha.91 a . The prominence of the cult of yogins in the Kaulaj nanirnaya appears unexcep a . tional in the Kaula scriptural corpus. Such is true, for instance, of the Kubjik texts of a the Western Kaula (pacim mn ya), some examples from which were cited in cons a a nection to temples of the yogins in chapter two. More comprehensive analysis of the relevant sourcesas yet poorly surveyed and largely unpublishedis unfortunately beyond the present study, although eminently worthwhile.

3.4 yogins in early buddhist tantric literature i

Parallel to the Saiva tantras of the Vidy ptha and Kaula emerged a corpus of Tantric a . Buddhist scripture devoted to a cult of yogins, deities whose signicance the tradi tion makes explicit by classifying this literature, according to one of the most common schemas, as yogintantras: Tantras of the Yogins.92 This corpus of scripture and its exegetical traditions represent the last major wave of Buddhist literary production in India, and the liturgies, deities, and meditational systems of the yogintantras dom inate the latter centuries of Indian Buddhismthe form in which the religion was transmitted to Tibet. Much as the literature of the Saiva yogin cult is marked by a shift from Sad siva to Bhairava as supreme deity, the mandalas of the Buddhist a ..
yasmin nispadyate pindam raktaukram pibet sad | s a . .. . . siddh n m yoginn ca ima carum priyam sad 32 a a. a a . . One should always drink [menstrual] blood and semen, in [i.e. from] which the body (pinda) is produced; this caru is dear to the siddhas and yogins. .. The same chapter refers to two types of caru consisting of ve substances: for the daily and occasional rites (nitya and naimittika), caru consists of the ve nectars of feces, urine, semen, [menstrual] blood, and marrow (vis. ham dh r mrtam sukram raktamajj vimiritam, Kaulaj nanirnaya 11.11ab). For the rites aa . . a s a .t . . . for special aims (k mya), the caru consists of ve cow products: beef, ghee, blood, milk, and yogurt a (gom msam goghrtam raktam goksra ca dadhim tath , Kaulaj nanirnaya 11.12cd). The same chapter mena. . a a . . . . . . tions numerous other powerful substances used as offerings, or consumed. 91 Although apparently optional, skulls do retain a place in ritual: after listing a number of alternatives, chapter twelve describes the Skull of Vi v mitra (i.e. a brahmin) as the best of ritual vessels s a (p tra; Kaulaj nanirnaya 12.13). a a . 92 On the complex subject of the classication of the Buddhist tantric canon, see Anthony Tribe, Mantranaya/Vajray na: tantric Buddhism in India, 20217. a

163 yogintantras center not upon Mah vairocana, supreme Buddha of the earlier yoga a tantras, but upon divinities of the vajra family (kula) presided over by the Buddha Aksobhya. The iconography of these deities is frequently k p lika, while their mana a . . dalas attest increasing emphasis on goddesses, including consorts of the Buddhas. It . is within the scriptures and practice systems centered upon Aksobhyas subsidiary . deities, especially erotic, k p lika Buddhas such as Cakra amvara, that a Buddhist cult a a s . of yogins comes into evidencemodelled in signicant ways, Sanderson argues, on that of contemporaneous Saivism.93 A distinctive aspect of the Buddhist yogin cult is terminological: while in Saiva and earlier Buddhist literature the term dakin generally describes a vile, often vam. piric variety of female being, the Buddhist yogintantras by and large treat this word as a synonym of yogin. This elevation of the dakin is consonant with Buddhist . precedents for conversion and incorporation of hostile deities, noteworthy examples of which include the early traditions assimilation of yaksas and yakss, and the . . Mother goddess H rit. Within tantric Buddhist literature, transformations in cona ceptions of dakins and related female deities, especially the Seven Mothers, appear . to provide key indicators for the emergence of a Buddhist cult of yogins. Not a specialist in this material, in the following pages I nonetheless attempt a provisional mapping of aspects of this process, limited by my reliance upon the scholarship of others and lack of competence in Tibetan and Chinese. Signicant uncertainties surround the chronology of Buddhist tantric literature, though attenuated by the assistance Chinese and Tibetan sources offer in dating specic works. Of particular value, we know the periods of early learned authors such as Buddhaguhya and Vil savajra, active in the mid and late eighth century, respectively, a who quote or comment upon tantric scriptural sources; for extant, datable Saiva commentaries, we must on the other hand wait until the tenth century, although SadyoSanderson, Vajray na: Origin and Function, passim. Some of the textual evidence for his thesis a is discussed in the subsequent section.

164 jyotis probably lived considerably earlier.94 As is well known, proto-tantric Buddhist literature of the variety later classied as kriy tantras survives from the early cena turies of the common era, often only in Chinese translation. Concerned largely with accomplishing worldly aims, this literature contains much that is characteristic of later tantric ritual, yet without articulating mantra-practice within a Mah y na sotea a riological framework.95 Evidence for a developed tantric literature and eye-witness reports concerning the prevalence of tantric Buddhist traditions in India emerge only in the middle or latter half of the seventh century.96 No cult of yogins is yet evident in the Mah vairocan bhisambodhisutra, one of the a a . few extant Buddhist texts of the transitional variety classied as cary tantras, similar a in many respects to the subsequent yogatantras but appearing to lack a developed soteriological vision of tantric ritual.97 Composed, according to Stephen Hodge, around 640 c.e. or somewhat earlier, this survives only in Chinese and Tibetan translations.98 Prominent in the mandala of the supreme Buddha Mah vairocana, as delineated in a ..
94 On the dating of Buddhaguhya, see Stephen Hodge, The Mah -vairocana-abhisambodhi Tantra with a . Buddhaguhyas Commentary, Introduction, 2223. Concerning Vil savajra, I follow Ronald Davidson, a The Litany of Names of Maju r: Text and Translation of the Majurn masamgti, 67. Although s s a . almost certainly a pre tenth-century author, little concerning the dating of the prolic, inuential, and perhaps quite early Saivasiddh nta exegete Sadyojyotis can be said with certainty. He was known to a Som nanda (early tenth-century), and his commentary on the Sv yambhuvasutrasamgraha appears to be a a . paraphrased in the Haravijaya (circa 830 c.e.), while in his critique of the Advaitaved nta, he displays a no awareness of the vivartav da or illusionism associated with Sankara (. c. 800 c.e.?) that came to a dominate this school. On this and other issues pertaining to the dating of Sadyojyotis, see Alex Watson, The Selfs Awareness of Itself: Bhat. a R makan. has Arguments Against the Buddhist Doctrine of No-self, 111 .t a .t 14. Watsons conclusion is that a seventh or early eighth century date is more likely than a late eighth or early ninth. Ibid., 114. 95 Hodge provides a valuable account of the chronology of the Chinese translations of early tantric literature. Mah -vairocana-abhisambodhi Tantra, Introduction, 58. The Buddhist kriy tantras in all a a . likelihood drew upon ancient and perhaps nonsectarian magical traditions, such as the vidy practices a discussed in the previous chapter in the section on the Vasudevahind. . . 96 Hodge points out that a Chinese traveller, Xuan-zang, gives no indication that tantric traditions were prevalent in India in the period up to 645 c.e. On the other hand, there are rst-hand reports concerning tantric practices and scripture from the latter half of the century. Ibid., 911. 97 See Tribe, Mantranaya/Vajray na, 20710. Hodge, offering a different assessment of the soteria ological dimension of the Mah vairocanasutra, considers this text likely to have been one of the rst, if a not actually the rst fully developed tantra to be compiled, that has survived in some form to the present day. Mah -vairocana-abhisambodhi Tantra, Introduction, 29 (quotation), 3339. In my discussion of this a . text, I rely entirely upon Hodges English translation from the Chinese and Tibetan. 98 Concerning the dating, see Hodge, ibid., 1417. Translated into Chinese in 724 c.e., it appears that a copy of the Mah vairocanasutra was among the manuscripts collected by Wu-xing in India at some a point during the eight years prior to his death in 674.

165 the second chapter, are goddesses such as T r . More signicant to the present study aa are references to the Mother goddesses: in the same mandala appear wrathful Moth.. a a ers headed by the goddess K lar tri, who form the retinue of Yama, lord of Death and guardian of the southern direction. This set is elsewhere identied as K lar tri, a a Raudr, Brahm, Kaum r, Vaisnav, C munda, and Kauberan unusual heptad of a . . .. a Mother goddesses.99 That they are tantric divinities, however minor, is evidenced by occurence within the mandala and their invocation by mantra.100 K lar tri and seven a a .. a unspecied Mothers also gure in the entourage of S kyamuni,101 while elsewhere Mothers are included in an enumeration of potentially dangerous spirits.102 Chapter six links them to mantras for causing illness, bridging the goddesses roots in the mythology of Skandas grahas with tantric magical practices.103 Furthermore, as do the Niv satattvasamhit and a variety of other tantric sources, the Mah vairocanasutra s a a . a lists Mother shrinesas well as temples of Sivaamong the places appropriate for performing solitary s dhana, but without cultic emphasis on these deities.104 a Besides Mother goddesses, the Mah vairocanasutra contains several references a to dakins and female divinities such as the yaksin, while the texts supplement . . . tantra (uttaratantra) describes rites for bringing the latter and female denizens of the
Mah vairocanasutra ii.50 mentions wrathful Mothers in the retinue of Yama; these deities are a named in xiii.89. Even without consulting the Tibetan or Chinese, I would assume that wrathful Mothers translates the Sanskrit rudram tarah. That this refers specically to the Seven Mothers is suga . gested by Ksemar jas explanation of the term as it occurs in Netratantra 2.13c (he glosses rudram tarah a a . . with brahmy dy sBrahm, etc.). The present heptad is unusual insofar as C mundas preeminent a a a . . position is usurped by K lar tri, who appears to be identied with Y m, the female counterpart of a a a Yama. The identication of Y m with K lar tri is suggested in the Chinese translation of i.19; see a a a Hodges note thereon (p. 63). Y m and V r h alternate in textual accounts of the Seven Mothers, a aa while sculpted sets appear as a rule to depict V r h; see chapter 2, n. 9. Also unusual is that Kauber aa replaces Indr n/Aindr in the present heptad. a. 100 Note also their association with a series of drawn insignia (mudr ), as with the other mandala a .. deities (xiii.89). While K lar tri is invoked with her own mantra, the others are paid reverence with the a a . . generic namah samantabuddhanam matrbhyah svaha (iv.11). . . 101 See Mah vairocanasutra iv.11. a 102 Mah vairocanasutra xvii.13; also mentioned are, e.g., piacas and r ksasas. a s a . 103 Cf. vi.15: Then, for example, the Asuras manifest illusions with mantras. Or, for example, there are [mundane] mantras which counteract poison and fevers. Or else there are the mantras with which the Mothers send sickness upon people. . . . 104 Lists of suitable locations are present in v.9 and vi.30. In Mah vairocanasutra, Uttaratantra iii.2, a Mother shrines are listed among the places appropriate for re sacrice having as its goal subduing (Sanskrit vakarana, presumably). s .

166 netherworlds under ones power.105 While in yogintantras of the subsequent period dakins would become prominent deities, identical with yogins, the Mah vairocana a . sutra groups them with minor, potentially pernicious beings such as the r ksasa, a . yaksa, and piaca. This appears consistent with early non-Buddhist conceptions of s . the dakin. No evidence for the gure of the yogin is present, although the vocative . case epithets yogini and yogevari appear in a mantra; the deity is not named.106 In s the Mah vairocanasutra we hence nd evidence for interest in some of the divinities a prominent in the cult of yogins, particularly a limited appropriation of the Mothers as tantric deities. This accords with roughly contemporaneous sculptural evidence for Buddhist interest in these goddesses, for a shrine of the Mothers is present in the Buddhist cave temple complex at Aurangabad.107 The Majurmulakalpa attests a similar, yet broader range of female deities and s spirits. Classied within the tradition as a kriy tantra, a portion of this heterogeneous a text has been shown to herald from the middle of the eighth century, the period in which some sections appear in Chinese translation.108 In its opening chapter, the Majurmulakalpa enumerates a vast pantheon of divine, semi-divine, and human s beings who assemble to hear the Dharma, among whom are an array of female di a . a a . vinities that include putan s, bhagins, dakins, rupins, yaksins, and ak sam trs, Sky . . . Mothers.109 This list is highly suggestive of the range of female divinities described in literature of the yogin cult. Although they are not prominent in the ritual of this text, the Majurmulakalpa, like the Mah vairocanasutra, positions the Seven Mothers s a
105 A short series of mantras for minor divinities and spirits such as r ksasas, dakins, and asuras is a . . provided in iv.16, while mudr s and mantras for a larger series, including dakins, are listed in xi.98 a . 99. A list of dangerous beings in the uttaratantra includes both dakins and what Hodge translates as . witches (iv.1). As described in iii.9 of the uttaratantra, through re sacrice one may draw to himself yaksins and likewise girls of the subterranean realm with the male and female assistants. . . 106 xv.10; the mantra for the Mudr of Upholding the Bhagavats Yoga is given as namah samantaa . . buddhanam mahayogayogini yoge vari khajalika svaha. s 107 See the discussion of post Gupta-era Mother temples in chapter 2 of this dissertation. 108 Yukei Matsunaga, On the Date of the Maju rmulakalpa, in Michel Strickmann, ed., Tantric and s Taoist Studies in Honour of R.A. Stein, 22 (1985): 88294. 109 Each of these beings is said to have ordinary and great (mah -) varieties, and many of the latter a are listed by name; the Great [Sky] Mothers include the standard Seven augmented by Y my , V rua a a n, Putan , and others, with retinues of innumerable nameless Mothers. Majurmulakalpa 1, vol. 1, p. a s . 2021 (Trivandrum Sanskrit Series edition).

167 in the retinue of Yama among the non-Buddhist deities in the outer layers of the mandala.110 The effort to give them a Buddhist identity is suggested by the addition of .. Vajrac mundi to their ranks.111 In general, however, the depiction of the Mothers is a .. more consonant with the ancient cult of Skandas countless grahas, with whom their connection is made explicit.112 As for dakins, their characterization is entirely that of . pernicious, possessing female spirits, against whom one requires mantras for protection; no indications are present of the positive associations and prominence assigned to them in yogintantras. One vidy -mantra, for instance, is said to have the power to a conjure a yaksin, or else to destroy dakins.113 Of additional interest in this tantra is its . . . incorporation, as tantric deities, of Tumburu and the Four SistersJay , Vijay , Ajit , a a a Apar jit the core pantheon of the early yet largely lost Saiva v matantras. Chapters a a a forty-seven to forty-nine are devoted to practices connected with these deities, and include the tale of their conversion to Buddhism.114 Further developments towards a cult of yogins are evident in the Sarvatath a gatatattvasamgraha, among the earliest extant scriptures classied as yogatantras and . representative of a developed Buddhist soteriological vision of tantric ritual. Its composition had apparently commenced by the last quarter of the seventh century, while
The Seven Mothers (precise identities unspecied) occupy a position in the southeastern direction, adjacent to Yama in the south, and are also among the deities around the perimeter of that layer of the mandala; their company includes major brahmanical gods, gana-lords such as Mah k la, sages, a a .. . Tumburu and the Four Sisters, the Planets, and so forth. Majurmulakalpa 2, vol. 1, p. 4445. s 111 Majurmulakalpa 45 provides mudr s connected to and named after the Mothers, and includes s a both C mundi (45.229cd30ab) and Vajrac mundi (45.228cd229ab). Vol. 2, p. 510. a a .. .. 112 Most of the Majurmulakalpas copious references to the Mothers point toward their identity as s dangerous female spirits, and only rarely the seven brahmanical goddesses. The Mothers are mentioned among the spirits by whom one may become possessed, alongside beings such as the piaca and dakin; s . see for example Majurmulakalpa 3, vol. 1, p. 53, and chapter 9, vol. 1, p. 82. Cf., e.g., 22.229, in a s vivid description of the activities of Mother goddesses (verse numbers here and elsewhere as per the reprint edited by P. L. Vaidya; vol. 1, p. 249 in the tss edition). The Mothers of Skanda (skandam tr) a. are mentioned in 22.24b (tss edition vol. 1, p. 233)a chapter rich in its accounts of beings fabulous and dangerous. 113 Majurmulakalpa 2.45, vol. 1, p. 30. Among the many other references to dakins, note for instance s . a curious rite to remove the breasts and genitalia of proud, wicked dakins and women. Used on a man, . it removes the penis and facial hair, and causes breasts to appear. Chapter 52, vol. 3, p. 56364. 114 The vidy -mantras of these deities are rst given in 2.1517, where they are said to be attendants a of the Bodhisattva (bodhisattv nuc rik [h], 2.16b). Vol. 1, p. 32. Majurmulakalpa 47 presents a brief a a a . s narrative of their taking refuge in the Dharma, after which begin instructions on their worship.

168 the text as we have it was translated into Chinese in 753.115 Although the Tattvasamgraha thus does not necessarily postdate the Majurmulakalpa, it takes the cons . version of goddesses considerably further, and its range of female deities even more clearly intimates that of the yogintantras. Here, for instance, we nd reference to Mother goddesses classied under the categories antarksac ri (aetherial), khecar . a (aerial), bhucar (terrestrial), and p t lav sin (denizens of the netherworlds) aa a closely related to categories applied in later classications of yogins. Along with a host of other erstwhile hostile deities, headed by Siva, Vajrap ni confers upon them a. tantric initiation and initiatory names; thus J tah rin becomes Vajramekhal , M ran a a . a a . becomes Vajravilay , Kauber becomes Vajravikata, and C munda becomes Vajrak l, a a a . . . to name one from each respective class.116 The latter k p lika goddess is once ada a dressed as Vajradakin.117 Leaving behind their identities as grahas of Skanda or as . maternal, brahmanical goddesses, the Mothers here take on identities as goddesses of the Adamantine Vehicle, the Vajray na. a In the Tattvasamgraha, we are presented with perhaps the earliest narrative of the . conversion and accommodation of dakins. Charged with quelling wicked beings, . Vajrap ni utters the Heart Mantra for Drawing Down All Dakins and other Wicked a. . Possessing Spirits, upon which the dakins and other grahas assemble in a circle and . supplicate. Undoubtedly concerned by the dietary restrictions their new allegiance will entail, they beseech, we eat meat; hence order [us] how [this matter] should be understood.118 Advised by Vajrasattva, the supreme Buddha, the compassionate
115 Elements of this text were introduced in China by an Indian, Vajrabodhi, who would have learnt the teachings around 700 c.e.; Amoghavajra translated the text in 753. See the discussion of Hodge, Mah -vairocana-abhisambodhi Tantra, Introduction, 1112. a . 116 Tattvasamgraha 6, p. 173 (lines 321). I cite the text from the edition of Isshi Yamada: Sarva. tath gata-tattva-sangraha. A critical edition based on a Sanskrit manuscript and Chinese and Tibetan translations a (New Delhi: Satapitaka Series, vol. 262, 1981). On the classication of yogins into aerial, terrestrial, . and so forth, cf., e.g., the Saiva Kulas ra, as discussed in the entry dikcar, etc. by Judit Trzsk, in a T ntrik bhidh nakoa, vol. iii (forthcoming). a a a s 117 Tattvasamgraha 14, pp. 3067 (lines 1014, 14); C munda/Vajrak l is also addressed as e.g. kap laa a a . . . m l lankrt (adorned with a garland of skulls) and vajrakhatv ngadh rin (bearer of a vajra and skullaa . a a . . a staff). 118 Tattvasamhgraha 6, p. 18081 (lines 817, 13): .

atha vajrap nir mah bodhisattvah punar api sarvadakiny didus. agrah karsanahrdayam abh sat a. a a a . . . a. . . .t i | om vajrakarsaya sghram sarvadustagrahan vajradharasatyena hum jah . . . .. . .

169 Vajrap ni provides appropriate means, saying thus: through this mudr , you may a. a extract hearts from all living beings and eat them.119 The episode, a conversion story of sorts, suggests growing concern with the gure of the dakin, and perhaps . also the entry of mantra techniques associated with them into the battery of those available to practitioners. An early eighth-century Chinese commentary on the Mah a vairocanasutra provides a closely related narrative, wherein the association of dakins . and their practices with Siva and Saivism is made explicit.120 While this signals a process of providing Buddhist identities to dakins and connected practicespresumably .
ath smin bh sitam tre dakiny dayah sarvadus. agrah h sumerugirimurdhni b hyato mandala a. a a a. a . . .t .. a bhutv vasthit iti a atha vajrap nir mah bodhisattvah t m dakiny dn sarvadus. agrah n a. a a. . a t a . . ahuyaivam aha | pratipadyata m rsah pr natip tavairamanyaiksasamayasamvare m vo vajrea . . a. a s . a . . nadptena pradptenaikajv lbhutena kul ni nirdaheyam | atha te dakiny dayah sarvadus. agrah a a a a . . . .t yena bhagav n ten jalim baddhv bhagavantam vij pay m asuh | vayam bhagavan m ms sinas a a a a a . a . a . . . a tad aj payasva katham pratipattavyam iti . Then Vajrap ni, the great Bodhisattva, again spoke the Heart Mantra for Drawing Down a. All Dakins and other Wicked Possessing Spirits: om vajra quickly draw down all wicked . . possessing spirits by the word of Vajradhara hum jah! Then, as soon as this had been . . uttered, all the dakins and other wicked possessing spirits formed an outer circle on . the summit of Mt. Meru and remained there. Then Vajrap ni, the great Bodhisattva, a. summoned the dakins and other wicked possessing spirits, and said, Resort, O friends, to . the assembly of the pledge of teaching abstention from slaughter, lest I should incinerate your clans with my burning vajra, [when it has] become a single, blazing ame. Then the dakins and other wicked possessing spirits, folding their hands to where the Lord was, . entreated the Lord: O lord, we eat meat; hence order [us] how [this matter] should be understood. Concerning vairamanya, see its lexical entry in Edgerton, Buddhist-Hybrid Sanskrit Grammar and Dictio. nary, vol. 2. 119 Tattvasamgraha 6, p. 181 (lines 412, 1518): . atha bhagav n vajrap nim evam aha | pratipadyasva vajrap ne esam sattv n m mah karua a. a. a a. a . . nam utp dyop yam d tum iti | atha vajrap nir mah k runika idam sarvasattvamarananimittaa a . a a. a a . . . . j namudr hrdayam abh sat | om vajra pratigrhna hrdayam akarsaya yady ayam a a . a. . . . . . . . sattvo masad ardhena mriyate tad asya hrdayan niskramatu samaya hum jjah . . . ath sya mudr bandho bhavati | . . . anay mudray bhavadbhih sarvasattvahrday ny apakrsya a a a a a . . .. bhoktavy nti | atha te dakiny dayah sarvadus. agrah hulu hulu praksvedit ni krtv svabhavanam a a a . . .t . . a . a gat iti a Next, the Lord spoke to Vajrap ni thus: O Vajrap ni, after generating great compassion a. a. for these beings, assent to give them a means. Then Vajrap ni, possessing great compasa. sion, spoke this, the Heart Mantra of the Mudr for Knowing the Deaths of All Living a Beings: om vajra seize extract the heart if this being dies within a fortnight then let its . . heart emerge samaya hum jjah. Now this is the binding of the mudr : . . . Through this a . mudr , you may extract hearts from all living beings and eat them. Then the dakins and a . other wicked possessing spirits made clamorous hulu hulu sounds and returned home.
120 This passage from the commentary of Subhakarasimha and his disciple Yixing is translated and . discussed by Gray, Eating the Heart, 4749. The commentators remarks concern Mah vairocanasutra a iv.16, mentioned above (n. 105).

170 similar in nature to those described in the lost Saiva dakintantras referred to by Dhar. makrtithere is as yet little indication in the Tattvasamgraha of their transformation . into the wild and ambivalent, yet supremely powerful and potentially benecent sky-wanderers of the yogintantras. A scripture composed perhaps in the latter half of the eighth century, the Guhyasam jatantra evidences a marked increase in engagement with the erotic and the ima pure, intimating developments carried even further in the yogintantras. Its ritual has a signicant k p lika dimension and incorporates both coitus and ingestion of impure a a substances, while erotic imagery distinguishes the iconography of its deities.121 Focused upon the Buddha Aksobhya, patriarch of the vajra-clan deities, the transitional . status of this and closely related literature is reected in its classication, frequently, as neither yoga- nor yogin-, but mah yogatantras.122 In chapter seventeen of the Guhya a sam ja occurs an important early reference to vajradakinstransformations of these a . hostile beings into wielders of the vajra sceptre, marking their entry into the Vajray na pantheon. Vajrap ni discloses a series of initiatory pledges (samaya) connected a a. with specic deities, among whom are female beings: yaksins, queens of the snake . . women (bhujagendrar j), asura maidens, r ksass, and vajradakins.123 The Pledge of a a . . All Adamantine Dakins binds one ever to the consumption of urine, feces, blood, . and alcohol, and to magical slaying;124 this suggests as yet little fundamental trans On the dating of the Guhyasam ja, I follow Isaacson, who cites the discussion of Yukei Matsunaga, a The Guhyasam ja Tantra, Introduction, xxiiixxvi; this edition has not been available to me. Isaaca son points out that the iconography of the Guhyasam ja is not yet k p lika, although erotic. Tantric a a a Buddhism in India (from c. ad 800 to c. ad 1200) (unpublished lecture transcript, Hamburg 1997), 4. 122 See Tribe, Mantranaya/Vajray na, 21013. a 123 Guhyasam ja xvii, p. 130 (Gaekwad Oriental Series no. 53, Benoytosh Bhattacharya, ed.). a 124 Guhyasam ja xvii, p. 130: a atha vajrap nih sarvatath gat dhipatih sarvavajradakinsamayam svak yav kcittavajrebhyo a. . a a a a . . . nic ray m asa | s a a vinmutrarudhiram bhaksed mady dms ca pibet sad | a . a . . . vajradakinyogena m rayet padalaksanaih a . . . . a svabh venaiva sambhut vicaranti tridh tuke | a a acaret samayam krtsnam sarvasattvahitaisina . . . . . sarvatraidh tukasamayasamavasarano n ma sam dhih | a a . . a Next, Vajrap ni, lord of all Buddhas, sent forth from the vajras of his body, speech, and a. mind the Pledge of All Vajradakins: .

171 formation in conceptions of dakins, despite their conversion. . Some evidence points toward the emergence of actual yogintantra material in the eighth century, separated little in time from the yogatantras. Amoghavajra wrote a description of the Sarvabuddhasam yogadakinj lasamvara, probably the earliest of this a . a . genre, after his return to China in 746 c.e.125 But this transitional text, referred to in some scholarship as a proto-yogintantra, was only retrospectively grouped with the yogintantra corpus;126 it seems likely that most of the yogintantra literature dates to the ninth century and beyond. David Gray, however, suggests that the late eighthcentury commentator Vil savajra quotes one verse and paraphrases another from the a Laghucakraamvaratantra or Heruk bhidh na, probably the earliest and most authoras . a a tive scripture in the cycle of yogintantras focused upon Cakra amvara. While this s . is not implausible, the evidence awaits publication; Vil savajra does not apparently a name the source for the verses in question.127 This issue is of considerable importance, not only for the history of Tantric Buddhism, but because the terminus ante quem of the BraY might depend upon the dating of the Laghuamvara, as discussed a s . subsequently. The Buddhist yogintantras and their exegetical literature constitute a vast corpus,
One should always eat urine, feces, and blood, and drink wine and so forth. One should slay () through the vajradakin yoga, through padalaksanas. Arisen . . . by their very nature, they [dakins?] roam the triple universe (?). One should . observe this pledge wholly, desiring the good of all beings. [Then Vajrap ni entered?] the meditative trance called The Assembly of the Entire Triple a. Universe. Aspects of this seem puzzling; vajradakinyoga might refer to the invasive yogic processes by which . dakins prey upon victims. See the discussion of the M latm dhava in chapter 2. padalaksanaih suga a . . . . gests no plausible interpretation to me, while the interpretation of the next verse-quarter is unclear as well. Candrakrti, commenting on this verse, glosses vajradakinyogena with the yoga of Gaur, . etc. (gaury diyogena). His remarks on padalaksanaih are unfortunately corrupt, but include clear refera . . . ence to the parasitic practices of dakins (padalaksanaih dus. an m udyarakt k[r]s. y diprayogaih marayet, a . .t a . . t a . . . . . One should slay with padalaksanas, i.e. the application of . . . extraction of blood from the wicked). . . Pradpodyotana, p. 206. 125 Rolf Giebel, The Chin-kang-ting ching y-chieh shih-pa-hui chih-kuei: An Annotated Translation, Journal of the Naritasan Institute for Buddhist Studies 18 (1995): 17982. 126 English, Vajrayogin, 5. 127 Gray, Eating the Heart, 54 (n. 38). He refers to his forthcoming study and translation of the Laghuamvara for a detailed discussion, remarking that most of the Laghuamvara quotations Ronald s . s . Davidson had identied in Vil savajras work come, in fact, from the Sarvabuddhasam yogadakinj laa a . a samvara. See Davidson, The Litany of the Names of Majur, 67. s .

172 much of which survives only in Tibetan translation and relatively little of which has been published, in cases where the Sanskrit original is preserved. Among the most important yogintantras are the Laghuamvara and Srhevajradakinj laamvara (i.e. the s . . a s . Hevajratantra), texts considered foundational to the systems of practice and cycles of scripture focused upon the Buddhas Cakra amvara and Hevajra, respectively. Other s . important texts of this genre include, for instance, the Candamah rosanatantra and Krsa . . .. .. nayam ritantraalthough the latter is technically considered a mah yogatantra128 a a . texts teaching the cults of their namesake Buddhas. While the dating of the major yogintantras is problematic, they undoubtedly belong to the period prior to the Laghuk lacakra and its important commentary, the Vimalaprabh , which date between a a 1025 and circa 1040 c.e., as John Newman shows convincingly.129 Perhaps one of the earliest of all, the Laghuamvara might have existed in the latter eighth century, s . as Gray suggests, while its earliest commentator, Jayabhadra, probably wrote in the mid-ninth century.130 The cult of yogins thoroughly permeates the literature and ritual of the Cakra samvara tradition. By way of illustration, I shall take the Laghuamvara as an exam . s . ple of the content of the yogintantras, for this happens to be a text with a signif icant relationship with the BraY a relationship addressed in the subsequent seca tion. In the Laghuamvara, the cult deities comprise a k p lika Buddha, Cakra amvara s . a a s . or Heruka, and his sow-faced consort, Vajrav r h or Vajrayogin, who preside over aa a mandala primarily of twenty-four goddesses referred to as dakins, vajradakins, .. . . or duts (consorts).131 While the mandala dakins have male counterparts in the .. . twenty-four heroes (vra), the latter have only secondary signicance.132 The Laghu samvaras dakins are fully representative of the yogin typology described in chapter . .
Isaacson, personal communication (May, 2007). Newman, The Epoch of the K lacakra Tantra, Indo-Iranian Journal 41 (1998): 31949. a 130 David Gray presents evidence suggesting that Jayabhadra, the third abbot of Vikramala, was s active in the mid-ninth century. Eating the Heart, 62 (n. 65). 131 The primary mandala is described in chapter 2 of the Laghuamvara, while the twenty-four dakins s . .. . are listed in chapter 4. 132 Mentioned rst in 2.19cd, the vras are not named until chapter forty-eight.
129 128

173 1 of this thesis, combining in their k p lika, theriomorphic iconography images of a a power and eroticism. They pervade the universe,133 a wild horde with names such as Khag nan (Bird-face), Sur bhaks (Drunkard), Cakraveg (Wheel-speed), a a a a . V yuveg (Wind-speed), Mah bal (Mighty), Mah n s (Big-nose), and Cana a a a a a a . daks (Grim-eyes). All but the rst two of these names are held in common with . . goddesses mentioned in the BraY , indicative of the shared Saiva-Buddhist image of a the yogin or dakin.134 . As goddesses of the clan of Vajrayogin/Vajrav r h, the Laghuamvaras twenty aa s . four mandala dakins represent a single class from a broad spectrum of female beings .. . with which the cult is concerneddeities whose principle varieties are the yogin, dakin, rupin, l m , and khandaroh .135 Collectively, they comprise the web or maa . . a a .. trix (j la) of dakins that pervades the universe. This has its reection in the great a . mandala of deities (mah cakra) described in chapter forty-eight, the abode of all a .. dakins (sarvadakiny laya); based upon the heart mantra of all yogins, this incora . . porates goddesses of the ve classes together with the twenty-four male heroes. The whole constitutes the Assembly of the Matrix of Dakins (dakinj lasamvara),136 . . a . and the supreme Buddha himself, Vajrasattvathe highest Blissconsists of all dakins.137 The nature of the goddesses manifestation and movement (sac ra) on a . the earth forms a central focus, reected in the several chapters the Laghuamvara s . devotes to typologies of the clans (kula) of goddesses: chapters sixteen to nineteen, and twenty-three. While in the yogatantras deities were organized according to clans
133 Laghuamvara 4.1ab, . . . dakinyo bhuvan ni vijrmbhayanti. Cf. 41.16ab, caturvimsatidakiny vy ptam s . a a a . . . . . trailokyam sacar caram. a . 134 The names of the twenty-four are given in Laghuamvara 4.14. While Khag nan has no precise s . a a counterpart in the BraY , for avian imagery, note Lohatund, Iron-beak. Sur bhaks too does not a a . . . gure in the BraY ; however, the principal Six Yogins are said to be fond of alcohol (madir savapriy a a a nityam yoginyah sat prakrtit h, liii.15ab). a. . . . . 135 Lists of the ve goddess classes occur in e.g. 13.3 and 14.2. Additional subcategories of dakins are . described in chapters 1619 and 23. The twenty-four mandala dakins are said to belong to the v r hkula aa .. . in 2.18cd (dakinya ca [ca]turvimsa v r hy h kulasambhav h). s a. . . a a a. 136 The description of the sarvadakiny laya (abode of all dakins) begins in 48.8, and is based upon a . . the pantheon of the hrdaya mantra stated in 48.3. The great cakra is described as the dakinj lasamvara . . a . in 49.16 (purvoktena vidh nena yajed dakinj lasamvaram | mah cakra[m] sarvasiddhy layam tath ; here I read a a a a . . a . . . as per the Baroda codex, f. 35v). 137 Laghuamvara 1.3ab: sarvadakinmayah sattvo vajrasattvah param sukham. s . . . . .

174 (kula) of the ve Buddhas of the Vajradh tu mandala, the Laghuamvara and similar a s . .. systems introduce new, matriarchal deity clans, much as Saiva yogins were classied according to clans and subclans of the Seven Mothers. The Laghuamvara devotes sevs . eral chapters to the subject of chomm as well, the secret verbal and nonverbal codes a for communication between practitioners and the deities, or between initiates mutually.138 Sacred geography forms a concern as well, a mapping of the powerful places where the goddesses are said to manifest.139 As with the Saiva Vidy ptha, the yogin cult of the Laghuamvara is thoroughly a . s . k p lika in character,140 and this texts rites of re sacrice utilize a battery of meats a a and other things impure, largely with aggressive magical aims.141 Prominent among the goals of ritual is attainment of encounters with dakins; to the heroic s dhaka, they a . may bestow the power of ight and freedom from old age and death.142 Enabled by the dakins, the s dhaka comes to traverse the entire world as their master.143 Signia . cant attention is devoted, furthermore, to rites of bodily transformation, a domain of magic characteristic of the shapeshifting, theriomorphic yogin.144

138 Chapters on chomm include Laghuamvara 15 (single-syllable chomm s), 20 (communication a s . a through pointing at parts of the body), 21 (similar gestures plus their correct responses), 22 (gestures made only with the ngers), and 24 (single-syllable and other verbal codes). 139 Lists of p. has occur in Laghuamvara 41, which associates specic sets of goddesses with these; and t s . Laghuamvara 50.2027. s . 140 Note, for instance, that the initiatory mandala described in chapter 2 is constructed with mortuary .. materials such as cremation ashes. 141 Particularly noteworthy are the homa rites described in Laghuamvara 50. s . 142 See for instance the brief chapter thirty-nine; the heroic s dhaka is promised attainment of the state a of a Sky-wanderer (nyate khecarpadam, 4b), and freedom from old age and death (na jar mrtyuh sarvatra a . . s dhako mantravigrahah, 5ab). a . 143 Laghuamvara 3.16: s . . dakinyo l maya caiva khandaroh tu rupin | a s a . .. etair vicared jagat sarvam dakinyaih saha s dhakah 16 a . . . . sarv h kinkars tasya s dhakasya na samsayah | a. a . .

Highly irregular grammatical forms such as etaih (masculine, for the feminine et bhih) and dakinyaih (for a . . . . . dakinbhih) are none too rare in this text, while the metrical irregularities of 16c and 17a are even more . typical. 144 Note in particular the rituals of Laghuamvara 49, which promise the yogin the power to transform s . himself at will (k marupo mah vrya yog sy n n tra samsaya, 49.14ab). a a a a .


3.5 buddhist and saiva yogintantras: the case of the i brahmayamala and laghucakra amvaratantra s .
In a pioneering article of 2001, Alexis Sanderson identied extensive parallel passages in tantric literature, within and across sectarian boundaries, and argued that substan tial portions of important Buddhist yogintantras were redacted from Saiva sources, largely unpublished.145 This constitutes some of the most important evidence mar shalled in support of his thesis concerning the historical relationship between Saivism and the esoteric Buddhism of the yogintantras, rst argued in an article of 1994 as serting that almost everything concrete in the system is non-Buddhist in origin even though the whole is entirely Buddhist in its function.146 While Sandersons examples concern several Buddhist texts, the most remarkable case is that of the Laghucakraamvaratantra or Heruk bhidh na, nearly half the contents of which he holds can s . a a be seen to have been redacted from Saiva originals found in texts of the Vidy ptha a . division of the bhairavatantrasthe BraY , Siddhayogevarmata, Tantrasadbh va, and a s a the Yoginsac raprakarana of the Jayadrathay mala.147 The implications are consider a a . able, for this would mean that one of the most fundamental scriptures of the latter phase of Indian Tantric Buddhism took shape, in large measure, through appropria tion of material from tantras of the Saiva yogin cult. Undoubtedly some of the most signicant historiographic questions concerning the cult of yogins lie in the dynamics of Saiva-Buddhist interaction, and the forma tion of parallel tantric ritual systems across sectarian boundaries focused, to a large degree, upon the gure of the yogin. For while there is much that is similar in older forms of Saivism and Tantric Buddhism, it is with the cult of yogins that parallels in ritual, text, and iconography reach their most remarkable level. Assessment of the enormous body of evidence relevant to these questions and its interpretation in light of the social and historical contexts of early medieval India shall require sustained
Sanderson, History through Textual Criticism in the Study of Saivism, the Pacar tra and the a Buddhist Yogintantras, especially 4147. 146 Vajray na: Origin and Function, 92. a 147 Sanderson, History through Textual Criticism; 4147 (quotation on 42).

176 scholarly engagement, admirably begun in the works of Sanderson, to whose work Ronald Davidson has offered a signicant rejoinder.148 In the present discussion I shall conne myself to aspects of textual history, primarily as concern the BraY , rather than attempt to address the larger picture of a Saiva-Buddhist interactions. The passages Sanderson identies as shared by the BraY and Laghuamvara belong to the rst portion of BraY lxxxv, a chapter ena s . a titled The Section on the Pledges,149 and the greater part of chapters twenty-six to twenty-nine of the Laghuamvara. He notes that chapter forty-three of the Abs . hidh nottaraa text of the Cakra amvara cycle, to which the Laghuamvara is fundaa s . s . mentalhas parallels in BraY lxxxv as well, while the Buddhist Samvarodayatantra a . has a section concerning the classication of skull-bowls parallel to a section in BraY a iv.150 To the passages identied by Sanderson I can add the nal ve verses of BraY a lxxxiv, which correspond to the opening verses of Laghuamvara 26 (table 3.1).151 s .
Davidson, Indian Esoteric Buddhism, 20218. The colophon reads, in Bya , samay dhik ro n ma c stimah patalah. a a a a . . . 150 Sanderson, Vajray na: Origin and Function, 95. In the case of chapter fteen of the Samvarodaya, a . the p tralaksananirdeapatala that Sanderson describes as closely related to a section in BraY iv, the a s . a . . content indeed overlaps considerably. The verses in question are 74787 of BraY iv, which concern the a specications for ritual vessels (p tras) and more specically skull-bowls. For the Samvarodaya, I have a . consulted Tokyo University Library manuscript no. 404. 151 Sanderson notes that in the period between penning the article History through Textual Criticism and its publication in 2001, he identied additional passages from Saiva scriptures redacted into the Laghuamvara; the parallel I identify above might belong to this category. History through Textual s . Criticism, 42 (n. 52). Note also that Laghuamvara 26.13cd (aprak syam idam guhyam gopanyam prayatnatah), which occurs s . a . . . . again as 31.14, is parallel to BraY 87.2cd (aprak syam idam devi gopanyam prayatnatah); variants upon a a . . . this phrase appear in chapters xxi, xxii, xlv, and xlvi of the BraY as well. Note the absence of the a (contextually inappropriate) vocative devi in the Laghuamvara version. There are other similarities of s . idiom too: another phrase shared by the BraY and Laghuamvara, and not with other Buddhist sources a s . I am aware of, is n tah parataram kicit trisu lokesu vidyate. This occurs as Laghuamvara 5.25cd, 38.7ab, a . s . . . . 49.14ab, and 50.14ab (cf. also 26.1ab and 49.16), and BraY xiv.262ab and lxxxiv.222ab. Other idiomatic a expressions shared by the BraY and the Laghuamvara include variations upon the following (Laghua s . samvara 3.20cd21ab): . adrs. amandalo yog yogitvam yah samhate ..t .. . . hanyate mus. in k sam pibate mrgatrsnik m | . t a a . . .. . a
149 148

pibate ] Baroda codex; pibati Ed. (unmetrical)

Striking the sky and drinking from a mirage are proverbial expressions of futile endeavor. My attention was rst drawn to this verse by Isaacson in a Sanskrit seminar in the autumn of 2003. Compare e.g. BraY lxxxviii.44: a aviditv -d- imam sarvam yah puj m kartum arhati | a . . . a. hanate mus. in k sam hate mrgatrsnik m . t a a . .. . a

177 Hence, Laghuamvara 2629 corresponds, more or less in sequence, to the last several s . verses of BraY lxxxix and rst fty-odd verses of lxxxv, although individual verses a and several short sections in both have no parallels in the other. Unfortunately, the only codex in which the original Sanskrit of the Laghuamvara appears to survive is s . lacunose from the third verse of chapter twenty-two up to the colophon of twentynine.152 Janardan Shastri Pandey has however made what appears to be a creditable reconstruction of the Sanskrit, utilizing the Sanskrit commentary of Bhavabhatta, the .. Tibetan translation, and parallels in the Samputatantra and Abhidh nottara.153 Table a . . 3.1 places the previously unidentied parallel passage from Brahmay mala lxxxiv a alongside the corresponding verses of Laghuamvara 26, the latter from Pandeys edis . tion (reporting variants he notes are suggested by the Tibetan). In the BraY , this passage concludes the rst chapter of the Uttaratantra, an ada dendum tantra to the BraY possibly belonging to a later stratum of the text.154 This a chapter comprises, primarily, a long and comparatively sophisticated description of yoga for which the passage in question forms the concluding statement. Parallels for the some of the obscure terminology occur earlier in the chapter and elsewhere in the text. In the received Laghuamvara, this passage instead opens chapter twenty-six, s . giving the appearance of having been awkwardly rewritten as an introduction; in verse six, it then shifts to the subject of the Eight Pledges, with a passage parallel to BraY lxxxv.142.155 The product, or so it seems to me, is a tract of decontextuala ized text cobbled together with scant regard for meter and still less for grammar, the interpretation of which challenges the imagination. There are multiple and clear indications of the dependence of Laghuamvara 2629 s .
Verses with remarkable similarities occur as BraY iii.5, xi.44cd45ab, xxii.106, lxxii.212, lxxxii.50, and a lxxxvii.56. These parallels are not however unique to the BraY ; note also Tantrasadbh va 28.88ab and a a Nivasak rik 241cd (hanate mus. in k sam pibate mrgatrsnik m). s a a t a a . a . . .. . 152 Oriental Institute of Baroda manuscript no. 13290. 153 Janardan Shastri Pandey, ed., Srheruk bhidh nam Cakrasamvaratantram with the Commentary of a a . Bhavabhat. a (2 vols.). See the editors note on Laghuamvara 22.3d, vol. 2, p. 121. t s . . 154 On the structure and possible stratication of the BraY , see the next chapter. a 155 Preceding BraY lxxxv.1 is a short series of mantras, the text of which is badly damaged. These a have no precise counterpart in the Laghuamvara. There may however be a loose structural parallel, for s . the short chapter preceding Laghuamvara 26 consists of a single long mantra, introduced by a verse. s .


Table 3.1: A parallel passage in Brahmay mala lxxxiv and Laghuamvara 26 a s . Brahmay mala lxxxiv.22228: a n tah parataram kicit a . . trisu lokesu vidyate | . . j tv picumatam tantram a a . . sarvatantr n parityajet 222 a carv h ravibh ge pi a a a t lak r dhake tath | a aa a sarv tmake ca yogo yam a . sarvatah sv nurupatah 223 a . . dutyog tmayog c ca a a prakriy yogayojan t | a a . sarvatra ca caturnam tu . yogo yam parikrtitah 224 . anulomavilomena dutayah samvyavasthit h | a. . . adhordhvasiddhid devi a atmadut tu sarvad 225 a taddravyam sarvad siddham a . . dar an t spar abhaksanat | s a s . . cumban d guhan c caiva a a sivapthe vi esatah 226 . s . . y vanto dravyasamgh t h a . a a. sarvasiddhikar h param | a. d tavyam mantrasadbh vam a a . . n nyath tu kad cana 227 a a a m t ca bhagin putr aa bh ry vai dutayah smrt h | a a . . a. yasy mantram daden nityam a . . tasyaiso hi vidhih smrtah 228 . . . .
222b trisu ] corr.; trsu Bya .. . 223a carv h ravibh ge ] em.; carv h r vibh go Bya a a a a aa a a a 223b ar dhake ] em.; ar dhane Bya 226a siddham ] em.; siddha Bya . 226c cumban d ] em.; cumban Bya a a guhan c ] em.; guhana Bya a 227a y vanto ] em.; y vato Bya a a a . samgh t h ] em.; sangh tah Bya . a a. 227b siddhikar h ] em.; siddhikarah Bya a. . param ] conj.; parah Bya . 227d cana ] em.; canah Bya .

Laghuamvaratantra 26.15: s . atah param mantrapadam . . . trisu lokesu na vidyate | . . srherukamantram j tv . a a sarv n mantr n parityajet 1 a a

anulomavilomena yoginyah1 samvyavasthit h | a. . . adhordhvam siddhid nityam a . . atmadutm tu sarvag h 2 a. tam dut tu sattv rthasiddhidam a . . dar anam spar anam tath | s s a . . cumbanam guhanam nityam . . . yogaptham vi esatah 3 . . s . . a a. y vanto yogasangh t h a sarvasiddhikaram smrtam | . . sarvasadbh vam deyam ca a . . n nyath tu kad cana 4 a a a m t bhagin putr v aa a bh ry vai dutayah sthit h2 | a a a. . yasya mantram daden nityam . . tasya so hi vidhih smrtah 5 . . .
1 The Tibetan supports reading dutayah . 2 The Tibetan supports reading smrt h . a.

179 upon BraY lxxxivlxxxv, for the redactors appear to have been less than successful a in removing traces of technical terminology distinctive to their source text. Sanderson has discussed one case in detail: a reference to the smarana, a word in ordinary par. lance meaning recollection, but in the BraY , a technical term for the seed-mantra a . of Kap lsabhairava (hum). An ostensibly neutral word, the Buddhist redactors ala lowed this to remain, perhaps unaware of its signicance in the source text.156 In addition to the smarana, I would single out another case in which characteristic jar. gon from the BraY has not been redacted out of the Laghuamvara: 26.14cd15ab, a s . which corresponds to BraY lxxxiv.9. This verse concerns a typology of the s dhaka a a that is as far as I can determine distinctive to the BraY and certainly alien to the a Laghuamvara. The text of the Laghuamvara version of the verse is as follows, in s . s . Pandeys reconstruction: suddh suddh tha miram vai s dhakas trividh 157 sthitih a a s . a a . a ar dhako viuddha ca dpako gunav n narah | s s . a . 14

Jayabhadra, the earliest commentator on the Laghuamvara, recognized that this verse s . should concern a classication of practitioners, and offers the following interpretation: The man of virtue (gunav n narah)the yoginhas a threefold division. . a . a Ar dhaka means one in whom understanding has not arisen; viuddha s means one in whom capacity has arisen; dpaka (light) means the madhyadpaka (middle light): one in whom some understanding has a arisen, and who enlightens himself and others. Or else, ar dhaka means worshipper of the deity through practice of mantra and yoga, gunav n . a means one who understands the meaning of scripture, [while] dpaka means capable of fullling the goals of all living beings, like a lamp (pradpa).158
156 Sanderson, History through Textual Criticism, 4447. The term smarana occurs in Laghuamvara s . . 29.3c in the critical edition. See also, in the present dissertation, the section in chapter 5 on the BraY s a titular epithet nav ksaravidh na. a . a 157 The commentator Bhavabhatta instead reads s dhak s trividh h. a a a. .. 158 Jayabhadra, Cakraamvarapajik : aradhako vi uddha ca dpako gunavan nara iti gunavan s . a s s . . naro yog tridha bhidyate [em. Isaacson; vidyate Ed.] aradhaka ity anutpannapratibhah vi uddha s . ity utpannas marthyah dpaka iti madhyadpakah kimcidutpannapratibhah svapar rthabodhaka ca athava a s . . . . aradhako mantrayog bhy sena dev t radhakah gunavan sastr rthavett dpakah pradpavat sarvasattv rthaa a aa a a . . a . kriy samarthah a .

180 Jayabhadras creative yet incongruent attempts to nd three s dhakas in the second a line testify to the fact that this verse lacks context; a threefold classication of this nature is otherwise absent from the Laghuamvara. s . In contrast, the triad of pure, impure, and mixed comprises a key conceptual framework in the BraY : practitioners, ritual, scripture, and the Three Saktis are pata a terned accordingly.159 Ar dhaka too has a specic, contextually germane meaning. In the BraY , the verse in question occurs as lxxxv.8, in a passage which follows an a enumeration of the initiatory Pledges (samaya): . . . . . . . . . . . . . . . . . . . . . ity as. au samay h par h 7 a. a. .t j tavy h s dhakair nityam s dhan r dhanasthitau | a a. a a aa . s m ny h sarvatantr nam na hantavy s tu hetubhih 8 a a a. a. . a . suddh suddhavimiras tu s dhakas trividhah smrtah | a s a . . . a ar dhako viuddhas tu dpak digunair vin 9 s a a . gr me gr me vratam tasya devat rupalaksanam | a a a . . . unmattam asidh ra ca pavitraksetravarjitah 10 a . . s dhakas tu dvidh prokta carum rgo tha t lakah | a a s a a . t lam rgarat n m tu na carur naiva samyamah 11 a a a a. . . vidy vrataviuddhis tu trisas. ivratam eva ca | a s . .t abhedyatvam tatas tasya t l dau s dhane vidhau 12 aa a . carum rgaikadeo hi t lah sarv tmako bhavet | a s a . a ksetrasth n ni siddh ni yoginyo yatra samgat h 13 a a a a. . . tesu sthitv japam kury c carum alabhate dvijah | a a . . .
7d samay h ] corr.; samay Bya a. a 8a j tavy h ] corr.; j tavy Bya a a. a a s dhakair nityam ] a . 9b em.; s dhakai nnityam Bya (tops damaged) a 9a vimi ras tu ] em.; vimuktas tu Bya s . a trividhah ] corr.; trvidhah By 12b trisasti ] em.; ttrisasthi Bya . . . . .. . ..

. . . these are the supreme eight Pledges. [7d] They should always be known by s dhakas when in the state of [mantra-]s dhana or [deity] wora a ship ( r dhana). They are common to all the tantras, and should not be aa assailed with reasoned arguments. [8] The s dhaka is threefoldpure, ima 160 while the ar dhaka is very pure, free from the quali a pure, and mixed ties () light and so forth (?).161 [9] From village to village, his observance (vrata) is [that of taking on] the form and characteristics of the deities, and
159 On the classication of scripture in relation to the saktis, see chapter 5; see below concerning the s dhaka. a 160 There are strong grounds for emending suddh suddhavimuktas to vimiras, as I have done, for a s this threefold classication of s dhakas based upon degrees of purity pervades the BraY and ts the a a present context. Cf., e.g., xxxii.331c, suddh suddhavimiresu. Furthermore, the Laghuamvara offers some a s . s . conrmation for the emendation, reading suddh suddh tha miram, as reconstructed by Pandey. a a s . While the sense of 9b is certain, one could consider emending to s dhakah trividh smrtah, or to a a . . . s dhak s trividh sthit h; the latter is supported by Bhavabhattas reading of the Laghuamvara parallel. a a a a. s . .. 161 I am unfortunately unable to determine the probable intended sense of 9d, dpak digunair vin , a a .

181 the madman and razors edge [observances],162 avoiding the locations of sacred elds. [10] But the s dhaka is [actually] twofold: the one fola lowing the path of caru (food offerings), and the t laka. For those on a the t laka path, there is neither caru nor self-restraint. [11] [After engaging a in] purication by the vidy -mantra observance and the sixty-three obsera vance,163 he then [reaches] the state of [making] no distinction between the ritual procedures of the t laka, etc. [12] Following the way of the caru, a having a single location, the t laka would become a sarv tman (univera a 164 Remaining in the sacred, empowered places where the sal) [s dhaka]. a yogins assemble, he should perform his mantra recitation in those; the twice-born one obtains caru [from the yogins].165 [1314ab] a Here ar dhaka, worshipper, refers to a specic category of practitioner. Typically, the BraY describes a threefold typology of the s dhaka: pure, impure, and impurea a
as transmitted in Bya . The parallel text in Laghuamvara 26.15b provides no assistance obvious to me. s . Jayabhadras interpretation of the latter, quoted above, might invoke the rhetorical gure of the dpaka. 162 The unmattakavrata is fourth of the Nine Observances described in BraY xxi, involving feigned a insanity, as the name implies. The asidh ravrata (observance of the swords edge) for its part comprises a the subject of BraY xxxix. a 163 While the various observances taught in BraY xxi are referred to collectively as vidy vratas, oba a servances of the [nine-syllable] vidy , this term is primarily used for the nal and most important of a these, a k p lika observance also called the mah vrata (108a) or bhairavavrata (109ab). As for the trisas. ia a a . .t vrata, this appears to be connected with a mantra-deity pantheon (y ga) of the same name; yet while the a y ga of the sixty-three and its vrata are mentioned in several chapters, I have not identied a detailed a description. 164 The implication is that the sarv tman s dhaka is bound by no single discipline and may engage at a a will in practices associated with the lower grades of initiate. This is consistent with the description of the sarv tman found in BraY xciv. a a 165 The notion that one may attain siddhi through consuming food offerings (caru) given directly by the yogins is mentioned in BraY ci 29, and is in all likelihood alluded to here in 14b. For a more a explicit description, note Kaulaj nanirnaya 11.7cd10: a . yad icchet kaulav siddhih pr sya pamc mrtam param 7 . a . a . . tad sidhyati yoginy m siddhimel pakam bhavet | a a. a . dadante ca tad devi carukam pacabhir yutam 8 a . yoginbhih sakrd dattam tatksanat tatsamo bhavet | . . . . . atha v pr sayej j tv yogayuktas tu kaulavit 9 a a a a sidhyate n tra sandeho vighnaj lavivarjitah | a a . yoginganas m nyo manas cintitam labhet 10 . a a a .
7d pr sya ] em.; pr sya Kjncod ; pr pya Kjned a a. a 9a yoginbhih ] Kjned ; y[e?]ginbhih Kjncod 10 . . sandeho ] Kjned ; sandoho Kjncod 10b vivarjitah ] em.; vivarjitamh Kjncod ; vivarjitam Kjned 10c . . . s m nyo ] Kjncod ; s m ny Kjned a a a a a 10d manas cintitam labhet ] Kjncod ; manahsu cintitam bhavet a . . . ed Kjn

One who desires Kaula siddhi then obtains siddhi after consuming the ultimate ve nectars. There would transpire a siddhi-[bestowing] encounter with the yogin[s]. And they then give [him] the food offering (caru), mixed with the ve [nectars], O goddess. [If he consumes the caru] immediately when rst given by the yogins, he becomes equal to them. Otherwise, if he would consume it after thinking [rst], the knower of the Kaula, disciplined in yoga, will undoubtedly attain siddhi, free from the web of obstacles. Equal to the horde of yogins, he would obtain whatever he thinks about. One should perhaps emend to sidhyanti yoginyah in 8a. .

182 cum-pure, for which the designations are t laka,166 carubhojin (one who consumes a ritual offerings), and mira (mixed), respectively.167 This classication receives s detailed elaboration in the texts massive forty-fourth chapter, the section on the s dhaka (s dhak dhik ra). However, the latter chapters of the BraY lxxxivci, coma a a a a prising the Uttara- and Uttarottaratantras168 introduce a new fourfold taxonomy of a initiates: the ar dhaka, carubhojin, t laka, and sarv tman (universal), whose activities a a and subdivisions comprise the respective subjects of BraY xcixciv. This typola ogy differs from the threefold insofar as the category of miraka, the practitioner of s mixed purity, appears to be recongured as the highest grade, the sarv tman a a above the t laka.169 On the other hand, the ar dhaka represents a variety of housea holder practitioner.170 That the redactors of the Laghuamvara had intended to remove references to a s . Saiva typology of practitioners is suggested by comparison; in table 3.1, note that BraY lxxxiv.22324, which makes specic reference to the classication of s dhakas a a
166 The word t laka appears non Indo-Aryan, and Sanderson (personal communication, May 2004) a suggests a connection with the Tamil t. , energy, effort, perseverance, application. The University of al Madras Tamil Lexicon, 1885. Accessed online (April 2007) through the Digital Dictionaries of South Asia project ( 167 The terms for the threefold s dhaka are provided in BraY xliv.10cd11ab: a a

suddhas tu t lakah prokta carubhoj tv auddhakah 10 a s s . . suddh suddho bhaven mirah s dhakas tu na samsayah | a s . a . .
prokta ] em.; proktams Bya s .
168 169

mi rah ] em.; misram Bya s . .

The structure and possible stratication of the text are discussed in the subsequent chapter. It is evident from the descriptions in BraY xliv that the miraka, as one might expect, constitutes a s the middle grade of s dhaka. Hence in xliv.472, it is said that a miraka puried through constant a s practice may become a t laka (kad cin mirako devi karmayogena nityaah | t lam rga[m] sam pnoti yad a a s s . a a a a . suddhah praj yate). However, the sarv tman s dhaka is mixed in an entirely different sense: he is free a a a . from all regulations, engaging at will in the disciplines associated with lower practitioners, including consumption of the impurest of substancesthe domain of the carubhojin. 170 a It appears that the ar dhaka might not be considered a s dhaka, per se; their characteristic modes of a a ritual, ar dhana (worship) and s dhana, are placed in contrast. See e.g. lxxxv.8b above. Nonetheless, a a a the term ar dhaka gures in later Saiva typologies of the s dhaka. In the Kulas ra, the ar dhaka features a a as fourth of the ve grades of s dhaka, above the t laka, cumbaka, and c rv ka (=carubhojin, presumably); a a a a a transcending the ar dhaka is the sivodbhuta: t lako cumbaka caiva c rv k r dhakas tath | a s a a aa a . sivodbhuta -m- atah proktah p mcabhedo pi s dhakah | a. a . . .
c rv k r dhakas ] em.; c rv kor dhakas ms a a aa a a a sivodbhutam ] em. Vasudeva; sivobhutam ms

I am grateful to Vasudeva for providing me with his transcription of this material. Given the terminological continuities, it seems possible that this vefold typology develops out of the threefold a classication present in the BraY , the addition of the ar dhaka reecting an intermediate stage. a

183 in question, is absent from the Laghuamvara. Yet Laghuamvara 14cd15ab refers s . s . nonetheless to what is, in the BraY , the same typology expressed with different a terminology.171 In the case of the Laghuamvara, I believe there can thus be little doubt concerns . ing Sandersons proposal: that this text has incorporated material from the BraY , a whether directly or through another derivative source, seems the only plausible explanation for the relationship between the passages in question. Derrivation from an unknown common source is perhaps not impossible, but this would in all likelihood have been a text intimately related to the BraY , to the extent of sharing unusual a terminological similarities. The case is similar with chapter forty-three of the Abhidh nottara, another text of the Cakra amvara tradition;172 as Sanderson points out, a s . this corresponds to the same material shared by the BraY and Laghuamvara. This a s . begins with text corresponding to Laghuamvara 26.6 and BraY lxxxv.9, omitting the s . a ve verses parallel to 26.15 and lxxxiv.22228 of these respective works. Though the text of Abhidh nottara 43 closely parallels Laghuamvara 2629fortuitously so, given a s . that this section of the Laghuamvara does not survive in Sanskritthe former cons . tains none of the latters divisions in chapters, being hence closer to the BraY . This a in fact appears true of the early Laghuamvara as well, for the commentator Jayabs . hadra shows no awareness of the chapter divisions known to the later commentator Bhavabhatta.173 Sanderson contends that several other sections of the Abhidh nottara a .. derive from Saiva sources as well, for which Judit Trzsk has provided convincing evidence in the case of its relationship to the Siddhayogevarmata.174 s
A reference to the fourfold typology of practitioners is clearly present in BraY lxxxiv.223, ala a though out of sequence: carv h ra (=auddha or carubhojin), t laka (=uddha), ar dhaka (by emendation of a a s a s a ar dhane; =viuddha), and sarv tmaka (=mira). While the interpretation of 224ab remains unclear to me, s a s the point of 224cd is that the yoga expounded in this chapter is applicable to all four (caturnam) types . of practitioner. 172 I have consulted two manuscripts of the Abhidh nottara: Institute for Advanced Studies of World a Religions, Buddhist Sanskrit Manuscripts lm-strip no. mbb-1971-100 (Nepalese script, dated to the equivalent of 1138 c.e.); and a late Devan gar manuscript that has been published in facsimile: Lokesh a Chandra, ed., Abhidh naottara-tantra: a Sanskrit Manuscript from Nepal. a 173 This has been pointed out to me by Isaacson, personal communication (April, 2007). 174 Trzsk discusses parallels between Siddhayogevarmata 29 and Abhidh nottara 38, identifying ins a dications of the direction of redaction from the former to the latter. Her examples include changes of

184 Ronald Davidson has voiced skepticism concerning a number of Sandersons claims pertaining to the relationship between Tantric Saivism and Tantric Buddhism. Of particular relevance to the present discussion, he questions the plausibility of extant Saiva texts being signicant sources of material found in the Buddhist yogin tantras. One of his principal objections is chronological: he considers problematic the evidence attesting specic, extant works of tantric Saiva literature prior to the ninth and tenth centuries. He questions, for instance, whether the mid eleventh-century Cambodian Sdok Kak Thom inscription should be taken as an accurate record for . the existence in the ninth century of the Saiva texts it mentionsfour v matantras a which the inscription associates with a brahmin in the court of that period. While such caution is laudable in principle, here it appears excessive: the existence of Saiva v matantras prior to the ninth century is suggested in multiple manners, and the texts a mentioned in the inscription are known to have been fundamental scriptures of this genre.175 In fact, Davidsons objection appears somewhat imbalanced considering that he draws upon a single reference in the late medieval K lik pur na for recona a a. structing the alleged pre-Buddhist origins of the deity Heruka, relying heavily on a mythological text for reconstructing history at a remove of well more than half a millenium.176
non-Buddhist references to Buddhist ones, alterations which render a metrical verse in the Saiva text unmetrical in the Buddhist, and Saiva iconographic features left unchanged in the Buddhist version. See Doctrine of Magic Female Spirits, appendix 3 and its supplement, Parallels of Siddhayogevars mata 29 in the Abhidh nottara Patala 38, 19295 (quotations on 19495). a . 175 Davidson addresses Sandersons remarks on this inscription as they were presented in History through Textual Criticism, 78 (n. 5). Sanderson has discussed this material in greater detail more recently, in Saiva Religion among the Khmers (Part I), 35557. On other early evidence for the v maa tantras, see the discussion of these in section 2. 176 Davidsons attempts to show that Bhairava and his Buddhist counterpart, Heruka, have (independent) roots in tribal or local divinities seem unconvincing. The K lik pur na, a text that in its current a a a. form is unlikely to predate the sixteenth century, associates a cremation ground called Heruka with K m khy , and Davidson identies this (plausibly) as the modern site called Mas nbhairo (maanaa a a a s s bhairava). He postulates that Buddhists apparently appropriated a local term for a specic Assamese ghost or cemetery divinity and recongured it into the mythic enemy of evil beings in general Heruka. Indian Esoteric Buddhism, 21116 (quotations on 211, 214). On the dating of the K lik pur na, a a a. see Sylvia Stapelfeldt, K m khy SatMah m y : Konzeptionen der Groen Gttin im K lik pur na, 3540. a a a a a a a a a. Assuming that the K lik pur nas Heruka cremation ground is indeed the site Mas nbhairo, this in fact a a a. a tells us no more than that Heruka and Bhairava were at some point prior to the sixteenth century considered cognate, to the point that their names could be interchanged. To argue that the Buddhist Heruka was originally an Assamese cremation-spirit deity on this basis calls to mind what David-

185 Critiquing Sandersons thesis of the Buddhist yogintantras indebtedness to Saiv ism, Davidson counters that a more fruitful model would appear to be that both heavily inuenced the nal formations of the agonistic other and that each had alternative sources as well.177 A model of mutual inuence certainly has appeal when considering Buddhist-Saiva interactions broadly over the course of the rst millienium,178 yet such cannot be assumed a priori in any particular case; indeed, most of the texts Davidson cites as examples of Tantric Saivisms syncretic sources appear to be post twelfth-century works, and accordingly have little bearing upon the relation between the Saiva Vidy ptha and Buddhist yogintantras. An exception is a . the Jayadrathay mala, which as Davidson points out mentions Buddhist tantras in its a account of the scriptural canon, apparently naming the Guhyasam ja.179 The Jayada rathay mala, Sanderson suggests, is an historically layered composition that took its a nal form in Kashmir prior to the period of Ksemar ja (. circa 100050).180 That a .
son elsewhere describes as sustained special pleading about single reference citations, a questionable method of arguing history. Indian Esoteric Buddhism, 206. As for Bhairava, Davidson asserts that he seems to have been little more than a local ferocious divinity at one time. . . eventually appropriated by Saivas, much as they aggressively appropriated so much other tribal and outcaste lore for their own ends. The only evidence he cites for this assertion are origin myths in the K lik pur na for a linga called Bhairava near Guwahati, and tantric Buddhist a a a. lore associating a series of twenty-four p. has with the same number of bhairavas as guardians. Indian t Esoteric Buddhism, 21113 (quotation on 211). While the roots of Bhairava remain obscure, it is worth noting that Mah bhairava is named as a Saiva place of pilgrimage in the Niv sa (Mukh gama 3.21d a s a a and Guhyasutra 7.115d) as well as the old Skandapur na (chapter 167); the latter source makes clear that a. the site is named after the form of Siva enshrined there (cf. Mah k la of Ujjayin). A fourth-century a a V k taka king is described as a devotee of Mah bhairava in an inscription of the fth century, on a a. a which see Sanderson, Saivism among the Khmers, 44344; and Peter Bisschop, Early Saivism and the Skandapur na, 19293. The emergence of Bhairava in the tantric Saiva pantheon, whatever his roots may a. be, appears to have involved some degree of identication with Aghora, the southern, erce face of Sad siva who is said to reveal the bhairavatantras. a 177 Davidson, Indian Esoteric Buddhism, 217. 178 Note for instance Davidsons plausible suggestion that P supata monasticism is a response to the a sramana ascetic orders. Ibid., 18386. One might also mention the possible inuence of the Buddhist . Yog c ra school upon nondualist Kashmiri Saiva thought, although this requires further investigaa a tion. 179 Indian Esoteric Buddhism, 217, citing Mark Dyczkowski, The Canon of Saivism, 102. 180 Sanderson, The Visualization of the Deities of the Trika, 32 (n. 6); and Remarks on the Text of the Kubjik matatantra, 2. The Jayadrathay mala is a heterogenous, layered work; Sanderson sees within it a a s multiple texts that might originally have been independent: the Siracheda, an early v matantra; and the a M dhavakula, a text cited by Abhinavagupta and incorporated into the fourth book (satka) of the Jayada . . s rathay mala. On the Siracheda, see History through Textual Criticism, 3132 (n. 33), and Remarks a on the Text of the Kubjik matatantra, 12. On the M dhavakula, see the latter. His claim, to which a a Davidson responds and for which the evidence has not yet been published, is that the ninth chapter of the Jayadrathay malas third book, the Yoginsac raprakarana, was a source for Buddhist yogintantra a a .

186 sections of the text show awareness of Tantric Buddhism is hence neither surprising nor unusual, and Davidsons assertion that this suggests dependence on Buddhist tantras should require demonstration of the nature of such dependence.181 Among the other Saiva texts Davidson singles out is the Brahmay mala; but what he refers a to is in fact a late medieval east Indian composition by this title, rather than the early Vidy ptha scripture.182 It would indeed appear that the later sakta tradition a . of Saivism, particularly in east India, appropriated much from Tantric Buddhism during the centuries of the latters decline. This is dramatized, for instance, in tales of the brahmanical sage Va isthas sojourn to Mah cna (Greater China) in order s .. a to learn worship of T r from the inebriated Buddha, and evidenced by the emeraa gence of syncretic pantheons such as the Ten Great Vidy -mantra Goddesses (daa a s mah vidy h), who include T r . The old BraY , in contrast, mentions in its account of a a. aa a the scriptural canon works of agonistic others such as the Vaisnavas and Vaidikas, .. yet shows no awareness I can discern of Tantric Buddhism. The same may be said of the Niv sa, an even earlier composition which otherwise shows enormous interest s a in hierarchically ordering rival systems. Regrettably, Davidson goes so far as to suggest that Sandersons model of the Vidy ptha is informed by a curious theology of scripture, contending that while a . it is seldom that a received body of texts reects no inuence at all, this seems to be Sandersons ultimate position on the Vidy ptha Saiva scriptures.183 This assertion a . appears entirely unsustainable in light of Sandersons research into the layered genematerial. History through Textual Criticism, 4143. 181 Davidson, Indian Esoteric Buddhism, 217. 182 Davidson refers to the Rudray mala, T r tantra, and Brahmay mala as texts transmitting the legend a aa a of Va istha learning cn c ra (the Chinese method) from the Buddha. Indian Esoteric Buddhism, 216, s .. a a citing Benoytosh Bhattacharyas introduction to S dhanam l , vol. 2, cxiii (in fact cxliii); and Bhata aa tacharya, Buddhist Deities in Hindu Garb, in Proceedings and Transactions of the Fifth Indian Oriental Conference, vol. 2, 127798. In this matter Bhattacharya drew upon Sanskrit textual materials published in a particular volume, edited from Bengali manuscripts: Girsacandra Ved ntatrtha (ed.), T r tantram. a aa Srgirscandraved ntatrthasankalitam. With an Introduction by A. K. Maitra. As I discuss in chapter 1, a section 1, this publication includes excerpts from the rst two chapters of a certain Brahmay mala prea served in a manuscript of the Varendra Research Society. There is no indication that the text is related to the Vidy ptha scripture of the same name. a . 183 Davidson, Indian Esoteric Buddhism, 386 (n. 105).

187 ologies of Saiva scriptures, including those of the Vidy ptha. Concerning the Tantraa . sadbh va, a Trika text of the Vidy ptha, he demonstrates that it has incorporated and a a . expanded upon cosmological material from the Svacchandatantraan extensive tract of text which the latter, in turn, drew in part from the Niv saguhya, transforming this s a in the process within its own cultic system.184 He argues, moreover, that the Niv sa s a itselfperhaps the earliest extant tantric Saiva scriptureis heavily indebted to pre and proto-tantric Saiva sects of the Atim rga.185 Particularly noteworthy is Sandera sons more recent investigation into the formation of the Netratantra, a Saiva text he argues was produced in the milieu of an eighth- or early ninth-century Kashmiri court.186 Note also his demonstration that the Brhatk lottara, a Kashmiri-provenance a . tantra of the Saivasiddh nta, has incorporated material from a Vaisnava scripture of a .. the P car tra.187 In light of this obvious commitment to identifying agents, circuma a stances, and sources involved in the formation of Saiva scriptural literature, it hardly seems defensible to attribute bias to Sanderson for failing to unearth examples of the indebtedness of early texts of the Vidy ptha to tantric Buddhist sources. I am aware a . of none; yet given the current state of research, it is entirely possible that examples will surface.



The present chapter has attempted to trace the early development of the yogin cult in Saiva and Buddhist tantric literatures. It was shown that signicant elements of the Saiva cult of yogins have roots in earlier Saiva scriptural genres. In particular, aspects of the k p lika cult of Bhairavain association with which the Saiva cult of yoa a gins comes into evidencehave discernable precedents in the Niv satattvasamhit . s a . a
Sanderson, History through Textual Criticism, 2332. The windows afforded by the Niv sa into early Saiva systems and its own dependence upon these s a comprise the subject of Sandersons recent essay, The L kulas: New Evidence of a System Intermediate a between P c rthika P supatism and Agamic Saivism. See also Sanderson, History through Textual a a a Criticism, 29. 186 Sanderson, Religion and the State: Saiva Ofciants in the Territory of the Kings Brahmanical Chaplain, passim. 187 Sanderson, History through Textual Criticism, 3841.
185 184

188 This archaic scripture of the cult of Sad siva appears to predate clear differentiation a between siddh ntatantras and other tantric Saiva scriptural traditions. More direct a precedents for the cult of yogins appear to lie, however, in scriptural traditions for which the record is fragmentary: dakintantras, bhutatantras, and the bhagintantras of . the cult of the Sisters of Tumburu. The latter, attested by the seventh century, were shown to gure in the background of the BraY , one of four extant, early bhairavaa tantras of the Vidy ptha (Seat of Female Mantra-deities). These scriptural authora . ities for the Saiva cult of yogins distinguish themselves from bhairavatantras of the Mantraptha (Seat of Male Mantra-deities) through their emphasis upon goddesses . (vidy ). It was argued that the Vidy ptha yogin cult might have developed within a a . a cultic context of the variety represented by the Svacchandatantraa text sometimes assigned to the Mantraptha in which goddesses have a secondary cultic status, and . in which the cult of yogins registers a presence in only the nal chapter, probably a late addition to the text. While Vidy ptha literature represents the yogin cult in a . its earliest accessible form, the vast majority of Saiva sources connected with yogins belong, however, to subsequent Kaula traditions. As discussed in chapter two, it is Kaula conceptions of yogins that appear to inform the post ninth-century temple tradition connected with these goddesses. A Buddhist cult of yogins is attested by the late eighth or early ninth century with the appearance of yogintantras, Tantras of the Yogins. While heralding a remarkable shift in the character of Tantric Buddhism, the prominence of goddesses in the yogintantras is the culmination of developments observable in earlier litera ture. The Brahmanical Mother goddesses feature in Buddhist mandalas as early as .. the mid-seventh century, in the Mah vairocanasutra, while the range and prominence a of goddesses and female spirits given Buddhist identities grows considerably in subsequent literature, such as the Majurmulakalpa. It was shown that the emergence s of the Buddhist cult of yogins is linked closely to transformation of the dakinfrom . the pernicious female spirit represented in early Buddhist and non-Buddhist sources

189 to the powerful, potentially benign vajradakins of the yogintantras. Evidence for the . conversion and elevation of the dakin appears in eighth-century sourcesthe Tattva. samgraha and Guhyasam jaa process completed in the early yogintantras, such as a . the Laghuamvara. s . In general character, the Saiva and Buddhist cults of yogins possess remarkable similarities, and the relations between these religious systems in their historical contexts merits comprehensive inquiry. In the nal section of this chapter, I offered evidence in support of Alexis Sandersons proposal that the Buddhist yogintantras in several cases depended heavily upon Vidy ptha scriptures. It was shown that a . the BraY is indeed the source of a substantial tract of text redacted into the Laghua samvara and Abhidh nottara, texts of the Cakra amvara scriptural cycle. This is sig . a s . nicant, moreover, not only for the history of tantric Buddhist literature, but also for locating the BraY in time, as discussed in the subsequent chapter. While this case a of textual borrowing undoubtedly provides a limited window into complex historical processes, it nonetheless illustrates that despite Ronald Davidsons objections, Sandersons thesis concerning the dependence of the Buddhist yogintantras on Saiva models remains compelling.

Chapter 4
The Content, Structure, and Provenance of the Brahmayamala


introduction: select topics in the brahmayamala

This chapter and the next shift focus more directly to the BraY , beginning in the a present chapter with discussion of its content, structure, and provenance. The rst section provides an overview of the BraY s material on several major topics, includa ing mantra, initiation and consecration, and religious images. This is complemented by appendix A, a transcription of the texts chapter colophons as found in Bya , which provides some indication of the range of the texts subject matters. Section two of the present chapter analyses the BraY s structure. It is shown that the text has multiple a divisions, the nature of and disparities between which suggest the existence of at least two textual strata. In section three, I address the question of the BraY s dating, a both in absolute terms and in chronological relation to some of the extant literature. This section also explores the geographic and social horizons of the text on the basis of places and individuals named therein. The BraY consists of one-hundred and one chapters of considerable variety in a length, containing a total of more than twelve-thousand verses.1 It takes the form of a conversation between Kap lsabhairava and the Great Goddess (mah dev) or a a Supreme (par ) Sakti, designated in this text by the names Bhairav, Aghor, and a Canda K p lin. The Goddess poses questions, and Bhairava answers, his didactic . . a a
1 At one extreme, chapter seven contains only eleven verses, while BraY iv extends almost ninea hundred and fty in length. On the length of the BraY , see chapter 5, section 5. a


191 responses constituting most of the text. Ordinarily, a chapter begins with a new question, although there are numerous exceptions; within a chapter, further questions might inaugurate new subjects. The primary departures from the dialog mode are occasional third-person asides, such as the common thus did speak Bhairava (evam vai bhairavo bravt). Aside from several mantras given in prose, and a smatter . ing of verse in the sragdhar meter,2 the text consists entirely of the thirty-two syllable a anus. ubh verses typical of tantras, pur nas, and much other religious literature. a. .t The BraY commences with the narrative of its revelation. Following this, chapa ters two and three introduce key topics: the Nine-Syllable Vidy -mantra (nav ksar a a . a vidy ) of the Goddess and the pantheon of mantra-deities this embodies, in BraY a a ii; and in chapter three, the fully elaborated form of the deity mandala, associated .. particularly with initiation. While the text is not systematic in its organization, there are several clusters of chapters concerned with particular topics, especially the cycle on initiation and consecration (BraY xxxixxxvii). Other important chapter clusters a include the those on mantra (reviewed below) and the s dhaka,3 while a number of a chapters, clustered especially between fty-three and eighty-two, are devoted to the propitiation of particular mantra-deities, primarily forms of Bhairava (table 4.1).4 Chapters devoted to yogins are scattered throughout the second half of the text, sev eral of which are critically edited in part ii.5 A compendium of rituals, by and large, the cohesion of this sprawling text lies primarily in the leitmotiv of the nav ksar vidy a . a a
The verses in the sragdhar meter (the four quarters of which contain twenty-one syllables, with a caesuras after the seventh and fourteenth) are the opening benediction, and two verses and a stray quarter in the closing section of BraY lxv. a 3 The principal exposition on the nature and practices of the s dhaka is BraY xliv, a chapter of a a just over seven-hundred verses. Chapters xcixciv expound a fourfold typology of s dhakas, which as a mentioned in the previous chapter (section 5), differs from the threefold classication advanced in xliv. 4 A number of the treatises are specically called kalpas or mantrakalpas, which [set] out the procedure for the propitiation of a Mantra. Sanderson, History through Textual Criticism, 13. Many of these mantra-deities are connected with s dhanas of a radical nature. Besides the male deities delina eated in table 4.1, the BraY has chapters concerned with the goddess Aghore var (BraY lvii), the Six a s a Yogins (lxxviii), and s dhana on an individual yogin (prthakyogins dhana) (lxxix). a a . 5 BraY lv concerns the classication of the clans (kula) of yogins, as well as secret signs; these topic a are also treated in lxxiii. BraY xcviii too concerns the divisions of yogin clans, while xcix provides a a brief account of yoginmel patransactional encounters with the goddesses. Several chapters delineate a ritual concerned with particular yogin sets: BraY lvi with a conguration of twenty-four yogins, and a lxxviii with the Six Yogins of the primary mandala. ..

192 and its mantra-deities, which suffuse and pattern the basic rites and their countless inections. In Table 4.1: Forms of Bhairava in the Brahmay mala a deity Kap lsabhairava a Vijayabhairava Manth nabhairava a Rurubhairava Mah k la a a a Kank labhairava Phetk rabhairava a Picubhairava Gartt bhairava a Y malabhairava a Hairambhabhairava Mah mardakabhairava a Utphullakabhairava a Jhank ra-/Ekap dabhairava a chapter passim xxviii xlv liii liv lxii lxvii lxviii lxix lxxi lxxvi lxxvii lxxxii xc both bulk and emphasis,

the BraY is a prescriptive text a overwhelmingly concerned with ritual. Within this amorphous category must be made numerous distinctions: dksa (initiation), abhiseka (conse . . cration), y ga (deity worship), vrata a (observance), yoga, and s dhana a (special practices of the s dhaka), to a name some of the major categories. Actual practices of these ritual types often overlap, for ubiquitous are the

ny sa (installation) of the mantra-deities upon a substrate (e.g. the body), japa or a mantra-incantation, dhy na or meditative visualization, mudr or ritual signs, and to a a a lesser degree re-sacrice (homa). A number of chapters are devoted specically to these techniques.6
6 On the subject of ny sa, note in particular BraY xli, the ny sapatala (Chapter on Mantraa a a . installation), which provides general instructions, and BraY xii, which elaborates upon the ny sa a a of the extended mantra-deity pantheon taught in the preceding chapters (x and xi). General procedures for mantra-recitation, japavidh na, are expounded in detail in a chapter bearing this title, BraY a a xviii. This provides, for instance, a threefold classication of japa and technical terms for mantras when inected in particular ways. Mudr forms the exclusive subject of chapter xlii; a transcription a of this from the oldest manuscript has been made available by Somadeva Vasudeva (see chapter 1, n. 88). Several mudr s and their mantras are elaborated upon in the last section of BraY xxiii, such as a a the Ocean of Milk (ksroda) and Defeat of Death (mrtyujaya). Chapter livcalled the Mah k lamata a a . . and the Mudr p. h dhik ra (The Seat of Mudr s Section)describes mudr s associated with the Eight a t a a a a Mothers. BraY lv also outlines a number of mudr s. Several chapters and sections focus specically a a upon dhy na, meditation on the visual forms of the deities; note e.g. BraY viii, which is concerned a a with yogic visualization on the goddesses, with magical aims. Similarly, BraY vii (only eleven verses) a describes iconographic inections of the goddesses as mounted upon yantras (yantr rudh h). Chapters a . a. focusing in more detail upon aspects of yoga include BraY ix, which concerns laks[y]abheda (the types a . of desiderata). This entails meditation upon goddessesperhaps the Four Goddesses of the core mandala, although this is not speciedas connected in turn with an hierarchy of cosmological levels .. (tattva) and cosmic creative powers (kal ). (On the subject of laksyabheda, see Somadeva Vasudeva, The a . Yoga of the M linvijayottaratantra, 25392.) Fire ritual, homa or agnik rya, forms the subject of chapter a a

mantra-deity pantheons and their worship

A number of chapters are specically concerned with mantroddh ra, the extraca tion/derivation of mantras. The core mantra system of the BraY , based on the a a nine-syllable vidy -mantra of Aghore var, is explicated beginning in BraY ii, which a s outlines the vidy -mantra itself (on which see chapter 5, section 5). BraY x delineates a a the extended mantras of the principal mandala goddesses, while xi concerns those of .. Bhairava, Bhairav, the Three Saktis, etc. BraY xxiii elaborates upon numerous in a ections of the vidy , and in its nal section teaches the Defeat of Death (mrtyujaya) a . mudr and mantra. a Parallel mantra-systems are taught in several other chapters. Two seem particularly important: the khecarcakra of BraY xiv, from which are extracted the kulavidy a a and samayavidy (the Clan Vidy -mantra and Pledge Vidy -mantra, respectively), a a a as well as the p supat stra (the Weapon Mantra of Pa upati); and the kurmaprast ra a a s a (the Tortoise Chart) of BraY lxxxviii. Several chapters in the latter portion of the a BraY teach deities and ritual based upon mantras derived from this kurmaprast ra. a a BraY lxi, the Tilakatantra, contains a new mantroddh ra, although based upon the root a a pantheon.7 A number of other chapters throughout the text teach additional mantras, many of which are connected to specic forms of Bhairava. Mudr s too have associaa tion with particular mantras; thus for instance BraY xxvi teaches mudr mantroddh ra, a a a extraction of the mantras of the mudr s, while liv, called the Mudr p. h dhik ra or a a t a a Mah k lamata, delineates the mantras of mudr s associated with the Eight Mothers. a a a Several sections of the text teach the basic y gas or mantra-deity pantheons of a the BraY and their associated rites (y gavidhi/vidh na). The Nine Y gas, pantheons a a a a centered upon each deity, in turn, of the core mandalaBhairava(-Bhairav), the Four .. Devs, and Four Dutsare listed by name in the opening of BraY iii. These have a
xx. Note also BraY li, a brief chapter explaining the meats which should be offered in re sacrice for a particular purposes. 7 As discussed later, this chapter appears remarkably self-contained, containing its own descriptions of, for instance, initiation, mantroddh ra, construction of the mandala, purvasev or prelimary propitiaa a .. tion of the mantra-deity, the Nine Y gas, puracarana, and the topic from which it appears to draw its a s . name: substances used for applying a magical bindi upon the forehead.

194 their congurations specied in BraY xiii, the chapter on the Nine Pantheons a (navay gapatalah). Grhay ga, or the worship performed in a shrine utilizing the Nine a a . . . Pantheons, appears to be the normative form of daily worship (nityakarman). This subject is taught in BraY xxix (from verse 193), which explains installation of the a pantheons in the context of shrine worship.8 BraY xxixs primary subject is however a a the mulay gavidhi, ritual procedure of the root pantheon, for which it teaches an elaborate mandala distinct from that of the mah y ga of BraY iii. a a a .. Many occasional rites (naimittika) and those with special aims (k mya) utilize the a Great Y ga (mah y ga), the extended pantheon as installed in an elaborate mana a a . dala. The subject of BraY iii, the mah y ga is explicitly contrasted with with the a a a . ninefold y ga used in a shrine context. According to a denition in BraY xv, the a a mah y ga is distinguished by its construction in a cremation ground with a mandala a a .. containing eight directional cremation grounds, as delineated in BraY iii.9 BraY a a xxv, called the y ganirnayapatala (the chapter of denitive judgment on y ga), offers a a . . a detailed explication of the mantra-installation (ny sa) and ritual procedures cona nected with several of the systems important y gas. This chapter claims that there a are in total sixty-ve y gas to which all s dhakas are entitled; for another fourteen, a a only s dhakas of the upper two of three grades, the t laka and mixed, have entitlea a ment. The y ga of the Three Saktis, V m , Jyesth , and Raudr, comprises the subject a a a .. a of BraY lxxxiii, while chapter twenty-seven describes the y ga of a more unusual a a triad: the V m (left), Madhyam (middle), and Daksina (right) Saktis, who a a a . . in BraY xxxviii are said to preside over the three primary streams of scriptural reva
A passage in BraY xvii teaches the procedures for nityakarman, which in this (possibly incomplete) a account consist of installation and worship of the mandala of the vidy -mantra, i.e. the vidy cakra, and a a .. binding of the Skull and Skull-staff mudr s, or else the Pledge-mudr . Just prior to this are taught the a a naimittika and k mya worship based upon the vidy y ga. These descriptions appear incomplete, omitting a a a for example mantra-installation on the body of the practitioner; some of the expected material appears instead in chapter xxv. 9 BraY xv.11: a smas ne tu mah ghore yo y go kryate prye | a a a . . smas nair as. abhir yukto mah y ga sa ucyate 11 a a a .t That y ga which is performed in a terrible cremation ground, my dear, possessing the a Eight Cremation grounds, is called the Great Y ga. a

195 elation and their practitioners. Chapter thirty, the chapter on the different Sivas and rudras, teaches a y ga called navan bhamandala, the mandala of nine naves. a a .. .. Exceptionally, this is a conguration entirely of male deities. Elaborate alphabetical cakras and their associated rites form the subject of chapters xiv, xvii, and xix. BraY xiv, mentioned above, describes the khecarcakra or kua lacakra, which forms the locus of powerful rituals performed in a cremation ground, crossroads, a shrine of the Mother goddesses, etc., or is ritually etched upon cloth, metal, or another substrate. BraY xvii teaches an elaborate cakra based upon the a nine-syllable vidy . This vast chapter also describes a variety of connected magical a rites (karma), involving etching the cakra upon a moveable surface or the ground. The mantra-deity conguration called the bhautikacakra comprises the rst subject of BraY xix, a chapter that delineates a cakra based upon the nine-syllable vidy as well. a a
initiation and consecration

Much material pertaining to the important topics of initiation and consecration is clustered in chapters xxxixxxvii. Together, these comprise more than ten percent of the BraY . Chapter thirty-two purportedly covers dksa, initiation proper, but a . in fact focuses almost exclusively upon one important dimension: the initiations of the sixfold courses or ways (sadadhvan) of ascent to Siva: bhuvanadksa (28? . . . 158ab), padadksa (158cd245), bja- or varnadksa (24667), kal dksa (26888), mantra . a . . . dksa (28995ab), and tattvadksa (295cd307ab). The universe contains six adhvan, . . which the six initiations purify in relation to the initiand: the ways of the hierarchy of worlds (bhuvan dhvan), words (pad dhvan), phonemes (varna-/bj dhvan), the a a a . primordial creative powers (kal dhvan), mantras (mantr dhvan), and reality principles a a (tattv dhvan).10 This cosmological system is explained in some detail in the preceding a chapter, BraY xxxi.11 It appears that the BraY viewed the initiation of the Six Ways a a
10 The sequence of presentation of the six adhvan differs within the tradition, and does so even within the text of the BraY . On this notion of six adhvan, see the entry in T ntrik bhidh nakoa, vol. i, 11011. a a a a s See also the entries for kal dhvan and kal dksa in T ntrik bhidh nakoa, vol. ii, 7476. a a . a a a s 11 That the cosmological systems explicated in BraY xxxi bear upon initiation is stated explicitly: a

196 as central to destroying the impurity binding the initiand.12 The subject of tattvadksa . is taught in more detail in BraY xxxvi, which bears this as its title. a BraY xxxii hence does not give a full account of initiation. Indeed, it even lacks a mention of showing the mandala to the disciple (mandaladarana), an act so central s . .. that it is sometimes equated with initiation. This disjunct presentation of initation makes it difcult to determine precisely what ritual sequence was envisioned. The initiation mandala itself is described in BraY iii, , the elaborate mandala of the great a .. .. rite (mah y ga) mentioned above. In its closing section, this chapter also provides a a a brief description of aspects of initiation ritual, including preliminary worship and mandaladarana. In the latter sequence, the blindfolded disciple is made to cast a ower s .. into the mandala, the point on which it alights establishing his or her initiatory clan. .. A section of BraY iv provides a concordance indicating which Mother-goddess clan a an initiate belongs to depending upon where in the mandala his or her ower lands.13 .. This probably pertains to the initiands known as the samayin and putraka; a different a concordance is provided in BraY xxxiii, which applies to the ac rya and perhaps a
evam yo vetti devei sad sivapade sthitam | s a . adhv nam tu mah devi sa dksam kartum arhati a . a . .
deve i ] corr.; devesi Bya s Bya arhati ] em.; arhasi Bya

dksam ] em.; dksa . . .

adhv nam ] em.; adhv nan Bya a . a

One who thus understands the adhvan situated at the level of Sad siva is t to perform a initiation. BraY xxxi.47. a 12 a Note the initial instruction to the ac rya, BraY xxxii.3cd5: a paum p rsvesu samsth pya daksinen tmano budhah 3 s . a a . . . . a . sahaj gantuk n m tu samsargk n m tathaiva hi | a a a. a a. . paos tu grahanam kury n n diyojanapurvakam 4 s a a. . . An intelligent [ c rya] should situate the bound soul at his own side, to the right. a a He should then perform seizure of the innate, adventitious ( gantuka), and natural a (samsargika) [karma], preceded by fusion of the [ c ryas and initiands] n ds. a a a. . It seems p rvesu in 3c must be singular in meaning although plural in form. Compare for example as . the Sv yambhusutrasamgraha, which describes the sixfold dksa as accomplishing a series of transformaa . . tions of m y , primal matter, probably in the sense of m yyamala, the impurity associated with m y . a a a a a This nds support in Svacchandatantra 4.80, according to which adhvauddhi accomplishes p saccheda, s a severing of the fetters [of mala, impurity], fetters elsewhere specied as m yya, belonging to m y a a a (Svacchandatantra 4.129). 13 BraY iv.888898, quoted in part in the introduction to chapter 2. Occurence of this material in a BraY iv seems curious given the chapters focus upon ritual icons, and given that the immediate a context is description of ritual bells (ghan. a). . t

197 s dhaka.14 a A separate chapter, BraY xxxvii, treats what it calls samaykarana, the making a . of a Pledge-holder, suggestive of what other sources call samayadksa.15 Confusing . matters, this is said to be a procedure for s dhakas,16 suggesting that it is instead a s dhak bhiseka, the consecration which makes one a s dhaka. Included in this ritual a a . a is the initial initiatory sprinkling (proksana) of the candidate, and the rite of the . . ofciand placing his mantra-empowered hand upon himhere a saktihasta, akti s hand, rather than the sivahasta, Siva-hand mentioned in other layers of the tradi tion. These two acts normatively precede and follow mandaladarana, respectively.17 s .. It is possible that this chapter provides a brief outline of the combined samayadksa . and s dhak bhiseka, for the s dhaka is after all the primary practitioner to which the a a . a BraY addresses itself. Neither this chapter nor BraY xxxii provides an account a a of the initiatory Pledges (samayas), which are not ennumerated until BraY lxi and a lxxxv. The lists of the latter two chapters are moreover discrepant. BraY lxi, a long a a and remarkably self-contained chapter called the sutr dhik rapatala (chapter on who a . has entitlement to the teaching) or the Tilaka, opens with the menstruating Great Goddess, Bhairav, asking for a synopsis of everything which has been said and not said.18 In this chapters eclectic discussion of food offerings (caru), magical pills (gutik ), mantra, skulls, and much else, there occurs a list of Pledges (119cd127). . a Entitled samay dhik ra, the chapter on the Pledges, BraY lxxxv also ennumerates a a a Eight Pledges, and, furthermore, teaches mantramay dksa, the initiation consisting . of mantras. Mantramay dksa appears primarily to involve destroying the fetters .
The passage from BraY iv is quoted in chapter 2 (n. 14), while an excerpt from the passage in a BraY xxxiii is quoted in chapter 4 (n. 82). a 15 This chapter calls its second subject brahm ndotkarsana, drawing down the cosmos. What this a. . . . has to do with initiation is presently unclear to me. Curiously, in the colophon this chapter calls itself tattvadksapatalathe subject and title of BraY xxxviiiprobably in error. . . a 16 Viz. 18ab, s dhakasya sam khy to vidhir esa may tava (this procedure I have taught you is for the a a a a s dhaka. a 17 Sanderson, Religion and the State: Initiating the Monarch in Saivism and the Buddhist Way of Mantras (forthcoming). 18 The opening half-loka reads, rtumat mah dev bhairav v kyam abravt, while the nal half-loka of s a a s . the goddesss question is, ukt nuktam aesam tu samgrahastham bravhi me (6ab). a s . . . .

198 of impurity associated with the Course of Mantras (mantr dhvan), and ends with a bestowal of the s dhaka-consecration. a Consecration (abhiseka) forms the subject of BraY xxxiiis approximately foura . hundred verses. In contrast to the preceding chapter on initiation, here the description of the ritual appears complete and self-contained. The ritual is structured along the lines of Saiva initiation and involves mandaladarana (viewing the mandala), in s .. .. which the initiand is led blindfolded to the mandala, made to cast a ower there.. upon, and given a clan name accordingly; and the rite of incubation,19 involving prognostication using a tooth-pick (dantak s. ha), and afterwards based upon the cana. t didates dreams. That the ritual described concerns the consecration of the ofciant, a ac rya, is made explicit through the nature of the Pledges taken, which include giving initiation and explaininng the scriptures.20 The chapter ends with a description of ritual worship of the guru and bestowal of the sacredotal fee. Mantrasamkalanavidhi, the procedure for preparing the mantras, comprises the . subject and title of BraY xxxiv. This massive chapter outlines the tedious syllablea a by-syllable sequence of mantra-incantation and re-offerings by which an ac rya prepares or empowers the full pantheon of basic mantras.21 BraY xxxi, having the a generic title prakriy patala, is connected to dksa insofar as it describes the cosmologia . . cal systems necessary for performance of the sixfold adhva-initiation.22 It provides a detailed exposition of the hierarchy of worlds, tattvas, etc., including their presiding mantra-deities. These collectively comprise the brahm nda or universe. The chapter a. . also describes the hierarchy of rudras who lie beyond the brahmanda, bearing it (114). . . . BraY xxxvs stated title is n dsac rapatala, the chapter on the movement of the a a. a .
19 This term has been adopted by Sanderson from scholarship on Greco-Roman religion, where it refers to ritual sleep in a sanctuary in order to obtain a dream. Sanderson, Religion and the State: Initiating the Monarch in Saivism and the Buddhist Way of Mantras (forthcoming), quoting Hornblower and Spawforth. 20 BraY xxxiii.350363. a 21 a That this material is directed toward the ac rya is claried in the opening verse, which describes the mantrasamkalanavidhi as something by the mere learning of which one is t to perform initiation . (yena vij tam trena dksam vai karttum arhati, BraY xxxiv.1cd). a a . . a 22 See above (n. 11).

199 n ds, and it describes the movement of the supreme deity within the channels of a. the body. This chapter gives an exposition on the bodily channels and the principal goddesses which they embody, these forming a ninefold cakra. Its occurrence within the cluster of chapters pertaining to initiation perhaps has its basis in this knowledge a being necessary for the ac rya to accomplish yogic fusion of the disciples channels (n dsandh na), as well as the nal act of linking the disciple to the supreme deity a. a (ivayojanik ) after completing the initiation which puries the adhvans.23 s a
cult images (pratima, linga)

Iconography, iconometry, and rituals involving images (pratim karma) form the suba jects of the BraY s massive fourth chapter. One of the most signicant sections of a the text, BraY iv contains about a thousand verses on the subject of religious images a (pratim ) and other ritual objects, including their rites of empowerment and worship. a This unique iconometric and iconographic treatise merits extensive study in its own right. Characteristic of chapter four is its hierarchical classication of deities and their images as semi-divine (divy divya), divine (divya), and supra-divine (divy dhika). In a a the semi-divine are included the Hundred Rudras, female spirits called l m s, and a a the demonic r ksasas. Divine images comprise the sets of goddesses included in a . the BraY s mandalayogins, guhyak s/devs, kinkars/duts, and the Mothersas a a .. well as their male counterparts (pati, vra), rudras and yogins of the sacred elds, and lokap las. Also divine are the gana-lords Nandin and Mah k la, and the river a a a . a goddesses Gang and Yamun , deities who typically serve as guardians (dv rap la) a a a of a temples sactum. In the category of supra-divine come the high deities: the supreme Sakti, Sad siva, the Three Saktis (V m , Jyesth , and Raudr), Srkantha, a a a .. .. a and Um pati. The text contains detailed instructions concerning image measurea ment and iconography, rituals of image empowerment (pratis. ha), the specications . t. and empowerment rituals for skull-bowls, skull-staves, rosaries, etc., and much else.
On the processes of n dsandh na and sivayojanik , see Sanderson, Religion and the State: Royal a. a a Initiation (forthcoming).

200 Several other chapters contain material relevant to images and their ritual (pratim karma). In particular, BraY vi, a chapter of twenty-two verses, describes the a a icononography of the mandala goddesses sculpted from mud of the cremation .. ground and inected in various ways according to each of the nine pantheon congurations (navay ga) and the magical results sought. Specications and empowa erment rites for lingas and their pedestals (p. ha) are taught in BraY lxxxvi, the t a lingalaksanadhik ra, chapter on the characteristics of lingas. These are threefold: a . . manifest (vyakta) lingas, hidden/unmanifest (avyakta), and manifest-cum-unmanifest (vyakt vyakta), the latter being comprised of lingas with sculpted faces (mukhalingas). a Somewhat detailed iconometric and iconographic details are provided. This chapter, moreover, seems to be the only one expressedly concerned with temples, although the Goddesss request to learn the characteristics of temples (pr s dalaksana) meets aa . . with disappointingly little detail. BraY xciii discusses moveable lingas for private a use in the context of s dhakacary , ritual conduct of the s dhaka. a a a

4.2 structure and textual strata

In the form transmitted, the BraY contains several sectional divisions, but their dea marcation is in part problematic. The opening verses of chapter fty announce a new section of the text. This passage recapitulates the subjects of several earlier chapters, after which the Goddess poses a new set of questions; these new questions in turn presage the subjects of many subsequent chapters, up to and including BraY a lxxxii. Chapter ftys introduction also appears to refer to the BraY as twofold, a although not unambiguously, and it might use the term satka (sextet) to refer to . . its two halvespresumably designated thus because of consisting of six-thousand verses, in principle.24 In this respect, the BraY might have served as model for the a

BraY l.16ab: a devy uv ca a srutam dv daas hasram evam tu dvigunam vibho | a s a . . . . yantramantrasamopetam karmai ca bahubhi citam 1 s s . s dhan r dhanair yuktam mudr mantrasamanvitam | a aa a .

201 Jayadrathay mala, a text organized into four satkas. a . . The second major section of the BraY therefore commences with chapter fty. a This marks the halfway-point chapter-wise, yet in verse count is approximately twothirds into the text. Virtually all integral facets of the BraY and its ritual systems a are taught within the rst forty-nine chapters, from the primary mantra systems to initiation and consecration. On the other hand, the section beginning with chapter ftythe second satka?has a miscellaneous character, containing a large number . . of short, often untitled chapters, as well as quasi-independent texts of the kalpa genre devoted to deities often marginal to the BraY s basic mantra system. It nonetheless a contains some material of signicance to the whole, whether or not belonging to the text in its earliest form, such as discussions of the meanings of the titles Brahmay mala and Picumata.25 Adding to the impression of its second half having disa parate content, the BraY contains two further sections, the contents of which are not a intimated in the introduction to BraY l: with chapter eighty-four, apparently, begins a the Uttar dhik ra (Addendum Section) or Uttaratantra ( Addendum Tantra), and a a this is followed, from chapter eighty-eight, by the Uttarottaratantra (Latter Addenkarmai ca vividh k raih sadyapratyayak rakaih 2 s a a . a . srutam bhagavat purvam mantram mataprak sakam | a a . . . nav tmakavidh na ca navay gasamanvitam 3 a a a mudr mandalamantrai ca vratacary samanvitam | a .. s a s dhan r dhanair yuktam n n guna-m-alamkrtam 4 a aa . a a . . . nikhila ca srutam sambho tatra karm ny anekaah | a. s . . ::::::: id nm daas hasre dvibhi caranasamyutaih 5 a . s a s . . . satke tu prathame deva khy tam karmasahasrakam | a . . . ::::::: . khecarna ca sarv s m bhucarna ca s dhanam 6 . a a. a pratim laksanam n tha pratis. h tadanantaram | a . . . a .t a kalpas dhanakarma ca n n vastusamanvitam 7 a a a
1a s hasram ] corr.; s hasramm Bya 2a r dhanair ] corr.; r dhanaih Bya yuktam ] em.; yukta a a a a . . . Bya 4c yuktam ] em.; yukt Bya a 5a id nm ] em.; id n Bya a . a 5b karm ny ] em.; karm ny a. a . Bya aneka ah ] conj.; ake[tama?] Bya s . 6a prathame ] conj.; prathamam Bya 6c . sarv s m ] em.; sarvv s Bya a a. a a

1ab refers to the Dv daas hasra (a titular epithet of the BraY meaning The Tantra of Twelve-thousand a s a a Verseson which see chapter 5) as being dviguna, presumably in the sense of twofold. The possibil. ity of this passage using the term satka to refer to these two divisions depends upon the emendation of . . prathamam to prathame in 6a; but an adverbial prathamam (rst) might be possible. What follows this . is an index of subjects that appear already to have been covered, such as pratim laksana (BraY iv) and a . . a worship of the khecarcakra (BraY xiv). a 25 See the rst two sections of chapter 5 of the present thesis.

202 dum Tantra). Of all the material in the BraY , this arouses the greatest suspicion of a being a subsequent addition. In the opening of BraY lxxxiv, the Goddess restates a the subjects of several preceding chapters of the second satka. She then poses ques. . tions that seem to intimate content from a number of remaining chapters, including material from both the Uttara- and Uttarottaratantras.26 This suggests that the latter two sections were composed as a single unit. The basic structure of the BraY thus appears threefold or fourfold: 1) the rst a forty-nine chapters; 2) the section from chapter fty to eighty-three, which includes, for instance, kalpa-texts of particular deities, chapters closely connected with the cult of yogins, and discussions of the texts titles; 3) the Uttaratantra of chapters lxxxiv lxxxvi; and 4), the Uttarottaratantra, chapters lxxxviici, these latter two sections being closely connected. It is possible that the rst section was considered to constitute the rst satka, the second satka comprising section two, or else the nal three . . . . sections together. However, complicating the question of the BraY s structure, the a closing verses of its nal chapter make the additional claim that the text consists of ve sutras.27 One might expect these to be sections of the text, for the division into books called sutras has precedent in the ancient Niv satattvasamhit , which consists s a . a of a Mulasutra, Uttarasutra, Nayasutra, and Guhyasutra, prefaced by a fth section, the Mukh gama. Each of these contains multiple chapters. In addition, the Niv sak rik , a s a a a
In particular, note the reference to the conduct of s dhakas (s dhak n m vrttam, in 7a) in the list a a a a. . . of future topics, which appears to intimate chapters xcixciv. This passage is quoted below (n. 32). 27 BraY ci.3132: a brahmay malatantredam laksap d dhik gatam | a a . . a a satakotyujjval t tantr t s r t s ratarottaram 31 a a aa a . sthitam dv daas hasram pacasutrojjvalam matam | a s a . . . may te kathitam bhadre bhadrasiddhiprad yakam 32 a a
kotyujjval t a 26

31b p d dhik gatam ] em.; p d dhik m gatam Bya ; p d dhik msatah conj. Sanderson a a a a a a. a a a. . . ] em. Sanderson; kotyajval Bya a . .


O good woman, I have taught you the Brahmay malatantra, which grants felicitous siddhis, a and emerged from the Tantra of One and a Quarter Hundred-Thousand. The Tantra of TwelveThousand exists (sthita) as the highest essence of essences from the Tantra endowed with a billion [verses]. It is held to be endowed with ve sutras. The emendation laksap d dhik gatam is supported by the colophon of the Matas ra quoted in chapter a a . a a 5 (n. 140). Alexis Sandersons edition of these verses has been published in Dominic Goodall, Bhat. a .t R makan. has Commentary on the Kiranatantra. Vol. 1: Chapters 16, lxxiiiiv. a .t .

203 which appears to belong to a somewhat later period, posits itself as the fth sutra of the Niv sa.28 s a How precisely the hundred and one chapters of the BraY could be divided into a ve sections called sutras is never explicated. The text nonetheless contains several other nebulous references to containing or consisting of sutras,29 and even refers to specic sections of the text as such. For instance, the terms mulasutra and purvasutra occur throughout the BraY ; yet it is not entirely clear what either refers to. The latter a . occurs particularly in the xed expression purvasutrena coditam, which appears simply to mean stated previously.30 On the other hand, mulasutra might refer to a specic section of the textpresumably at least the rst forty-nine chaptersbut nothing in the BraY identies itself as such explicitly.31 In BraY lxxxiv, the beginning a a of the Uttaratantra, the Goddess asks to hear the denitive judgment on the Root Tantra (mulatantravinirnaya); recapitulating the subjects of much of what precedes, . this passage appears to contrast a mulasutra section (=mulatantra?), uttara[sutra], vinaya[sutra], and sangrahasutraalthough the uttara and vinaya sections might be identical.32 It seems possible that the vinayasutra is BraY lxii, for this chapter once a
See Sanderson, The L kulas: New Evidence of a System Intermediate between P c rthika a a a P supatism and Agamic Saivism, 15253. a 29 Note e.g. BraY lix.77: a ucchusmasambhavam tantram devya cocchusmasambhav h | s a. . . . . a tvayoktam tu mah deva sutr dibhi mahevara 77 a s .
devya ] conj.; devy Bya s a

O Mah deva, through sutras and so forth, you have taught the tantra arising from a Ucchusmabhairava, and the goddesses arising from Ucchusmabhairava, O Mahe vara. s . . Neither sutra nor adi (etc.) is clear in meaning here. 30 Note for instance the opening verse of BraY xiii: a . atah param pravaksy mi purvasutrena codit h | a. . . . a navay g mah devi tan me nigadatah srnu 1 a a a . . .
codit h ] em.; codit Bya a. a nigadatah ] em.; nigaditah Bya . .

Hereafter, I shall teach the Nine Pantheons indicated previously [lit. revealed with/by a . previous thread (purvasutrena codit h)], O Great Goddess; listen while I explain this. a. The Nine Pantheonsthe subject of the chapter quotedwere listed without elaboration in BraY iii.2 a 4, to which the present verse probably refers. 31 Problematizing the possibility that mulasutra refers to the rst fty-odd chapters, the phrase . mulasutrena coditam also occurs in BraY xxxiii (77b); though unclear, this might mean taught [prea . viously] in/with the root (mula) section/thread. 32 BraY lxxxiv.111: a devy uv ca a

204 appears to call itself the vinaya.33 As for the sangrahasutra, this is undoubtedly BraY lxi, a remarkably self-contained treatise called the Tilaka (discussed below). a By Uttarasutra, in contrast, the redactors presumably had in mind the Uttara- and perhaps Uttarottaratantras. While the BraY certainly contains multiple sections, the notion that it consists of a sutras appears contrived, and perhaps belongs to a late phase of its redaction. Indeed, the principal references to the BraY consisting of sutras belong to the concluding a sections of the text, probably its nal textual stratum: BraY lxxxiv, of the Uttaraa tantra; and ci of the Uttarottaratantra. Two factors might have contributed to this conception: a smattering of vague references in the text to sutras, and the sutra model of the Niv sa corpus. The latter could have stimulated some redactors to an unlikely s a interpretation of the former. In turn, the statement in BraY ci that the text consists a of ve sutras appears to have stimulated further speculation within the tradition, for such is evident in the Jayadrathay mala. As Sanderson notes, chapter forty-four a of the latters rst satka, called the Brahmay malanirnaya (A Denitive Judgment on a . . .
mulasutravibh gan tu uttaram vinay tmakam | a a . kalpaskandhavidh nan tu sasanan tu yath sthitam 1 a a kathitam sangraham sutram kulak lartujvalam | a . . . . mah mandalay gan tu tantre smin kathitam tvay 2 a .. a a . vauyogavi[vikt] / | s (hi?) tantre []smi cottare vibho 3 vivacakrasuobh dhya c rodayasamanvitam | s s a. a yogim trganopetam y ga ca picusambhavam 4 a. . a . mantroddh ra ca devea sarvv varanasamyutam | a s a . . samksiptatarayogena yogasadbh valaksanam 5 a . . . . sangrah rthavibh gena samay ye viesag | a a a s . a caravas tu yath prokt rtujam kusum mrtam 6 a a. a . . s dhak n n tath vrtta ca(s - a?)sth satkas dhanam | a a a a . a. . a . dksasesavidh nan tu sarvvatomukham eva ca 7 . . a y gaksettrevar na ca navay garas dhanam | a . s a. a a samastaikatra yogena mulatantravinirnnayam 8 .. kathayasva pras dena ukt nukta ca me prabho | a a bhairava uv ca a a a a mulasutr dik n n tu kramam s dhanalaksanam 9 . a . . durlabham ttrisu +lokesu+ samay c rap lanam | a a a . . . y gam vidhis tath j na cakram yoga ca sobhanam 10 a . a a . kathay mi mah devi yat tvay coditam balam | a a a . The text and interpretation of this passagepresented as is in Bya , the oldest codexare doubtful in several instances. 33 BraY lxii 64d, vinaye smim mah tmane. a a .

205 the Brahmay mala), claims that texts of the Brahmay mala cycle contain ve sutras: a a a mula-, guhya-, naya-, uttara-, and uttarottarasutra.34 This appears to be an attempt to reconcile several things: the claim that the BraY consists of ve sutras, the ve-sutra a model of the Niv sa corpus, and the actual sectional divisions of the BraY for s a a unlike the Niv sa, the BraY possesses both an Uttara- and Uttarottaratantra/sutra. It s a a is possible that the BraY s vinaya section was considered analogous to the Niv sas a s a Nayasutra. The Jayadrathay mala does not, however, speak of the BraY s sutras strictly a a as sections: on the contrary, it describes the mulasutra, for instance, as the words ath tah, with which begin many of the BraY s chapters and sections, while the a . a guhyasutra is said to consist of teachings on secret matters, such as the yogic bodys channels and the clans of yogins.35
Concerning the BraY s purported division into sutras, and the elaboration of this in the Jayadrathaa y mala, Sanderson expresses the following opinion: in the case of the BY the schema of the ve sutras a has nothing to do with the sequence of sections seen in the Ni v sa. For it has no sequential divisions s a other than its patalas. It seems, therefore, that the JY understands the sutras to be ve classes of teaching . within the otherwise unitary Tantra, in the same way that it claims to have four P das, though it lacks a a corresponding division into a sequence of four parts bearing their names. Quoted in Goodall (Bhatta .. R makanthas Commentary), lxxiiiiv. This assessment of the Jayadrathay malas understanding of the a a .. BraY s sutras seems convincing; note however that the BraY does contain several divisions, beyond its a a division into chapters. 35 Jayadrathay mala i.44 (folio unnumbered): a . ath tah sabdamulam tu sarvatantr rthasucakam | a . a . . tat mulam svalpasutram ca guhyam guhyavivecakam | . a nir c r vadhut rtham n dn digatih kram t | a aa a . a. a. . bindubjakal yoganavay gavivecakam | a a mantrajv mrt karsayoginkulakrttitam | a . a . . guhyasutram tad uddis. am b hy dhy tm rthav cakam | t . a a a a a .
gatih a a . kram t ] em.; gatikram t cod. navay ga a 34

] em.; vay ga cod. a

ath tah is the root of speech, indicating the meaning of all tantras; and it is the very a . pithy mulasutra. Guhya indicates secrets: the meanings of beyond regulations (nir c ra) a a and immaculate (avadhuta), the channels [of the subtle body] and movements of the channels, in sequence; it [also] indicates bindu, bja-mantras, the kal s, yoga, and the Nine a Pantheons; [it is that] by which are proclaimed mantra, extraction of the nectar of living beings, and the clans of yogins. This is called Guhyasutra, expressing [both] external and internal meanings. This chapter of the Jayadrathay mala makes the additional claim that the BraY contains seven sutras, a a from which derive eight Matatantras: brahmay malam etad dhi sutrair bhinnam tu saptabhih | a . . saptamam ca dvidh bhinnam ato jeyam mat s. akam | a a. t . . . For the Brahmay mala is divided according to seven sutras, and the seventh is divided in a two; because of this the Matatantras are known as an octad. Codex of Vi vabh rat University (Manuscript Division, Department of Sanskrit). This is a Devan gars a a script paper manuscript of Nepalese provenance, written on twenty-six folios and containing chapters

206 The evidence reviewed above concerning the structure of the BraY provides a some indications that this heterogenous text incorporated new material over time. In addition, a number of discrepencies in the BraY s content suggest redactional disa continuity. One of these, mentioned in the discussion of the BraY and Laghuamvara a s . in the previous chapter, concerns typologies of the s dhaka: while a threefold clasa sication of the s dhaka is presupposed in the earlier chapters, especially xxivxlv, a the Addendum Tantras explicate a fourfold typology that appears to be a subsequent innovation.36 Less suggestive of historical development, other disparities nonetheless point toward heterogeneity. For instance, the BraY has two chapters treating the a subject of chomm or secret signs: lv and lxxiii. Striking is the lack of relationship a between these chapters, which overlap in context and many particulars, but provide no clear indications of belonging to the same system.37 This sort of heterogeneity could point toward redactional discontinuity, but it might also reect the disparate sources drawn upon by the BraY s redactors. a Signicant to the question of the BraY s stratication is the conception of it cona taining twelve-thousand verseshence its epithet, Dv daas hasra[ka], The Tantra of a s a Twelve-thousand Verses. References to this conception of the BraY s length are a absent from chapters iixlix, aside from colophons, and it is conceivable that this notion of the texts size does not belong to its earliest textual stratum. The text as we have it consists of upwards of thirteen-thousand verses,38 fewer than eight-thousand of which belong to the rst forty-nine chapters. This suggests that the notion of the text consisting of 12,000 verses is unlikely to predate some or most of the chapters from fty onwards. On the other hand, the brevity of the Uttara- and Uttarottara3542 of Jayadrathay mala, satka i. Cf. the discussion of the Vi vabh rat codex of the BraY , in the a s a a . . introduction to part ii. 36 The closing verses of BraY lxxiv rst intimate the fourfold typology of the s dhaka explicated in a a a the Addendum Tantras, for here the practitioners called the ar dhaka and sarv tman are rst mentioned a ( r dhakas tu devee sarvaih sarv tmaka[h] smrtah, 211ab). aa s a . . . . 37 In particular, note that lv.10137 and lxxiii.1640 overlap considerably; the chomm s of these a sections often involve similar secret signs made in similar contexts, for which distinct ritual meanings are nonetheless provided. 38 See chapter 5, section 5.

207 tantraseighteen chapters, but only sixteen-hundred odd versesin no way rules out the possibility that these were added subsequently. As mentioned, several chapters of the second half of the BraY , from chapter a fty, appear remarkably self-contained. This is in part because a number belong to the kalpa genre: treatises setting out the practices of mantra-propitiation connected with a particular deity. However, at least two chapters arouse the suspicion of being independent tantras subsequently redacted into the BraY a phenomenon ata tested, for instance, in the case of the Jayadrathay mala.39 Two chapters in question a are BraY lxi, the Tilaka[tantra]; and BraY lxxxii, the Utphullakamata, the titles of a a which match texts quoted by Abhinavagupta. The Utphullakamata, Utphullakatantra, or Utphull dhik ra,40 a chapter of roughly 195 verses, teaches practices connected with a a the deity Utphullakabhairava and the nine-syllable utphullakamantra. A text by the .t name Utphullakamata is listed in the Srkan. hyasamhit as the seventh of eight mata. a tantras.41 Abhinavagupta unfortunately provides too little information concerning the Utphullatantra he cites to link it with BraY lxxxii.42 Sanderson suggests the ina triguing possibility that this chapter has been derived from an independent tantra of this name.43 Indeed, the Utphullakamatas elaborate Mandala of the Nine-syllable .. Mantra (nav tmaka cakra) contains astrological elements unique in mandalas of the a .. BraY .44 At the same time, the chapter in the form transmitted clearly identies itself a
See chapter 3, n. 180. The title Utphullakamata is rst intimated in BraY l, in its list of subsequent chapter subjects; two a references to this title occur within the chapter (11d and 17b). In addition, the chapter calls itself the Utphullakatantra in its opening verse (atah param pravaksy mi tantramm utphul[lu]kam param, 1ab), while . . . a . . its colophon gives the name Utphull dhik ra. a a 41 Srkan. hyasamhit , as quoted by Jayaratha, commenting on Tantr loka 1.18. The verses listing .t a . a .t the eight matatantras appear absent from the Srkan. hyasamhit manuscript transcription provided in a . Hanneder, Abhinavaguptas Philosophy of Revelation. Incidentally, heading this list is the Rakt mata, named a after the rst of the Four Devs in the mandala of the BraY s Kap lsabhairava. a a .. 42 In Tantr loka 29.166a, Abhinavagupta cites the Utphulla, identied by the commentator as the a Utphullakamata, alongside yogin cult scriptures such as the Siddhayogevarmata (assuming this is the s referrent of Srsiddh ) and the lost Nirmary datantra (The Tantra of No Constraints). He cites these as a a a sources for the erotic adiy ga (primordial rite) expounded in this chapter of the Tantr loka. There does a not appear to be a link between the BraY s Utphullakamata and a ritual of this type, however. a 43 Sanderson, handout from an unpublished lecture, K lkula: Abhinavagupta and the Kramaa system in the Light of a Newly Discovered Corpus (Hamburg University, 1981). 44 This elaborate cakra, based upon Utphullakabhrairava in a circle of four devs and four dutshence mirroring Kap lsas mandalaincludes (alphabetical representations of) the signs of the zodiac (r si), a a ..
40 39

208 as a section of the BraY ,45 suggesting a potentially complex history. a Chapter sixty-one of the BraY possesses several titles: Tilaka (perhaps, [Tantra] a a of the [Magical] Bindi?),46 Sutr dhik rapatala, Sangrahatantra, and Sangrahasutra.47 a . About two-hundred and seventy-ve verses long, this chapter claims to be the essence (s ra) of the Picumata/Brahmay mala,48 while nonetheless introducing novel matea a
calendrical dates (tithi), constellations (naksattra), and planetary conjunctions (yoga). . 45 Reference to the Brahmay mala occurs, for instance, in 101cd (laksap d dhike khy tam viesad a a . s . . a a brahmay male). a 46 The contextual signicance of this title word is unclear. The application of magical bindis (tilaka) is the focus of only one brief section of the chapter (237cd49), which might nonetheless be the source of its name. Less probable, the title could be connected to the term tilak , used as an epithet of the supreme a a Sakti in another chapter, BraY lvii.12 (tilak khy mah sakti j narup manonman | s siv sivasamyukt a a a a a a a a . aghor ghoran san). Tilaka might simply be intended in the sense of decorative forehead dot (bindi) a the chapter therefore being a decorative ornament to the BraY , which it claims to synthesize. Cf. the a a a title S rad tilaka[tantra]The Ornament of Sarasvat Tantra. 47 a The title Sutr dhik rapatala is given in the chapter colophon, while Sangrahatantra occurs in a . 145ab (tad atra samgrahe tantre samksiptara ucyate). In a synopsis in the opening verses of BraY lxxxiv, a . . . this material is referred to as the Sangrahasutra (kathitam sangraham sutram, 2a). . . . 48 See for instance BraY lxi.69ab: a bhairava uv ca a s dhu s dhu mah bh ge j navij nasampadam | a a a a a a samgrah rthaviesam tu kathay mi tav khilam 6 a s . . a a . j te picumate tantre daa dve ca sahasrake | a s tatra sthitam mah devi saktij namahodayam 7 a a . tantrasya s rabhutam tu tad atra tilakam matam | a . ye na j nanti tantraj s te bhramanti vidambak h 8 a a a. . k ryasiddhim na payanti na ca y nti par m gatim | a s a a. .
6b sampadam ] em.; sampado Bya 7c sthitam ] em.; sthita Bya .

Bhairava spoke: Excellent, O fortunate woman, excellent. I shall teach you (?) a par ticular digest treatise (sangrah rtha ?) in its entirety, replete with wisdom and knowledge. a After the Picumatatantra has been learnt, the Tantra of Twelve-Thousand [Verses], therein lies a treasure of knowledge of the Sakti, O Great Goddess, the essence of the tantraknown here as this Tilaka. Those who do not know [this] wander about as impostors, [even if they] know [other] tantras. They do not see the fruition of their rites, nor do they attain to the highest destination. (On the possible use of the term artha as treatise or tantra, see the annotation to the translation of BraY i.65.) Compare also verses 25051: a j tv picumatam tantram tilakam c py apacimam | a a s . . . a pal lam iva dh ny rth tyajet mantr n aesatah 250 a a a a s . . mulatantr rthasadbh vam samastkrtalaksanam | a a . . . . samksepavistaram j tv tato mantr prasidhyati 251 . . . a a
250d mantr n ] em.; mantr m Bya a a 251a sadbh vam ] conj.; sa - - Bya a .

After learning the Picumatatantra, and the Tilaka as well, () which has nothing afterwards (apacimam) (?), one should give up [all other] mantras without exception, like a vegetarian s [would give up] meat. After learning the essence of the meaning of the Root Tantra, with [all] its characteristics brought together, both in abbreviation and at length, the mantrin then attains siddhi.

209 rial.49 The Goddess opens her questions asserting that she has already heard the Y malatantra,50 while other passages too draw a distinction between the Tilaka and a the Picumata or BraY , clear indications of the independence of this section.51 Consisa tant with its self-identication as the essence (s ra) and digest (sangraha) of the a BraY , the Tilaka has the appearance of a self-contained treatise, containing concise a accounts of topics spanning initiation, the Pledges, a mantra-system (mantroddh ra) a and its associated mandala and worship, substances and objects used in ritual, yoga, .. and more. Regardless of whether it was originally written as part of the BraY , the a Tilaka is deeply anchored in this tradition, as is illustrated by its focus upon several characteristic topics, such as the Nine Pantheons and the smarana-mantra.52 It is pos. sible that the Tilaka had an independent life; however, there is insufcient evidence for identifying it with the Srtilakaastra cited by Abhinavagupta.53 A Tilakatantra s
49 A striking example is the introduction of the nav tmamantra, a ninefold mantra-pantheon characa teristic of the Niv satattvasamhit , which BraY lxi appears directly or indirectly to draw upon. See the s a a a . discussion in section 3 of this chapter. 50 BraY lxi.1abc: rtumat mah dev bhairav v kyam abravt | pr k srutv y malam tantram (The mena a a a a a . . . struating Great Goddess, Bhairav, spoke the following words: having earlier heard the Y malatantra, a . . . ). 51 See verses 78 and 250, quoted above (n. 48), and 35:

a. naimittike ca devee caravo y gasangat h | s a y male tu pur siddh [s] tilake prakatkrt h 35 a a a . . a. And the food offerings connected with the [Nine] Pantheons in the occasional rites, established earlier in the Y malatantra, have been made explicit [here] in the Tilaka. a . On the Nine Y gas, see the previous section of this chapter; on the smarana-mantra (hum), see a . section 3 of chapter 5. 53 In the Par trimsik vivarana, Abhinavagupta attributes to the Srtilakaastra and the Srbhargaikh the a . a s s a . idea that assiduous practice of the nondual rites of heroes causes ones gl ni (lassitude, inhibition?) a to vanish suddenly, effecting the merger of individual identity ( vea) into the heart of Bhairava. After a s quoting the pratka of Spandak rik 3.8 (gl nir vilumpik dehe), he remarks, seyam yad jhatiti vigalit a a a a a a . . bhavati tad nirastap savayantranakalanko bhairavahrday nupravis. o bhavatti sarvathaitadabhy se yatitavyam | a a a t a . . . srtilakaastre yam bh vah | srbhargaikh y m api uktam (when this very [gl ni] suddenly dissolves, then, s a . s a a a . being one who has cast off the blemish that is the bound souls afiction, he becomes merged into the a heart of Bhairava. One must hence in every respect endeavor in this practice [of the adiy ga]. This idea is present in the Srtilakaastra; this is also stated in the Srbhargaikh ). From the commentary s s a on Par trimsik 9cd18ab (p. 235). BraY lxi enjoins one to perform ritual with a nondual mental a . a a disposition, but no particular statement correlates closely with the idea and phrasing Abhinavagupta attributes to the Tilakatantra. Note for instance BraY lxi.26-27ab: a kulasiddhiprasidhyartham devn m agratah sthitah | a. . . . advaitabh vasampanna carukarmany aankitah 26 a s . s . tena pr sitam trena khecarsiddhim apnuy t | a a . a
26b agratah ] corr.; agrata Bya sampanna ] em.; sampanna Bya s . Bya a ankitah ] corr.; asamkitah Bya 27a pr sita ] pr sita Bya s a a . . . 26d karmany ] em.; karmamny . .

210 .t in fact gures in lists of scriptures in the Siddhayogevarmata and Srkan. hyasamhit , s . a suggesting that there might have existed an ancient scripture by this title distinct from the BraY .54 a Evidence for the independent existence of the BraY s Utphullakamata and Tilaka is a hence inconclusive, and it cannot be said with certainty that the BraY incorporated a previously independent treatises. Another chapter tooBraY liv, the Mah k laa a a matapossesses a title matching a scripture early enough to be mentioned by B na;55 a. but in this case as well, the grounds are insufcient for linking the texts. Nevertheless, internal evidence from the Tilaka chapter, in particular, suggests that new material was added in the course of the BraY s transmission, and other chapters from this a portion of the BraY merit similar suspicion. a Altogether, the BraY has the appearance of a layered and composite text, the a production of which involved multiple individuals potentially separated in time and place. Nonetheless, while its language is utterly non-classical, it appears consistently so, betraying no obvious linguistic or stylistic discrepancies from section to section.56 This relative stylistic uniformity might point toward production within a single textual community, our understanding of the geographic, social, and chronological parameters of which remains nebulous. Simplistic as it may seem to suggest that its primary sections represent consecutive strata in its development, this possibility appears to have merit, particularly in the case of the Addendum Tantras. As a working hypothesis, I would suggest that the core of the old text consists of much or most of BraY ixlix, to which, in the next stage, material from chapters llxxxiii a
For the sake of attaining the clan siddhis, [one should remain] standing before the goddesses, having reached a state of nonduality, without apprehension about the rites with [impure] gruel; by the mere consumption of this, one would obtain the siddhi of the ying yogins (khecar). .t Siddhayogevarmata 29.16d, and Srkan. hyasamhit 223, 245246 (numbering as per the manuscript s . a transcription in Jrgen Hanneder, Abhinavaguptas Philosophy of Revelation: An Edition and Annotated Translation of M linsloklav rttika I, 1-399); the latter two lacunose verses refer to a Vidy tilaka and a a a Bhairavatilaka, suggesting that tilakatantras comprised a scriptural genre, like the y malas or matas. a 55 See the discussion of B na in chapter 2, section 3. a. 56 On the language of the text, see the remarks in the next section, and particularly the annotation to the critical edition in part ii.

211 was incorporated. The nal stage of redaction is probably represented by the Uttaraand Uttarottaratantras, chapters lxxxivci.

4.3 on the provenance of the brahmayamala


Nothing I am presently aware of makes it possible to date the BraY with precision. A a variety of factors, however, points towards the sixth to eighth centuries as the period within it would most plausibly have been composed. Quotations of the BraY in the a Tantr loka of Abhinavagupta, who ourished in the decades before and after the turn a of the eleventh century, establish the existence of the text by this period, at the very latest. The distribution of Abhinavaguptas citations, which are drawn from most sections of the BraY , suggests that he knew the text in a form close to that preserved a in the Nepalese manuscripts.57 The oldest of these was in any case copied not long
57 Abhinavagupta cites the BraY fteen times in the Tantr loka; precise references are provided in a a the next chapter (nn. 12). In what follows, several of the passages he makes reference to are identied, instances spanning chapters iv, xxi, xliv, lv, and (probably) lxxxiv of the BraY : a

1. In Tantr loka 27.2123ab, Abhinavagupta paraphrases BraY iv.30815; Tantr loka 27.22ab is in a a a fact a direct quotation of iv.308ab. In the ksts edition, Tantr loka 27.22ab reads, ture yogah sad a a . sastah siddhido dosavarjite; yogah is however certainly a corruption of y gah. Codex Bya of the a . . . . BraY reads ture y gam sad sastam siddhid m dosavarjitah. The original text was perhaps ture a a . a a. . . . y gam sad sastam siddhidam dosavarjite, and Abhinavagupta thus appears to have corrected the a . a . . gender of y ga from neuter to masculine. In addition, Abhinavagupta refers in 27.29 to a typology a of ritual skulls, for which the source is BraY iv.74755from a section on the subject of the a characteristics of skulls (kap lalaksana). a . . 2. Tantr loka 4.55cd65 makes reference to both BraY xxi and lv, closely paraphrasing a passage a a from the latter (see below). The relevant passage from BraY xxi concerns the observance (vrata) a connected with the goddess Rakt , i.e. the rakt vrata. Abhinavagupta draws on this in advancing a a the idea of self-consecration ( tm bhiseka), closely paraphrasing BraY xxi.69cd70ab. Note the a a . a text of Tantr loka 4.63cd65: a tatraiva ca punah srmadrakt r dhanakarmani 63 aa . . vidhim proktam sad kurvan m sen c rya ucyate | a a a a . . paksena s dhako rdh rdh t putrakah samay tath 64 a a a . . a . dksayej japayogena rakt dev kram d yatah | . a a . guror al bhe proktasya vidhim etam sam caret 65 a a . And furthermore, in that very text, [the BraY ,] in the [section on the] ritual proa cedure for worship of Sr Rakt , this procedure [of self-consecration] is taught; by a practicing constantly, after a month one is called c rya, by a fortnight, s dhaka, a a a from a quarter [month], putraka, and likewise [in half that time] samayin. Since the goddess Rakt would bestow initiation, in due course, owing to [ones practice of] a mantra incantation and yoga, this is the procedure one should follow in the absence of the aforementioned guru.


This cites BraY xxi.69cd70: a dvibhih karmasamarthas tu m sen c rya ucyate 69 a a a . paksena s dhako hy esa bhutale mantravigrahah | . . a . . prathaman tu vratam hy etad rakt y h parikrtitam 70 a a. . By two [months of following the observance], he is capable of [any] action. Through one month, he is called c rya. By a fortnight, he becomes a s dhaka, having a body a a a of mantra on this [very] earth. This is known as the rst observance, of Rakt . a 3. Note BraY lv.26: a evam di-r-anekai ca prak rais tu mahevarah | a s a s . kurute nugraham pums m yasm sau sarvatomukhah a . . a. . 26

26a anekai ] Byb ; anekai Bya Byd 26c kurute ] Bya Byd ; kumrute Byb nugraham ] em.; nus s . . . grah m Bya Byb Byd a. pums m ] Byb ; puns m Bya Byd a. 26d yasm sau ] Bya ; yasy Byb ; a a . a. yasm (m?) sau Byd a .

And in these and many other ways, Mahe vara bestows grace upon souls, since he s is all-seeing. Abhinavagupta rewrites this as Tantr loka 4.56cd-57ab, correcting its irregularities: a evam dyair anekai ca prak raih paramevarah a s a . s . sams rino nugrhnati vivasya jagatah patih | s . a . . . . . 56

Note the rewriting of 26a to avoid internal-hiatus breaking -r-, the removal of the meaningless tu in 26b, and the total avoidance of 26d, with its non-standard yasm for yasm t. a a 4. Tantr loka 28 makes reference to a passage in BraY xliv, paraphrased in 28.383cd84ab: a a srmatpicumate coktam adau yatnena raksayet 383 . praveam sampravis. asya na vic ram tu k rayet | s a . a . .t And it is said in the revered Picumata that at rst, one must guard entry [to the ritual assembly] carefully. However, one should not deliberate over one who has been admitted. Compare with BraY xliv.228cd29ab: a a ad v eva na vai dady t praveam kasya cit priye 228 a s . pravis. ena sahaikatvam bhaksitavyan na samsayah | .t . . . . At the very rst, one should not admit just anyone, my dear. [But] with someone who has been admitted, one should feast together as one, without a doubt. 5. In Tantr loka 15, Abhinavagupta cites the authority of the BraY on the inseperability of internal a a and external worship, which he links to the dichotomy of gnosis (j na) and ritual action (kriy ). a a It seems possible that he had in mind a passage from BraY lxxxiv. Note Tantr loka 15.43cd44: a a n dhy tmena vin b hyam n dhy tmam b hyavarjitam a a a a a . a . a siddhyej j nakriy bhy m tad dvityam samprak sate | a a a. . a .


srbrahmay male deva iti tena nyarupayat 44 a Not without the spiritual (adhy tma) would the external succeed, nor the spiritual a devoid of the external; () the pair nds expression through gnosis and ritual action (?)with this [statement] the Lord has explained in the revered Brahmay mala. a tad dvityam (the second one [i.e. adhy tma?]) or taddvityam (having that as its second?) is . a . problematic; Isaacson suggests, as one possibility, reading tad dvitayam (the/that pair), which . is adopted in the translation above. (Personal communication, autumn 2003.) Compare with BraY lxxxiv.140: a

213 afterwards, in 1052 c.e.58 Considerable historical development separates the Saivism of Abhinavagupta from that of the BraY most notably the entire corpus of Kaula a scripture, which nds no place in the BraY s account of the Saiva canonbut this a separation is difcult to quantify. Several sources of evidence nonetheless suggest with a high degree of probability that the BraY existed two centuries or more prior a to Abhinavagupta. By the mid-tenth century are attested works of tantric literature that place themselves within the tradition of the BraY , illustrating that it was by this time considered a a an important authority. A Nepalese manuscript of the Bhairavamangal , a scriptural text ascribing itself to the tradition of the BraY , appears on paleographic grounds a a to date to this period.59 Furthermore, the Pingal mata, a scripture of the pratis. h .t a tantra genre that places itself in the tradition of the BraY ,60 appears to have been a
adhy tma cintayed b hyam b hyam adhy tmik m tath | a a a a. a . a cakre sam nabh vena tato viny sam arabhet 140 a a a One should meditate upon the internal [cakra] as external, and the external likewise as internal. Considering [these] to be the same, one should [only] then commence installation [of the deities] on the cakra. In the BraY , the pair adhy tma and b hya simply refer to yogic processes and external ritual a a a performance, respectively. In this case the correspondence with the Tantr loka citation is only a suggestive, and Abhinavagupta might have had in mind other passages of the BraY . Note for a instance lxxxvii.101 (also from the BraY s Uttaratantra). Here too, the text does not link j na a a and kriy to the adhy tma-b hya dichotomy in the manner of Abhinavagupta: a a a anena vidhin devi japahom dikarmasu | a a b hy dhy tmeva mantrajah puj m kurvan prasidhyati a a a . a.
101b karmasu ] corr.; karmasu . Bya


a. a. 101d puj m ] em.; pujy m Bya

Through this procedure, O goddess, in mantra incantation, re sacrice, and other rites, the knower of mantras achieves siddhi, practicing both external and internal worship [b hy dhy tmeva=b hy dhy tm m eva ?]. a a a a a a a In the future, I intend to publish a more comprehensive discussion of Abhinavaguptas citations of the BraY . a 58 See chapter 1, section 1. 59 a Bhairavamangal , nak 5-687 (ngmpp reel b27/21); regarding the dating of this manuscript, I am grateful for the learned opinion of Diwakar Acharya (personal communication, January, 2007). I quote a the Bhairavamangal s references to the BraY in chapter 5 of this thesis (nn. 68, 1034). This text is a .t potentially identical to the Bhairavmangal listed in the Srkan. hyasamhit as rst of the mangalatantras, a . a .t described as picutantrasamudbhav (arising from the Picumata [i.e. Brahmay mala]). Srkan. hyasamhit a a . a 276 (in the numbering of Hanneder, Abhinavaguptas Philosophy of Revelation). 60 a Note for instance the colophon of the (incomplete) Vi vabh rat codex of the Pingal mata (f. 11r): iti s a a brahmay male jayadrath dhik re pingal mate pratim dhik ro n ma prathamaprakarane [em.; prakarana cod.] a a a a a a . . s a caturthah (Thus ends the fourth [chapter], entitled the Section on Images, in Book One of the Pingal . mata, in the Jayadratha-[y mala?] Section of the Brahmay mala). a a

214 commented upon by an important mid tenth-century Saiddh ntika exegete, Bhata . ta N r yanakantha.61 Finally, the rst of the Jayadrathay malas four books gives aa . a . .. much importance to the BraY in its description of the canon of Saiva scripture, even a containing a chapter entitled brahmay malanirnaya, A Denitive Judgment on the a . Brahmay mala. The dating of this heterogenous work is problematic; portions seem a likely to be quite early, while its nal form might postdate Abhinavagupta.62 It is unclear how much prior to the mid-tenth century these works existed; one or more might belong to the ninth century, or even earlier. In any case, their attestation in the mid-tenth century suggests that the BraY probably existed by the end of a the ninth century. Its terminus ante quem can be pushed back somewhat further, however. As discussed in the previous chapter, there is strong evidence that the Buddhist Laghuamvaratantra incorporates a large tract of text from the BraY . Should s . a it prove correct that Vil savajra quotes the Laghuamvara, the latter must have existed a s . by the late eighth century; in any case, the Laghuamvara appears to have received a s . commentary in the mid-ninth century.63 Signicantly, the section incorporated into the Laghuamvara belongs to the Uttaratantrain all probability a late stratum of the s . BraY suggesting that the BraY existed in a redaction close to its extant form by a a the mid-eighth or early ninth century, depending upon the date of the Laghuamvara. s . Besides the aforementioned cases, which pertain with little ambiguity to the extant BraY , there are more nebulous early references to a text by this title. As a discussed in chapter two, the old Skandapur na provides a list of Saiva m trtantras, a. a. Tantras of the Mother Goddesses, that includes the BraY (br hmam y malam). Transa a . a mitted in a manuscript dated 810 c.e., it is the working hypothesis of its editors that the Skandapur na took shape in the sixth or early seventh century.64 It remains posa. sible that the Skandapur na chapter referring to the BraY is a comparatively late a. a
Alexis Sanderson, The Saiva Religion among the Khmers (Part I), 441. See section 3 of chapter 3 in the present thesis. The Jayadrathay malas brahmay malanirnaya chapter, a a . mentioned in the previous section, is fortieth in the rst satka. . . 63 See the discussion of the Laghuamvara in section 4 of the previous chapter. s . 64 See the discussion of the Skandapur na in chapter 2, section 2. a.
61 62

215 addition to the text, while it is also conceivable that the text it refers differs from the extant BraY . Nonetheless, the probability seems high that this passage intimates the a existence of some form of the extant BraY in the eighth century, if not considerably a earlier. The passage contains clear reference to the Yogin cult and is to this minimal extent compatible with the extant BraY . As for other early references to a Brahmaa y mala, a hymn entitled Bhairavvardham naka, of which several manuscript folios a a appear to date to the early ninth century, refers to the Goddess as Brahmay mal , a a apotheosis of the scripture by this name.65 The grounds for establishing the BraY s terminus post quem appear tentative. a While it is difcult to imagine that the BraY existed as we have it in the sixth century, a this cannot be ruled out entirely: the old Skandapur na, the inscription of Gangdh r, a. a Dharmakrtis reference to dakintantras, and allusion to tantric goddess worship in . early seventh-century literary sources leave open the possibility that a tantric Saiva cult of yogins, and perhaps a Brahmay mala, existed in this period. It is not until a the early eighth-century M latm dhava of Bhavabhuti, however, that a yogin cult of a a the type described in Vidy ptha tantras nds detailed attestation in reliably dated a . sources. Buddhist yogintantras come into evidence only from around the mid-eighth century, as discussed previously; and the Laghuamvaracomparatively early in this s . corpusappears to draw not only upon the BraY , but on several Saiva scriptures of a the Vidy ptha.66 How much earlier the latter sources might have existed is unclear, a . however. A vast text, the BraY provides substantial material of potential value for situating a it in chronological relation to the extant Saiva literature. Being an inuential early scripture, the BraY is mentioned by name in a variety of sources, while its indirect a
Sanderson remarks that the Bhairavvardham naka, the hymn to the Goddess of which some folios a are preserved with the P ramevara codex of a.d. 827/28 . . . knows a Brahmay mala, Visnuy mala, and a s a .. a Rudray mala, since it refers to the Goddess as the embodiment of these (f. 53r1): tvam brahmay mal a a a . tvam visnuy mal tvam rudray mal . History through Textual Criticism, 19 (n. 21). a a a . .. a . 66 Sanderson identies borrowings in the Laghuamvara from the Siddhayogevarmata, BraY , Tantras . s a sadbh va, and the Yoginsac ra of the Jayadrathay mala, which appear to comprise the oldest major extant a a a scriptures of the Vidy ptha. See History through Textual Criticism, 4147. a .

216 inuence can also be identied in some cases. In addition, early though it may be, the BraY describes a vast and diverse scriptural canon, providing a detailed picture a of the forms of Saivism it claims to transcend. Internal evidence from the domains of ritual, doctrine, cosmology and so forth also provide indications of the texts relative archaism, although considerable work remains to be done in these areas. BraY xxxviiithe srotanirnayapatala,67 Chapter on the Streams [of Revelation] a . . maps out the canon of Saivism, providing a valuable catalog of tantric Saiva scriptures that it classies according to three primary streams (srota[s]): those of the v matantras, siddh ntatantras, and bhairavatantras (tables 4.24). Both of the latter are a a twofold; the siddh ntatantras include siva-division (bheda) and rudra-division scripa tures, while the bhairavatantras are divided into Mantraptha and Vidy ptha texts.68 . a . The BraY therefore presupposes varieties of tantric Saiva literature that appear to a have existed by the early seventh century.69 In addition, reference is made to Vais. nava tantras of the P car tra, the titles of which appear at least partly spurious a a . (table 4.3).70 Although the extant literature of the Vaisnava P car tra does not a a .. seem especially ancient, the P car tra tradition itself is mentioned in the Mah a a a bh rata, presumably in a pre-tantric variety; it does not seem possible at present to a determine the antiquity of its earliest tantric literature, making the relevance of its mention in the BraY unclear.71 Absent from the BraY s account of the Saiva canon a a are Buddhist tantras and Saiva scriptures likely to belong to Kaula traditions. These absences need not be read in strictly historical terms; yet the latter in particular would
67 In the vast majority of its occurences, srotas (stream, current) is thematized as an a-stem (srota) in the BraY ; srotra is also very common, although this is potentially a scribal corruption. a 68 On the BraY s conception of the Saiva canon, see chapter 5 of the present thesis, passim. Tables a 4.24 provide lists of the texts mentioned in BraY xxxviii. a 69 As reviewed previously, the early seventh-century Buddhist author Dharmakrti appears to refer to v matantras. Early-seventh-century inscriptions make reference to the initiation of kings into mandalas a .. apparently of the Saivasiddh nta, the tradition for which siddh ntatantras are the scriptural authorities. a a Sanderson, History through Textual Criticism, 810 (n. 6). 70 That the titles given for Vaisnava tantras are spurious is suggested by their generic naturee.g. .. Pacar travidh na (Tantra of the Pacar tra Rites) and Vaikun. havidhi (Tantra of the Rites of Visnus a a a .t .. Heaven)as well as their apparent lack of attestation. 71 On some of the problems of dating P car trika literature, see Sanderson, History through Texa a tual Criticism, 35, 38 (n. 50).

217 be difcult to reconcile with a ninth- or tenth-century period of composition, when Kaula traditions were prevalent.72 Additionally, the BraY s very model of the scripa tural canon seems archaic, for it displays no awareness of the ve-stream model of Saiva revelation that became normative.73 Table 4.2: Siddh ntatantras of the Middle Stream (madhyamasrotas) according to BraY xxxviii a a Sivabheda Sarvak mika a Yogodbhava Acintya K rana a . Ajita Dpt bha a . Suksma S hasra a Amsum [n] a . Suprabha Rudrabheda (list 1) Vijaya Ni v sa s a Sv yambhuva a . V thula a Candrah sa a Raurava M kuta a . Vre a s Agneya (conj.; tath c nyam cod.) a a . Candraj na a Bimba Prodgta Lalita Siddhi Sant naka a Sarvodgta Kirana . P rame vara a s Rudrabheda (list 2) Vijaya Ni v sa (em.; visv sa cod.) s a a Sv yambhuva a . Vrabhadra Agneya Raurava M kotya a . Candrabh sa a vramvaund J na a Mukhabimba Prodgta Lalita Siddhi Sant na a Kirana . Sarvodgta P rame vara a s

Little of the extant Saiva literature can be said on strong grounds to predate the BraY . Nonetheless, several siddh ntatantras that survive or are quoted in early a a
72 Incidentally, while the BraY does not refer to Buddhism, the Matas raa Kaula text placing itself a a in the tradition of the BraY in its fourth chapter makes clear reference to initiation into Buddhist a tantras:

siddh nte dksit ye ca vaisnave ye ca dksit | a . a . a .. pacasrotodbhave tantre sammohe caiva dksit | . a . saure ca g rude devi bauddhe ye caiva dksit h | a . . a ath pi p savah sarve asmin tantre na dksit | a a . . a
asmin ] em.; yasmin cod. tantre na ] em.; tantrena cod. .

Those initiated into the siddh nta-, and those initiated into the Vaisnava, Saura, G ruda, a a . .. or Buddhist [tantras]the tantra[s] arising from the Five Streamsand those initiated into the [v ma] Sammoha, all of them are still bound souls, [for they are] not initiated into this a . tantra. nak 3-379, f. 37, lines 12; transcription courtesy of Somadeva Vasudeva. 73 See chapter 5, section 4, and tables 4.24, 5.2.


Table 4.3: Tantras of the Left Stream (v masrotas) and Lower Stream (adhahsrotas) a . northern stream Sammoha . bhav a Nayottara Saukra lower stream Vibhuti Adhyaya M kuta a . subdivisions of the . makutatantra: N rasimhavidh na a a . vidh nam ksetrakalpan (?) a . . a vivar ntargatakalpas a Var havidhi (em.; yar hasya vidhi cod.) a a Pacar travidh na a a Vaikunthavidhi .. Garudavidh na a . bhutantr di a osadhikalpas . ras yanavidhi[s] a

Table 4.4: Bhairavatantras of the Right Stream (daksinasrotas) . . i. vidyaptha eight bhairavas: Svacchanda Krodha Unmatta Ugra Kap lin a a Jhank ra Sekhara Vijaya eight yamalas: Rudray mala a Skanday mala a Brahmay mala a Visnuy mala a .. Yamay mala a V yuy mala a a Kuberay mala a Indray mala a other: Prapacayoginj la a Yoginj la a Yoginhrdaya . Siddh a Mantram lin a Aghore s Aghore var s Krdaghore . s L kinkalpa a a M r Mah m r a a Ugravidy gana a . Bahurupa (twofold) Aghor stra a mantraptha i. Vrabhairava Candabhairava .. Gudik bhairava (conj.; . a Mah vre abhairava a s
gudak bhairava cod.) . a

219 sources are listed in BraY xxxviii: the Niv sa, Kirana, P ramevara, Raurava[sutrasaa s a a s . ngraha], and Sv yambhuva[sutrasangraha].74 In only one case thus far identied does a . a parallel passage shed further light on relative chronology: it would appear that the BraY has incorporated a passage from the Niv satattvasamhit , whether directly a s a . a or through an intermediary source. The relevant text is Niv sottara 1.1013, and s a Table 4.5: A parallel passage in the Brahmay mala and Niv satattvasamhit a s a . a Ni vasatantra, Uttarasutra 1.1013 s svara uv ca | a arah prakrtir jey uk . a . yak rah purusah smrtah | a . . . . . vak ro niyatir vidy l a a lak rah k la ucyate 10 a . a m y tattvam mak ras tu a a a . ksak ro vidya eva tu | a . rak ra svaro jeyo a hak ras tu sad sivah 11 a a . dehavy p ca navamo a sakti ca da am smrt | s s a .a akathya ca arup ca s k rana sa sivah parah 12 a . . . ete tattv h sam khy t h a. a a a. sad sivasamudbhav h | a a. r eva jagat sarvam . pre y pre ya car caram 13 s a s a Brahmayamala lxi.272cd75ab a . uk rah prakrtir jey a . yak rah purusah smrtah 272 a . . . . . vak ro niyatim vindy l a a . lak rah k la ucyate | a . a m y tattvam mak ram tu a a a . . ksak ro vidya eva ca 273 a . repham sam iti prokto hak ras tu sad sivah | a a . dehavy p tu navamah a . k ranah parame varah 274 a . . s . etat tattve varam devam s . . tritattvoktam prakalpayet | .
272c uk rah ] Byapc ; uk ra Byaac a . a prakrtir ] em.; prakrti Bya . . 273a vak ro ] em.; vak ra Bya a a vindy l ] em.; vindy Bya a a 273c tattvam ] em.; tatva Bya . 274b hak ras ] em.; hak ros Bya a a 274c navamah ] em.; navamo Bya . 275a tattve varamem.; conj.; tatte varam Bya s s . . 275b tattvoktam ] em.; tattvokt Bya a .

BraY lxi.272cd75 (table 4.575 ), the latter belonging to the BraY s sangrahasutra a a chapter referred to earlier. This passage places the nine syllables of a mantra called The Ninefold (nav tman) in correlation to a hierarchy of reality levels (tattva)a a nine-tattva series that appears characteristic of the Niv sa corpus (table 4.6a).76 In s a
On the evidence for the antiquity of these ve siddh ntatantras, see Goodall, Bhat. a R makan. has a .t a .t Commentary, xxxvixlvii. 75 The text tabulated from the Niv sottara is as given in the provisional edition circulated among s a participants of the Workshop on Early Saivism (Pondicherry, January 2007); see the discussion of the Niv sa in chapter 3, section 2. s a 76 This series of nine tattvas is, for instance, presented in relation to the nine constituents of the letter

220 the BraY , however, this particular nine-tattva series (table 4.7a) is unusual; svaraa tattva, the seventh of the series, gures only rarely in the BraY s accounts of the a pervasion (vy pti) or purication (odhana) of the tattvas. In this position, the a s BraY normally places the saktitattva, between the vidy tattva and sad siva (tables a a a 4.7cf). The particular nine-tattva series correlated with the nav tman mantra occurs a elsewhere only in BraY lv (table 4.7b). More signicantly, the Niv sas nav tman a s a a mantra is wholly anomalous in the BraY , despite the latters afnity for all things a ninefold. This combination of factors suggests that the BraY has assimilated material a concerning the nav tman mantra from another source, the obvious candidate being a the Niv sas Uttarasutra. The date of the latter could thus provide the terminus post s a quem for the composition of the BraY . This reveals relatively little about the period a of BraY s composition, unfortunately, for some sections of the Niv sa could date a s a even to the fth century c.e.77 a period improbably early for the BraY . a Table 4.6: The Nav tman Mantra in the Niv sottara and the Vidy r ja of the Svacchandatantra a s a aa
(a) The tattvas of the nav tman a mantra, according to the Niv sottara s a (b) Purication of the Thirty-six and Nine Tattvas with the syllables of the King of Vidy a mantras (vidy r ja), according to Svacchandaaa tantra 5.411

tattva paraiva s sakti dehavy pin a sad siva a svara vidy a m y a a k la a niyati purusa . prakrti .

aksara .

36 tattvas

aksara .

9 tattvas

h? . ha ra ksa . ma la va ya u

sakti sad siva a svara vidy a m y a a k la, kal a a niyati, vidy a purusa, r ga a . ksitiprakrti . .

pranava? . ha ra ksa . ma la va ya u

siva sad siva a svara vidy a m y a a k la a niyati purusa . prakrti .

While it is hence likely that at least one of the earliest siddh ntatantras predates the a
a, and nine parts of the body, in the Niv sas Nayasutra (1.120). s a 77 See chapter 3, n. 3.


Table 4.7: The Nine Tattvas in the BraY a

(a) The Lord of Nine Tattvas (navatattvevara) s according to BraY lxi a (b) Nine goddess clans (kula) and corresponding tattvas in BraY lv a

tattva paramevara s sad siva a svara vidy a m y a a k la a niyati purusa . prakrti .

aksara . dehavy pin a ha ra ksa . ma la va ya u

tattva siva sad siva a svara vidy a m y a a k la a vidy a purusa . prakrti .

kula devs bhagins siv s a rudradakins? . dakins? . duts yogins m trs a.

(c) Pervasion of the tattvas by the Nine Saktis according to BraY xxix.22528 a

(d) Pervasion of the tattvas by the Nine Saktis according to BraY xxix.22934 a

tattva paraiva s sad siva a candra aditya k la a prthv . apas tejas v yu a

mantradevata Sad siva a Mahocchusm (pa) . a . Candaks (ka) .. Kar lin (de) a . Rakt (ca) a Kar l (li) aa Dantur (ni) a Bhmavaktr (sva) a Mah bal (ha) a a

tattva sad siva a sakti vidy a m y a a k la a kal a niyati purusa . prakrti .

mantradevata Mahocchusm (pa) . a . Candaks (ka) .. Kar lin (de) a . Rakt (ca) a astras Kar l (li) aa Dantur (ni) a Bhmavaktr (sva) a Mah bal (ha) a a

(e) Purication of the tattvas according to BraY xxxii a

(f) Purication of the Nine Pantheons according to BraY xxxvi (tattvadksa) a .

tattva siva sad siva a sakti vidy a m y a a k la a niyati purusa . prakrti .

mantradevata Bhairava Mahocchusm . a Candaks . . . Kar l aa Rakt a astras Duts Yogins M trs a.

tattva siva sad siva a sakti svara vidy a m y a a k la a niyati purusa . prakrti .

mantradevata Siva Bhairava-Bhairav M he var a s Brahm n a. Vaisnav .. Kaum r a Vivasvat M hendr a C munda a .. Par Sakti a

222 BraY , the chronology of the BraY relative to other early Saiva literature is elusive. a a As mentioned in the previous chapter, the BraY refers by name to several v maa a tantras, including the extant Vnasikhaa text not among the earliest of its genre, . yet nonetheless potentially quite old. There are, furthermore, faint indications of inuence from the cult of the Four Sisters upon aspects of the BraY .78 As for the a relationship between the BraY and the Svacchandatantra, the evidence I am currently a aware of is not especially strong. The former does list the latter in its account of the canon; but the evidence from texts lists must be treated with caution, as the case of the Siddhayogevarmata and BraY illustrates (discussed below). On the other s a hand, preliminary analysis of cosmological materials suggests that the BraY could a be archaic in comparison to the Svacchandatantra. For instance, like the BraY , the a Svacchandatantra draws upon the Niv sas nav tman mantra, placing a series of nine s a a tattvas in relation to the syllables of the nav tman. However, the Svacchandatantra a version also correlates the series to the thirty-six tattva system normative in later Saivism (table 4.6b)a system absent from both the Niv sa and BraY .79 s a a There are possible grounds for considering the BraY s principal male deity, a Kap lsabhairava, secondary in the historical development of Saivism to Svacchandaa bhairava. Kap lsa is rst attested as an important rudra in the Niv saguhya, heading a s a the Hundred Rudras at the level of the re of time (k l gni), at the base of the hiaa erarchy of world levels (bhuvana).80 Kap lsa the rudra is presumably an early form a of the deity who gures as rst of the eight bhairavas in the mandala of Svaccha.. nda, according to the Svacchandatantra.81 He attains the apex of his cultic status as supreme Bhairava of the BraY , alongside Aghorewho is the goddess consort of a s
In the previous chapter, see n. 39 in the discussion of v matantras. a Note that although the BraY does not attest the thirty-six tattva series that becomes standard in a Saiva exegetical literature, all of the tattvas included in this schema do nd mention at one point or another in the text, in its various non-standardized tattva series. 80 The list of the Hundred Rudras (atarudra) begins, satarudr ni me srnu | kap lso hy ajo buddhah s a. . . a . vajradehah pramardanah (Niv saguhya 7.82bcd). s a . . 81 In the Svacchandatantra, the eight bhairavas (bhairav s. aka) forming the primary entourage of Svaca. t chanda are headed by Kap lsabhairava; the remaining seven are Sikhiv hana, Krodhar ja, Vikar la, a a a a Manmatha, Meghan da, Somar ja, and Vidy r ja. The names and mantras of the Eight are given in a a aa Svacchandatantra 1.76cd86; cf. 2.11722.
79 78

223 Svacchandabhairava as wellin the form of Canda K p lin. The BraY s convention a . . a a for naming a new ofciant suggests that Kap lsa might have usurped the position a of Svacchandabhairava: during consecration, when the ower cast by the candidate falls upon the central god, he receives the name Svacchandabhairava, rather than being named after the BraY s own Kap lsaan apparent carryover from the cult of a a Svacchanda.82 Within the Vidy ptha, the relation between the BraY and the Trika Siddhaa . a yogevarmata remains an open question.83 Possessing distinct pantheons, these texts s nonetheless share much in the domain of ritual, both being decidedly k p lika and a a siddhi-oriented scriptures instrinsically connected with the cult of yogins in what appears to be an archaic form. The Siddhayogevarmatas account of the Saiva canon s lists the Brahmay mala.84 However, as Judit Trzsk points out, the BraY might also a a refer to the Siddhayogevarmata, for it mentions a Vidy ptha text by the title Siddh s a . a one of several names by which Abhinavagupta cites the Siddhayogevarmata.85 This s circularity, which owes perhaps to ongoing revision or idealized text lists, suggests the need for weighing multiple types of evidence in determining relative chronology, evidence that appears lacking in this case. The BraY does, however, contain material a concerning the Three Saktis (V m , Jyesth , and Raudr) that might potentially shed a a .. a light on the cultic background of the Siddhayogevarmatas triad (trika) of goddesses.86 s

BraY xxxiii.16566: a

bhairave tu yad puspam patate purvvacoditam | a . . . saktn m tu tad tasya n mam vai kalpayed budhah 165 a. a a . svacchandabhairavo n ma tad tasya praj yate | a a a bhairavy m tu yad p ta saktibhairavasamjakah 166 a. a a . . There would seem to be a textual problem here, for 165cd appears misplaced (perhaps it followed 166cd?), or even interpolated; 166ab seems to intended to follow 165ab. In BraY iv, in the section explicating initiatory kinship based upon the ower-cast of the initiand, a the applicable clans are those of the Eight Mothers and Bhairava, of whom no particular form is specied. This passage is quoted in part in chapter 2 (n. 14). 83 As mentioned in the previous chapter, the Siddhayogevarmata survives only in a short recension s missing a signicant amount of the material attributed to it by Kashmiri authors. See Trzsk, Doctrine of Magic Female Spirits, ivv. 84 Siddhayogevarmata 29.18a. s 85 This is pointed out by Trzsk, ibid., ix (n. 42). 86 Worship of the Three Saktis comprises the subject of lxxxiii (the saktitrayavidh napatala). Potentially a . relevant material is found in BraY lxxx as well. In this fascinating chapters discussion of the origins a

224 More can be said concerning the relation of the BraY with another major Vidy a a ptha scripture: the Tantrasadbh va of the Trika, a text which explicitly situates itself . a in the tradition of the Siddhayogevarmata.87 There are several indications that the s Tantrasadbh va postdates the BraY as well. While the Tantrasadbh va makes no mena a a tion of a Brahmay mala, it does refer to the y malatantras as a genre.88 Given that the a a
of the skull-staff (khatv ngotpatti), the various elements of the archetypal skull-staff of Bhairava are . a described in homological relation (adhidaiva) to a hierarchy of divinities and cosmological spheres. In particular, the three prongs of the trident (triula) that caps the skull-staff are said to be presided over s by the Three Saktis. BraY lxxx.17273ab: a yat triuloparis. at tu tatra saktitrayam viduh | s . t . . v m jyes. h ca raudr ca v m daksinamadhyag h 172 a a .t a a a . . a. uparis. at sivam santam sthitam sarvasya murdhani | . t . . .
172a oparistat ] em.; oparista Bya 172c jyesth ] corr.; jesth Bya .. .. .. a .. a g h ] em.; madhyag m Bya 173a uparistat ] em.; uparista Bya a. a. .. .. 172d

There, above the trident, are known to be the Three SaktisV m , Jyesth , and Raudr, a a .. a on the left, right, and middle. Situated above, overhead all, is the quiescent Siva. Visualization of the triad of goddessesPar , Par par , and Apar upon the prongs of a trident a a a a features in the initiation mandala of the Siddhayogevarmata, a fundamental work of the Trika; see s .. Sanderson, The Visualization of the Deities of the Trika, 39. It is conceivable that the material cited above from the BraY provides an early precedent for this aspect of the Trika, the goddesses of which a might have had their identities grafted upon those of the Three Saktis of early SaivismV m , Jyesth , a a .. a and Raudr. Note that BraY xxvii, the saktitritayay gapatala, concerns a different triad of saktis: that of V m , a a a a . Madhyam , and Daksina, who in BraY xxxviii are described as presiding over the three primary a a . . streams of scriptural revelation and their practitioners. 87 See chapter 5 in the present thesis, n. 100. 88 Tantrasadbh va 1.1112, 34cd35ab: a srutv vedam may tantr rahasyam gudhagocar h | a a a a. . . . v madaksinam rg s ca y mal s tu anekadh 11 a a a a . . a a siddh nt s ca sureana daas. adaabhedatah | a a s s . t s . laksakotivibh gena kotayas tu anekadh 12 a a . . . siddhayogevartantre satakotipravistare | s . mulatantre mah sutre sutradvayavinirgatam 13 a tantraikam tu mah j nam yonyarnavasamudbhavam a a . . . :::::::::: na srutam srotum icch mi tantrasadbh vam uttamam 14 a a . ... ye may kathit s tantr v madaksinay mal h 34 a a a a a. . . a rudrabhed s tath devi sivabhed s tathaiva ca | a a a
11a vedam ] Tsk Tskh ; veda Tsg tantr ] Tskh ; rudr Tsk ; - dr Tsg ; rudra em. Dyczkowski a a a 11b . ] Tsk Tskh ; - - - Tsg 11c v ma ] Tsk Tskh ; - - Tsg a 13 sutradvayavinirgatam ] Tsk Tskh ; . ? Tsg mah j nam ] Tskh ; may j tam Tsk Tsg a a . a a . 14b yonyarnava ] em.; yony rnava a . . k Tskh Tsg samudbhavam ] Tsk Tsg ; - - - - Tskh Ts 34c kathit s ] em.; kathit m Tsk ; kathit a a. a Tskh Tsg
gocar h a ::::::::::::::::: :::::::::

mss as reported by Dyczkowski; the edition proposed above is highly provisional. Cf. Tantrasadbh va a 25.301: s m nyam sarvatantresu v madaksinay male | a a . . a . . a a. vidy p. he pi devei brhadgranthe tu sucit h 301 a t s .

225 extant lists of y malas almost all feature the BraY , this alone might suggest that the a a latter was known in some form to the redactors of the Tantrasadbh va. Furthermore, a the Tantrasadbh va once makes reference to a text called the Dv daas hasra ([Tantra] a a s a of Twelve-thousand [Verses]), an epithet of the BraY .89 So far I have identied only a three verses shared by the two texts, with no obvious indications of the direction of redaction. In the rst case, the context is that of the initiatory Pledges (samaya), which one might expect any number of scriptures to shareand indeed the Siddhayogevars mata contains the same verse.90 Similarly, the texts share a verse listing eight sacred sites (p. ha).91 In the third case, which concerns chomm , a garbled Tantrasadbh va t a a verse would be impossible to reconstruct but for its parallel in the BraY .92 None of a
89 Tantrasadbh va 1.234a. This passage comes at the end of a list of deities, which it links to a text a called the Dv daas hasra: a s a

anena kramayogena pac san m samudbhavah 232 a a . . . kathit devadevena k ryak ranabhedatah | a a a . . evam dv daas hasre prthagdh raprakrtitam | a s a a . . a n m ni rudrasamghasya sucit nha p rvati 234 a a a . Text as consituted in the draft edition of Dyczkowski. 90 BraY lxi.121: a na nagn m vanit m payen na c pi prakatastanm | a. a. s a . n lokayet paukrdam ksudrakarman na k rayet 121 a s . . . a
pa yen ] em.; pasye Bya s
krdam a . . ] em.; krda By .

This is identical to Tantrasadbh va 9.543: a na nagn m vanit m payen na c pi prakatastanm | a. a. s a . n lokayet paukrdam ksudrakarmam na k rayet 543 a s . . . a .

nagn m ] Tsk ; nagl m Tskh Tsg prakatastanm ] Tsk Tskh ; prakatasanm Tsg krdam ] em.; krdam a. a. . . . . . . . Tsk Tsg ; krda Tskh ksudra ] Tsk Tsg ; ksudre Tskh . . . (mss as reported by Dyczkowski.) This verse also occurs as Siddhayoge varmata 6.47cd48ab. s
91 92

See chapter 2, section 2 (n. 88). BraY lv.1034: a potangety abhiv danam pratipotange pratyabhiv danam | a a . yoginn m tu vr nam n rsety abhiv danam | a. a . . a . a pratin rsaabdena procyate prativ danam 103 a . s a ek ngulidaran t sv gatam dv bhy m susv gatam | a s a a a a. a . kos. hapravis. en ngus. hena ksemamudr vidhyate 104 t t a .t a . . .

103a potangety ] Bya Byc ; potange(tt?)y Byb 103b pratipotange ] Bya Byb ; pratipotanga Byc pratyabhiv danam ] Byb ; praty bhiv danam Bya Byb Byc 103d n rsety ] Bya Byb ; n risebhy Byc 104c a a a a . a . . kostha ] Byb ; kosta Bya .. .. Notes: In Byb , 104ab is missing, while 104cd is in the lower margin, possibly by original scribe. Byc skips from 103c (pratin ( a?). . . ) to 105c (. . . smrse p dam). as a . .

Cf. Tantrasadbh vatantra 18.18cd19: a pottangety abhiv danam pratyottunge pratikrtam 18 a . . ek mgulidaran t sv gatam dv bhy m susv gatam | a. s a a a a. a .

226 these isolated cases seems to warrant the assumption of direct borrowing. More substantially, the twenty-rst chapter of the Tantrasadbh va includes in its a treasury of vidy -mantras (vidy koa) the following: om camunde kapalini svaha, a a s . .. described as the root mantra (mulamantra) of Aghore. This is remarkably similar s to the Nine Syllable Vidy (nav ksar vidy ) of the same Goddess in the BraY : om a a . a a a . . hum cande kapalini svaha. The Tantrasadbh va describes several inected forms a .. of Aghores vidy utilizing the vocative cande, rather than c munde, echoing still s a a .. .. more closely the BraY s nine-syllable vidy .93 It seems likely that the Tantrasadbh va a a a draws on the tradition of the BraY with these mantras, although this cannot be a stated categorically. Even more suggestive of a direct link between the texts is Tantrasadbh va 19. This chapters rst section concerns sacred topography, mapping out a a network of eight elds (ksetra), secondary elds (upaksetra), and meeting points . . (sandoha), for which presiding goddesses, male eld guardians (ksetrap la), and a . sacred trees are enumerated. This systems eight ksetras map precisely to the eight . sacred cremation grounds enumerated in BraY lxxxiv (table 4.10).94 In addtion, a
kos. hapratibimb c ngus. [h]am ksemamudr vidhyate 19 a a .t a .t . . nak 5-445 (ngmpp reel a44/2), f. 122r. It seems likely that the text of Brahmay mala lv.103ab and 104 a underlies this, with the signicant absence of lv.103cdef. Among the sources at my disposal presently, the text of 103ab is found elsewhere only in the Laghuamvara, in which potangin, potang, and pratipotas . ng occur as mudr s in 22.5, and in a closely related chomm passage of 24: potangyabhiv danam pratipotaa a a . ng pratyabhiv danam (printed as prose in the Sarnath edition, v. 1, p. 126). On the latter Bhavabhatta a .. rather unhelpfully comments, potangy abhiv danam iti | abhiv danasya samj potangsabdah | pratipotang pratyabhiv danam a a a . a . iti pratyabhiv danam punar vandan viesah | a a s . . . . . . The word potang is a convention for greetings; . . . pratyabhiv danam is again a partica ular type of paying respects. Ibid. (text repunctuated). 93 For instance, the hrdaya or heart mantra is described as having the t raka (savior) in the bea . gining, then the word cande, ending with n da (=namah?) (candeti t rak dyam ca n d ntam hrdayam a a a a a .. . .. . . . . . param, Tantrasadbh va 21.156cd). That the t raka is hum is suggested in BraY xxiii, which describes the a a a t raka as the Root Mantra (mulamantra) of Bhairava, when conjoined with its ancillaries (t rakam yan a a . may proktam vaktranetr ngasamyutam | bhairavasya sam khy tam mulamantram na samsayah, 54). a a a a . . . . . . 94 Tantrasadbh va 19.4cd5ab: a pray g varuna koll at. ah s jayantik 4 a a a .t a a a . caritraik mraka caiva kotivarsam tu c s. amam | a s a. t . . . aindrdii s m rabhya y vad sanagocaram 5 s a a a j tavyam ksetraviny sam mantrina tu yath kramam | a a . a . . .
aik mraka a s

] em.; ek mraka mss a s

6b mantrina ] em.; mantrinam mss . . .

(mss as reported by Dyczkowski.) On the sacred geography of the BraY , see the next section of this a

227 the Tantrasadbh va organizes the goddesses of these and the subsidiary sacred places a into clans (kula) presided over by the Four Devs and Four Duts of the BraY s core a pantheondeities who have no cultic signicance in the Tantrasadbh va. For instance, a belonging to the clan of Mah bal , fourth of the Duts in the BraY , are the goddess a a a Karnamot, who presides over the eighth ksetra, Kotivarsa; Cipitan s , of the eighth . . . . a a . . upaksetra, R jagrha; and Carmamunda at Pundravardhana, eighth of the sandohas. a . . . .. . On these grounds it appears probable that the Tantrasadbh va postdates the BraY . a a The dating of the Tantrasadbh va is itself problematic; according to Sanderson, it is a one of several Vidy ptha tantras, including the BraY , drawn upon by the Buddhist a . a Laghuamvara.95 This could suggest that a signicant interval of time separates the s . BraY and the Laghuamvara, with the Tantrasadbh va belonging to the intervening a s . a period. The BraY contains a single potential reference to a work of non-tantric Sanskrit a literature. A verse in the revelation narrative of chapter one predicts that the BraY a shall become as famous as the Saptaat ni (The [Text of] Seven Hundred [Verses]).96 s a The possibility seems signicant that the text alluded to is none other than the Durg saptaat (Seven-hundred Verses on Durg )the Devm h tmya of the M rkaa s a a a a ndeyapur na. A work exceptionally well-known in latter medieval India, it appears a. .. tting that the Devm h tmya would be singled out as emblematic of popularity a a although it is unclear how early it attained this status. There is little to base this identication upon, however; all other texts mentioned in the revelation narrative appear to be tantras.97 Would the BraY know the Devm h tmya, this could have a a a
chapter. 95 While this is certainly plausible, the primary textual parallel adduced in evidence is comparatively short, making the case less unambiguous than those of the BraY and Siddhayogevarmata. Sanderson a s identies nineteen p das from Tantrasadbh va 16 as the source of material in Laghuamvara 41. He a a s . also identies related passages in Laghuamvara 18 and Tantrasadbh va 16, and two verses redacted s . a from chapter seven of the latter into Laghuamvara 49. History through Textual Criticism, 44. In a s . presentation at the 13th World Sanskrit Conference, Ronald Davidson reportedly called into question the possibility that the Tantrasadbh va is the source of any material in the Laghuamvara; both his and a s . Sandersons arguments await publication. 96 BraY i.11516ab. a 97 It might also be mentioned that another famous workH las anthology of Prakrit versesbears a the designation Sattasa (=Sankrit Saptaat). s

228 signicant chronological implications: long held to be a work of the sixth century, or even earlier, Yuko Yokochi questions the basis for this, and argues that the Dev m h tmya might instead belong to the latter part of the eighth century.98 a a Signicant uncertainties thus surround both the absolute and relative chronologies of the BraY , further complicated by the fact that the text appears to have multia ple strata. It does seem highly likely that the BraY , in a form close to that preserved a in its oldest Nepalese codex, existed at some point in the eighth century, while I see no grounds for ruling out a seventh-century dating, especially for the older portions of the text. This period in fact appears quite plausible. The possibility that some form of the BraY existed in the sixth century also merits consideration, especially given a its mention in the old Skandapur na, while it is also not impossible that the BraY a. a continued to develop into the early ninth century; this possibility depends upon the dating of sources such as the Laghuamvara and Tantrasadbh va. s . a
geographic and social horizons

One might expect a text as vast as the BraY to offer clues into its region of origin, a and it is hoped that further study from a variety of perspectives will yield such information. At the present juncture, I would assert little more than that the text seems unlikely to herald from the far south or far north and northwest of the subcontinent. In its rst chapter, the BraY provides an unusually detailed account of its dea scent (avat ra) or revelation. Mentioning numerous individuals, this narrative taca itly acknowledges the role of human agencythrough the medium of the tantric guruin the production of scriptural literature. More will be said in the subsequent chapter concerning the model of scripture accommodating this. Idealized though the BraY s account of revelation certainly is, some of the personages and places mena tioned appear entirely realistic. More than twenty-ve individuals are referred to, the majority of whom have their castes and regions of origin specied (table 4.8). Most are designated by initiatory name alone; for several, however, additional in98

See chapter 2, n. 103, in the present thesis.

229 formation is provided: a pre-initiatory name, native village, and/or the name of a parent. These cases comprise the brahmin Srdhara of Kuruksetra, near Delhi, whose . initatory name is Kap labhairava; Candabhairava, an Atharvaveda-school ( tharvanaa a .. . sakh ) brahmin from the village Brhodar, of Sindh in modern Pakistan; Amantr of a . Ujjayin (modern Ujjain, M.P.), a.k.a. Svacchandabhairava, son of the brahmin woman Deik ; and Sattik or Santik 99 of Kanavra village, close to Pray ga (near modern a a a a . Allahabad, U.P.), the daughter of a chandoga-school brahmin, Meghadatta. Sattik is a none other than the goddess Bhairav herself, manifesting in the world of mortals on account of a curse. She recovers her divinity through ritual perfection, and then sets in motion the descent of scriptural wisdom (j na) that culminates in the BraY . a a The narrative of revelation places these and the other, more anonymous individuals mentioned within a cosmic temporal framework: Bhairava teaches the Goddess the scriptural wisdom in its unabridged form of 125,000 verses at the beginning of a particular Kaliyuga, and she transmits this to Srdhara or Kap labhairava in the a second Tret yuga thereafter. Kap labhairava transmits an abbreviated redaction of a a 24,000 verses to his disciple, Devadatta or Padmabhairava, who further reduces the text to twelve-thousand in the Dv parayuga, at the juncture of the Kali age. This a would appear to be the BraY itself, one epithet of which is Dv daas hasraka, The a a s a Tantra of Twelve-thousand Verses.100 Padmabhairava has fourteen disciples from throughout the subcontinent (table 4.8), among whom the texts circulation is apparently restricted for most of the age. In the nal quarter of the Kaliyuga, the initiate Svacchandabhairava comes to learn the scripture. Having had numerous miscarriages, a certain Deik of Uja

jayin prays for a son before the Mother goddesses, and they place in her womb the child called Without a Mantra (Amantr)an accomplished initiate who in
99 It is unclear whether the name given is Sattik or Santik , for tt and nt are often undistinguishable a a in the writing of the BraY s oldest codex. For a discussion, see the annotation on BraY i.28 in the a a translation. 100 See section 5 of the subsequent chapter.


Table 4.8: Individuals mentioned in BraY i a name(s) Sattik or Santik (daughter of a a Meghadatta) Krodhabhairava Srdhara/Kap labhairava a Devadatta/Padmabhairava
disciples of padmabhairava:

place of origin Kanavra village, . near Pray ga a Kuruksetra . Odrade a s . Madhyade a s Madhyade a s Madhyade a s Saur str a. . a Saur str a. . a Sindhuvisaya . Sindhuvisaya . Brhodar village, . Sindhuvisaya . (Kashmir?) (Kashmir?) Kashmir Lamp visaya a . K s a Oddiy na .. a Ujjayin Kal pa village, a Kum rdvpa a

caste/vedic sakha br hmana (chandoga) a . br hmana a . br hmana (bahvrc) a . . br hmana ( tharvana) a . a . br hmana ( tharvana) a . a . br hmana ( tharvana) a . a . sudra sudra ksatriya, r japutra a . ksatriya, r japutra a . br hmana ( tharvana) a . a . m tanga a m tanga a br hmana (chandoga) a . br hmana a . (v jimadhyamdina) a . br hmana (bahvrc) a . . br hmana (taittirya, a . apastambhabr hmana) a . br hmana a .

Raktabhairava Jv l bhairava aa Hel bhairava a V mabhairava a Vijayabhairava Bbhatsabhairava Gajakarnabhairava . Candabhairava (son of Yajasoma) .. Kum rabhairava a Krodhabhairava (ii) Tejabhairava Kar labhairava a Ucchusmabhairava . Yamabhairava Visnubhairava .. Daksinabhairava . . Sekharabhairava
svacchanda and his disciples: disciples of candabhairava: ..

disciples of padmabhairava (cont.):

Amantr/Svacchandabhairava (son of Deik ) a Candabhairava (ii?) .. Bindubhairava (=Vibhubhairava?) M y bhairava a a Anantabhairava Vibhubhairava Visnubhairava (ii?) ..

231 a previous birth had broken the initiatory Pledges and failed to achieve siddhi.101 Reborn, Amantr attains siddhi through practice of the vidy -mantra. Consecrated as a Svacchandabhairava, he learns the Tantra of Twelve-thousand Verses from Krodhabhairava, the primordial disciple of the Goddess.102 His own disciples preside over ever-diminishing redactions of the scripture at the twilight of the cosmic cycle, at the end of which yogins hide away the teachings altogether. Concealed throughout the Krta, Tret , and Dv para ages of the subsequent cycle, at the begining of the next a a . Kaliyuga the Goddess reveals the unabbreviated scripture of 125,000 verses to (the new incarnation of) Svacchandabhairava. He teaches a redaction of 12,000 verses to a certain Visnubhairava103 in the legendary village of Kal pa, renowned as an abode a .. of sages.104 Visnubhairava then transmits the text to the inhabitants of the Isle of .. Maidens (kum rdvpa)the civilized world.105 With this we arrive, unambiguously, a at the BraY . The narrative ends by predicting that the text shall achieve tremendous a popularity, being present in the homes of all worthy of siddhi. While this narrative contains tantalizing details concerning individuals and places, its idealized framework obscures potential historical data. Noteworthy is the fact that the text claims a pan-South Asian genealogy. This could reect the wish to ascribe a universal dimension to what was, in fact, a manifestly local tradition. However, it is entirely plausible that the textual community involved pan-South Asian lineages. Among the gures mentioned, Svacchandabhairava of Ujjayin appears pivotal to
See BraY i.78cd86ab. The narrative concerning Amantrin or Svacchandabhairava and his discia ples, spanning two Kaliyugas, comprises BraY i.78cd118. a 102 The verses in question, 7879, are somewhat problematic; see the annotation thereon. 103 Visnubhairava is mentioned twice in the revelation narrative: as a v jimadhyamdina-school braha . .. min from Lamp , one of the fourteen disciples of Padmabhairava (verse 73); and as the student of a Svacchandabhairava (verses 11214), medium for the dissemination of the Tantra of Twelve-thousand Verses to the residents of Kum rdvpa (see below). There is no suggestion that these are the same a individual; however, the second could be a subsequent incarnation of the former. 104 See, for instance, Bh gavatapur na 9.12.6, 9.22.17, 10.87.7, and 12.2.3738; and the Daavat ra of a a. s a Ksemendra (opening of the Kalkyavat ra section). In both sources, the site is associated with the a . Kaliyuga, its nal period in particular. I am grateful to Isaacson for these references. 105 On Kany - or Kum rdvpa, see Tantr loka 8.8592, especially verse 91 (n n varnasram c rasukhaa a a a a . a a . duhkhavicitrat | kany dvpe yatas tena karmabhuh seyam uttam : Because of the existence of [the system a a a . of] manifold castes and stages of life, and the variegation of pleasure and suffering on Kany dvpa, it a is the greatest land [for the performance] of pious acts (karman)).

232 the texts transmission; yet there are no strong grounds for assuming he or anyone else mentioned represents an historical gure. It is nonetheless possible that the BraY s revelation narrative preserves a record of some key individuals connected a with the scripture and its background, cast within an idealized temporal and geographic framework. The geographical horizons of the BraY , as indicated by the places it names, have a two distinct spheres. On one hand, the revelation narrative presents an expansive topographic vision: individuals involved in the transmission of scripture span from Odrade a in the eastpresumably related to todays Orissato Sindh (sindhuvisaya) s . . and the Swat Valley (oddiy na) in Pakistan, and Kashmir (kamra) and Lamp in the s a .. a far north. The far south is not represented, however, nor, e.g., Nepal. In contrast, the sacred places mentioned in the BraY suggest more restricted geographical horizons a (tables 4.910106 ). The BraY s primary deity mandala, as delineated in chapter three, a .. positions eight sacred sites in the cardinal and ordinal directions, referred to in this context as cremation grounds (maana). This mandala of eight cremation grounds s s .. encompasses central India and the Deccan, the North-Indian heartland, and Orissa and Bengal in the east. This suggests a largely central-eastern geographic sphere, the farthest point west being Kollagiri, perhaps corresponding to the Kolhapur in the Deccan (in modern Maharashtra). Orissa is disproportionately represented, while Kotivarsa of modern Dinajpur district in northwestern Bangladesh marks the far . . northeastern horizon. This mandala of eight cremation grounds has a close parallel .. in BraY lxxxivs list of eight p. has (sacred mounds; see table 4.10).107 Discrepa t ancies between the two can in part can be accounted for by synonymsJayantik is a
106 In identifying the probable regions of the sacred sites the BraY enumerates, I follow Sanderson, a History through Textual Criticism, 7 (n. 4). 107 BraY lxxxiv.81: a

pray g varuna koll at. ah s jayantik | a a a .t a a a . caritraik mraka caiva kotivarsam tath s. amam a a. t . . . .
attah s ] em.; hattah s Bya .. a a .. a a


Tantrasadbh va 15.21 is identical to this verse, offering as substantive variants at. ah s (adopted above) a .t a a and the corrupt caritrek mbukam (81c; mss as reported by Dyczkowski). a .

233 presumably Ujjayinbut not entirely: V r nas is replaced by Pray ga, while the a a. a synonymity of some site names is uncertain. It is difcult to say which of the two geographic spheres invokedthe pan-South Asian or the central and easternmight better reect the early textual community of the BraY . a The Table 4.9: The Eight Cremation Grounds in BraY iii a name(s) V r nas a a. Viraj a Kollagiri Prabh sa a Ujjain s Bhute vara Ek mra a Kotivarsa . . probable location Varanasi, U.P. Jajpur, Orissa Kolhapur, Maharashtra Somnath, Junagadh Dit., Gujarat Ujjain, M.P. ? Bhuvanesvar, Orissa W. Dinajpur Dit., Bangladesh

mandala of cremation grounds .. might have been inherited from

older sources, possibly shedding light upon the geographic horizons of the early traditionor, potentially, the geographic sphere of the BraY in its earliest form. a The more expansive geography envisioned in the BraY s revelation narrative could, a on the other hand, reect a broadening of the cults horizons by the period of the texts nal redaction. In addition, it is difcult to imagine obscure villages such as Brhodar of Sindh and Kanavra, near Pray ganeither of which seems traceable a . . nownding mention in the absence of a genuine connection to the text. Concerning the individuals mentioned in the revelation narrative, two facts stand out: the prominence of male brahmins in the production and transmission of scripture, and the simultaneous representation of a spectrum of other castes. Eleven brahmin men gure among the twenty-ve odd individuals named, representing a variety of regions and Vedic schools. The roster features two ksatriyas and two sudras, and . includes two members of the tribal m tanga community as well; information is not a provided concerning the remaining individuals. All of the more important gures are brahmins, with Sattik also a brahminthe single woman of signicant status. a I suspect that this points toward the simultaneous diversity of participants in the tantric Saivism of the BraY caste and gender are, in principle, not barsand the a reality that literacy, and therefore textual production, was undoubtedly a domain in

234 which male brahmins were particularly prominent. At the same time, the remarkably rustic Sanskrit of the BraY , heavily inuenced as it is by the Middle Indic vernacua lar, suggests redactors of little training in Sanskrit grammarthe study of which had pride of place in Brahmanical education. Indeed, the language of the BraY comprises a a body of data potentially useful for locating the text in time, place, and sociolinguistic community. Its evaluation faces limitations, though, for the paucity of Table 4.10: The Eight Sacred Mounds (p. ha) in BraY lxxxiv t a name(s) Pray ga a Varuna . Koll /Kolagiri a Attah sa .. a Jayantik a Caritra Ek mra a Kotivarsa/Devkotta .. . . probable location Allahabad, U.P. (=Varana, i.e. Varanasi?) . Kolhapur, Maharashtra Birbhum Dit., W. Bengal Ujjayin/Ujjain? Puri Dit., Orissa Bhuvanesvar, Orissa W. Din jpur Dit., Bangladesh a

manuscript evidence complicates the effort to distinguish between the vagaries of scribal transmission and genuine irregularity. Nonetheless, the sheer volume of material in a text of more than 12,000 verses partly compensates for this problem. In the present study, I do not offer a systematic survey of the language of the BraY , a although the philological notes to the critical edition discuss linguistic issues as they arise. Further study is required to determine whether the BraY possesses dialectical a features that could be linked to a particular region and period. In general character and most particulars, the language of the BraY bears coma parison with that of the Siddhayogevarmata, on which Judit Trzsk has written a s useful and detailed study.108 To provide a single illustration, note that in the BraY , a metrical concerns override grammar when the two conict, a principle consonant with other varieties of Middle-Indic inuenced Sanskrit. This is, however, taken much further than in Epic and the so-called Buddhist Hybrid variety. Particularly striking are cases in which a spurious visarga is sufxed for metrical reasons: note, for instance, the cases of the adverb upari (BraY xxxiii.86cd, sane c sane sth pya sapa a a

Trzsk, Doctrine of Magic Female Spirits, xxvilxix.

235 tadh nyoparis tath ), the locative plural noun diksu (evam diksu vidiksu ca asanam a a . . . . . s samnive[]ayet, xliv.646cd), and the verb raksyati (s dhak n m pad n m tu putravad s a a a. a a. . . raksyatis tath , BraY xi.125cd). Again driven by meter, nal consonants are opa a . tionally omitted, especially the nal -t of optative verbs; as this is characteristic of Prakrit, pronunciation is surely a factor.109 Additionally, the BraY accepts a degree a of metrical freedom by allowing for verse-quarters of nine syllables, provided that the nal four-syllable cadence pattern remains intact, and provided that several of the initial ve syllables are short (laghu).110 For a text which places considerable emphasis on attaining worldly power, the BraY makes surprisingly few references to political or military power as an aim a of ritual.111 Furthermore, there are few signicant indications of a courtly or even urban environment in the BraY . Although urban centers nd mentionK s/V r a a a a nas and Ujjayintheir civic status seems incidental, for these are important Saiva . places of pilgrimage. These circumstances seem suggestive of a rural social milieu. Among the possible regions of origin, Orissa might seem a strong candidate: note, in particular, the prominence of Orissan sites in the sacred geography of the BraY a (tables 4.910), and the texts references to Ekap dabhairava, an unusual one-legged a form of the deity common in Orissan-provenance sculpture but attested elsewhere, as far as I can determine presently, only in neighboring Andhra Pradesh.112 Temples suggestive of tantric goddess cults are, furthermore, widely attested in the sculpture of this region, particularly from the ninth century.113 In addition, Orissa, or a region
See the annotation on BraY i.60. a See the annotation on BraY i.20 in part ii. a 111 Exceptions include occasional reference to the good fortune of kingship or royal fortune (r jyasaubh gya) as one of several aims of ritual; cf., e.g., BraY xliv.314 (anena kramayogena a a a r jyasaubh gyam eva ca | arth ms ca vividh [m] caiva mantr sarv n av pnuy t 314). Kingship itself a a a. a . s a a a (r jya) is mentioned in a list of siddhis in BraY lxxxvii.203d. Chapter li species which meats should a a be offered in re sacrice depending on ones caste and station; in this context, kings are mentioned as individuals who must offer human esh. In addition, there are references to protection of ones army (nijasainyasya raksana) and terrifying the opponents army (parasainyasa tr sana) in lists of magical a . . objectives, with no particular emphasis. 112 The connection between Ekap dabhairava, the BraY , and Orissa has been suggested to me by a a Sanderson (personal communication, May 2003). 113 a Thomas Donaldson, Tantra and S kta art of Orissa, vol. 1, passim.
110 109

236 of comparable peripheral status in the Brahmanical cultural world, would seem compatible with the sociolinguistic milieu of the text.

Chapter 5
To Name a Tantra: Identity, History, and the Saiva Canon in the Epithets of the Brahmayamala

The text thus far referred to as the Brahmay mala in fact designates itself by a vaa riety of titular epithets. In my attempt to introduce this little-studied and unusually voluminous tantra, the present chapter attempts to explicate the functions and historical signicance of the texts various titular epithets. I demonstrate how these provide important lenses into the texts contents, history, and rhetorical world, especially the position its authors sought to articulate for it within a canon of Saiva scripture. Though possessing several titles, evidently the scripture was best known within the tradition as Brahmay mala. However, the title Picumata had considerable impora tance as well, and the text has two additional titular epithets: Nav ksaravidh na (Proa . a cedure of the Nine-Syllable Vidy -mantra) and Dv daas hasraka (Tantra of Twelvea a s a thousand Verses). Abhinavagupta cites the text as Brahmay mala on eleven occaa sions,1 also making four references to the Picumata.2 Within the body of the text, exclusive of chapter colophons, references to the titles Brahmay mala and Picumata are a clustered heavily in the latter sections of the text,3 while discussion of the signicance
1 Viz. Tantr loka 4.54b, 4.60b, 5.97c, 13.145ab, 15.44c, 18.9a, 23.43d, 27.29a, 28.419b, 28.423b, and a 29.11a. 2 Evidently preferring the title Brahmay mala, Abhinavagupta also mentions the Picumata in Tantra aloka 28.383a, while in 27.21d referring to Picuastra, and Srpicu in 28.409c. Note also the phrase s picuproktam, spoken in the Picu[mata], in Tantr loka 27.24b. In addition, Jayaratha, commenting on a . Tantr loka 1.18, quotes from a scriptural source that refers to the Picutantra. a 3 References to the title Brahmay mala occur in xxxviiii.26a, lxi.1c (y malam tantram), lxi.35c (y mala), a a a . lxx.100c, lxxv.36d, lxxvi.93a, lxxvii.1c (y malatantra), lxxxii.16b (y mala), lxxxii.101d, lxxxii.124b, a a lxxxvi.85b, and ci.31a, besides numerous references in lxx, lxxi, and lxxiv. References to the title Picumata[tantra] occur in lxi.7a, lxi.250a, lxxi.110d, lxxxiv.222c, while Picutantra occurs in lxi.39a,


238 of both titles occurs particularly in lxx and lxxi. Both of these titles, moreover, had the distinction of apotheosis, for the BraY describes the title deities Y malabhairava a a or Y male vara, Lord of the Y malatantra, and Picubhairava.4 a s a

5.1 Brahmayamala
Y mala has as its primary meaning pair, and in tantric literature, frequently has a the specic sense of coupled god and goddess, especially in contrast to ekavra or ekavr , a solitary deity.5 The y malatantras as a genre appear dened, in theory, by a a teaching the cult of a coupled supreme Godhead.6 On this basis I prefer the English rendering Union Tantra for the scriptures designated y mala or y malatantra. In the a a Brahmay mala, the supreme, paired divinity comprises the deities Kap lsabhairava a a and the Great Goddess, whose primary names include Canda K p lin (Fierce Skull. . a a bearer), Aghor (Un-Dreadful; also Aghore and Aghore var), and Bhairav. As s s noted already, Sanderson points out that the gender polarity of this supreme divinity is imbalanced, for the mantric being of the supreme Goddess, the Nine-Syllable Vidy -mantra, subsumes that of Bhairava and the mandala deities. a .. There might appear a degree of incongruity in a Saiva tantra cast as dialog between Bhairava and the Goddess bearing the title Brahmay mala, for the orthodox a creator-deity Brahm has remarkably little to do with the text. His role is conned to a a narrative episode in BraY lxxx, which contains an interesting tantric variant on the a Skull of Brahm (brahmakap la) myth.7 In this episode, Brahm has the distinction a a a
lxvi.1b, lxxiv.205c, lxxvi.93c, lxxxi.23d, lxxxii.16d, lxxxvii.257b, and lxxxvii.258b. 4 Y male vara and his mantra, mandala-deities, and the sakti in union with him (tady male) are a s a .. . taught in BraY lxxi, while Picubhairava and his lump (pinda) or heap (kuta) mantra are described a .. in lxviii. 5 Note for example Ksemar jas expression bhairavay mala, in the sense of Bhairava together with a a . Bhairav, e.g. ad Netratantra 10.12ab and 10.13ab (. . . caitad bhairavay malam yajeta, and bhairavay malam a a . . dhy yet, respectively). In the Tantr loka, Abhinavagupta uses the word y mala in the sense of malea a a female pair, for example in the well-known mangalaloka, where he speaks of his conception by the s y mala of his own parents. In 29.120, he appears to use y mala in the sense of conjoined siva and sakti a a (aktiaktimat). s s 6 Note Jayadrathay mala I.xxxiii.25ab: dampatyayogatah puj y maleti nigadyate, The word y mala a a . a a means worship of [/in accordance to] the conjoined (yogatah) [divine] conjugal pair. I am grateful to . Alexis Sanderson for providing this reference. 7 Compare this with, e.g., Skandapur na 67. a.

239 of losing one of his heads to Bhairava, thence providing him a skull as alms-bowl. Brahm s cosmogonic role, mentioned in a single passage, exists only by the blessing a of Bhairava8 and perhaps to the extent that his skull, lled with blood Visnu offers .. Bhairava as alms from his own body, serves as locus for the creation of the primordial saktis.9 Although few old y malatantras survive, extant text lists show that their a titles were formed by appending -y mala to the names of Brahmanical deities, or in a some cases other mythical gures. In this regard, the y malatantras appear to mirror a the naming conventions of Pur nic literature, which includes texts with titles such as a. Skandapur na, V yupur na, Garudapur na, and M rkandeyapur na. a. a a. a. a .. a. . More specically, it appears possible that y malatantras were, in principle, texts a named after the Brahmanical deities who serve as counterparts to and namesakes for the Mother goddesses (m tr). This possibility receives support from the fact that a. a number of sources describe or identify the y malatantras as m trtantras, Tantras a a.

BraY lxxx.15760ab: a etac chrutv tu tau devau pranipatya punah punah | a . . . janmamrtyubhayatrastau puna c sv sitau may 157 s a a a . datv varasahasr ni buddhik m nus ratah | a a. a a a . srs. im kurusva he brahma tvam p laya jan rdana 158 a ..t . . . a ajitas trisu lokesu subhaga ca bhavisyasi | s . . . mama tulyabalo vatsa maccharre bhavisyasi 159 . praj n n k ranam brahma vivekt vedav din m | a a a . . a a a
a 157a chrutv ] em.; chrutv Bya 157d c sv sitau ] em.; c sv sito Bya a a a a a balau Bya 160b v din m ] em.; v din h a a a a. 159c balo ] em. Isaacson;

After hearing this, the two gods [Brahm and Visnu] prostrated again and again. Tera .. ried by fear of [the cycles of] birth and death, I [Bhairava] again consoled them, after granting thousands of boons in accordance with their minds wishes: O Brahm , create! a You, Visnu, maintain! You shall be undefeated in the triple universe, and fortunate. My .. child, having power equal to me, O Brahm , you shall be the cause of the beings (praj ) a a () [arisen] from my body (?), the wisest of those adhering to the Vedas. The interpretation of 159cd is unclear. By Isaacsons suggestion, I have emended balau to balo, referring to Brahm . 159ds maccharre has been tentatively construed as a locative of source (cf. V. S. Apte, a The Students Guide to Sanskrit Composition, 77). Judit Trzsk suggests the emendation maccharro, which seems plausible (personal communication, November, 2006). 9 BraY lxxx, e.g. 225cd26ab: a brahmanasyottam nge tu visnuonitapurite 225 a . .. s . mama drs. inip tena utthit s c tra ramayah | t a a a s .. .
utthit s ] em.; utthit Bya a a ra mayah ] corr.; rasmayah Bya s . .

And when my gaze fell upon the head of Brahm , lled with the blood of Visnu, the a .. [akti-]rays arose within it. s One could alternatively understand 225cd as a locative absolute clause, viz. when the skull of Brahm a was lled with the blood of Visnu, . . . . ..

240 of the Mother Goddesses. Among these is the old Skandapur na, as discussed in a. chapter two. The eight y malatantras listed in BraY xxxviii correspond to six of the a a standard Seven Mothers, omitting the independent goddess C munda, and adding a . . y malatantras of Kubera and V yu.10 Although not normally included among the a a Seven or Eight Mothers, the BraY and other sources do attest the goddesses Kauber a and V yav, saktis corresponding to Kubera and V yu.11 While this possibility suga a gests a meaningful basis for the naming convention of y malatantras, evidence in a support of this view is inconclusive. None of the early lists of y malatantras, as a identied by Alexis Sanderson,12 map precisely to the Seven or Eight mothers, although y malatantras of Brahm /Brahm n, Rudra/Rudr n, Visnu/Vaisnav, and a a a. a. .. .. Skanda/Kaum r feature prominently. In addition, many y malatantra titles, whether a a of actual or idealized texts, do not seem compatible with this model. These include, for example, the Nandiy mala,13 bearing the name of Sivas prominent retainer (gana) a . Nandin, and the (extant) Jayadrathay mala, named after the ambiguous warrior and a Siva-bhakta of the Mah bh rata. a a The Jayadrathay mala nonetheless advances precisely this Mother-goddess framea work for understanding the y malatantras. In the thirty-sixth chapter of the rst book a (satka), a passage listing root scriptures of the Vidy ptha refers to the sevenfold a . . . Union Tantras of the Mothers (m try mala), beginning with the Mother goddess a. a Brahm n. In this conception, BraY is apparently rst of seven scriptures correa. a sponding to the Seven Mothers.14 Accounting for the diversity of titles, chapter forty10

BraY xxxviii.25cd27ab: a rudray malam anya ca tath vai skanday malam 25 a a a . brahmay malakam caiva visnuy malam eva ca | a . .. a yamay malakam c nyam v yuy malam eva ca 26 a a . a . a kuberay malam caiva indray malam eva ca | a a .
25d skanda ] em. (Sanderson); kanda Bya

BraY xxix.94a includes Kauber and V yav in a set of six yogins, which also includes the Braha a manical saktis Hut san (i.e. Agney), Y my , Nairrt, V run, and Sarkar (i.e. Sakr n, Indr n?). Cf., a a a a. a. . a . e.g., M linvijayottaratantra 8.9798 and Kubjik mata 14.81. a a 12 Sanderson, History through Textual Criticism in the Study of Saivism, the Pacar tra and the a Buddhist Yogintantras, 7. 13 Mentioned in e.g. Skandapur na 171.129b. a. 14 This passage also speaks of divisions of six and ve y malas corresponding to smaller groups of a


241 two describes the ve primary root-scripture y malas as Mother Tantras, from which a emerge secondary (upa-) y malatantras. From the ve upay malas in turn emerge vara a ious y malas of the duts (Consorts), yogins, and other minor goddesses.15 While a this model of the core y malas as Mother Tantras might have historical value, it apa pears also to reect a posthumous and idealized classication. Certainly it seems that many y malas and upay malas posited in this conception existed only in name. a a Naming conventions of y malatantras hence suggest that Brahmay mala has as a a an important, perhaps original meaning, the Union Tantra of Brahm (or of a Brahm -Brahm n), with the Jayadrathay mala and other sources conrming a strong a a. a association between the y malatantras and Mother goddesses. However, the texts a self-understanding of its title diverges considerably. In the opening of chapter seventyone, the Goddess asks, But why [the word] y mala, O God? I want to know truly.16 a Commencing to answer, Bhairava declares, this Brahmay malatantra emerged through a the sequence of the brahma-[mantra]s.17 Elaborating upon the link between the brahma- of the title and the archaic Saiva brahmamantras, these ve mantra-faces of Sad siva are identied with Tumburu and the Four Sisters, the core pantheon of the a v matantras. As such, the brahma-mantras consist of both siva (Tumburu) and sakti a (the Sisters). The same passage also offers an alternative interpretation of brahma-, understanding it in the sense of brahman, the formless absolute, rather than the brahma-mantras or deity Brahm . The scripture is called Brahmay mala because Siva a a
Mother goddesses. Jayadrathay mala I.xxxvi.1625 (text courtesy of Sanderson). a 15 a Jayadrathay mala, I.xxxxii.17 (text courtesy of Sanderson). In this model, the ve mulay malas are a s those of Brahm , Visnu, Rudra, Skanda, and Um , which give rise to the Vet lay mala, Ianay mala, a a a a a .. Atharvay mala, Sarasvaty mala, and Somay mala, respectively. a a a 16 BraY lxxi.1ab: y malan tu katham deva j tum icch mi tattvatah. a a a a . . 17 BraY lxxi.3ab: brahmay malatantredam brahmabhamgy vini[h]rtam, understanding tantredam as a a . . a . s. . tantram idam, with metri causa elision of the case ending. (The same expression occurs as BraY ci.31a.) a Brahmabhangi refers to the sequence of the ve brahma-mantras, beginning with that of Sadyoj ta. See a Dominic Goodall, et al, The Pac varanastava of Aghoraiv c rya: A Twelfth-century South India Prescripa s a a . tion for the Visualization of Sad siva and His Retinue, 13637. See also Ksemar jas explanation of the a a . term commenting on Svacchandatantra 1.46a. In the BraY , cf. lxi.181cd: pacabh gakrt [m] vidy [m] a a . a . a . a brahmabhangy niyojayet, One should employ the vidy -mantra divided in ve parts, in the brahmabhaa ngi sequence. That this sequence begins with Sadyoj ta appears to be stated in a corrupt half-loka, a s lxxi.3cd: sady nt divibh gena srs. im grena samyutam. Most probably, sady nt di should be emended to a a a a a ..t a . . . sadyoj di , beginning with Sadyoja/Sadyoj ta. a a

242 and Sakti, the y mala, are both vast (brhattv t) and cause expansion (brmhakatv t)an a a a . .. etymology of brahman.18 An explanation in BraY lxx focuses instead upon the term y mala. The text a a is the Brahmay mala because it encompasses within itself a variety of oppositions a (y mala): those of liberation and supernatural attainment (mukti and bhukti); ritual a and doctrine (kriy and j na); and teachings of both the pure and impure ways, a a as well as the mixed and unmixed ritual paths.19 This denition ties into the texts classication of scripture, ritual, and practitioners according to the threefold schemata of pure, impure, and mixed.20 Because the text teaches ritual paths for all

BraY lxxi.6769ab: a brahmam pacavibh gastham na pacarahitam kvacit | a . . . k mab na jay dy s ca devyas tumburupurvak h 67 a a. a a a. nand ditithayah paca sivaakty tmak h priye | a s a a. . brhattv d brmhakatv c ca sivasaktiv ubh v api 68 a .. a a . jagaty asmin sugtau tu brahmay malasamjay | a a .
67c k mav na ] em.; k mav na Bya a a. a a. 67d purvak h ] em.; purvvak Bya a. a va akty tmak h ] em.; siv sakty tmak Bya 68c brhattv d ] em.; b(a?)hatv d Bya s a a. a a a a a . 68b si

brahman always has ve divisions; in no case is it devoid of the ve. [These are] the ve arrows of K madeva, the [ve] goddesses who begin with Jay , preceded by Tumburu, a a and the ve [auspicious] dates beginning with nand , consisting of siva and sakti. Because a of being vast and causing expansion, both siva and sakti are well-known in this world by the designation brahmay mala. a Bya s k mab na in 67c is suspect; the connective ca suggests reading k mab na jay dy s ca. It seems a a. a a. a a unlikely that the arrows of K madevawhich number veare identied with the Four Sisters and a Tumburu, for this would make Tumburu himself one of Cupids weapons. In 68d, it seems that sivaaktiv s represents the dual sivaakt, with an unusual hiatus-breaking -v-. This could reect inuence of sandhi s of the type ubh v api for ubhau api. a Cf. the denition of the word brahman provided in BraY lix.24cd: brha[t]tv d brmhikatv c ca tad a a .. a . brahmam sabdat m gatam. This is a standard etymological explanation (nirukti) of the term; note, a. . e.g., Ksemar jas comments ad Svacchandatantra 1.46: brahmano brhattv t brmhakatv t viv tmanah saka a .. a s a . . . . . timurteh. In Par khyatantra 14.78cd, a similar etymology is provided for Brahm , as one of the ve a a Causal Lords (k ranevara): brhattv d brmhakatv c ca brahm v brahmayogatah, [He is called] Brahm bea . s a .. a a a a . . cause he is great, because He lls, or because He possesses brahman (translation of Dominic Goodall, The Par khyatantra. A Scripture of the Saiva Siddh nta, 374). a a 19 BraY lxx.99cd102ab: a suddh suddhe tath m rge mir mire ca tatsamah 99 a a a sa s . bhuktimuktikriy j nam suddh suddhasya laksanam | a a . a . . brahmay malam ity uktam etad devy matam tath 100 a a a . . suddh suddhavibh gena j nam uktam var nane | a a a a . mantraaktivivekena paravidy mayena ca 101 s a suddh suddhavibh gastham mirakam y malam smrtam | a a s . . a . .
100a kriy ] corr.; kry Bya a . a

Intriguingly, Devy mata in 100d appears to be another epithet of the BraY . cf. Devy y mala, the title of a a a a a text quoted by Abhinavagupta. 20 See, e.g., the annotation on BraY i.3839 in part ii. a

243 three grades of s dhaka, characterized by the basic opposition of pure and impure, it a is the y malatantra. a

5.2 Picumata
Alongside Brahmay mala, Picumata (and Picutantra) gures as foremost among a the titles and epithets provided in the texts colophons. A number of other tantras also bear the designation -mata, Doctrine [Tantra], such as the Siddhayogevarmata. s This epithet actually occurs with twice the frequency of Brahmay mala in colophons, a and the text is cited by this name and variations such as Picutantra in the Bhairavama a ngal 21 and Tantr loka.22 Sanderson highlights the epithets importance by referring a to the scripture as the Picumata-Brahmay mala.23 This choice has merit, for instance a clearly distinguishing the text from later tantras claiming the title Brahmay mala. a Nonetheless, I deem Brahmay mala the primary title and generally refer to the text a as such, giving weight to the balance of evidence from citations and text lists. The relative priority of a texts titles is of course a modern concern; the tradition itself readily accommodated multiplicity. In Sanskrit, picu ordinarily means cotton, while picumarda and picumanda refer to the neem tree. It cannot be assumed forthright that picu has in the BraY precisely a the same meaning, but there are some indications of a botanical identication: we nd references to picu trees, owers, and leaves.24 It is conceivable that these refer to neem (Sanskrit nimba), picuvrksa being, that is, identical in meaning to picumanda and . . picumarda. More likely, perhaps, is that the picu tree refers to the cotton plant, or . even the much larger silk cotton tree (Sanskrit apuran). Irrespective of the identity of the picu tree, a botanical referent is difcult to reconcile with most references to picu (as opposed to picu tree, etc.) in the BraY . In a
See below (nn. 68, 104). See above (n. 2). 23 See for example Saivism and the Tantric Traditions, 672. 24 Cf. BraY lxiv.51cd (picuvrksad ato devi krtv mandalakam subham), lxxvi.98ab (arkapatram picoh a . . . a .. . . patra[m] durbhag puspam eva ca), and lxxvi.104cd (arkapuspam picoh puspam r sabhasya tu sekajam). a . . . . . . . a
22 21

244 many passages, it appears to be a liquid, a substance one lls ( pur) into a vessel or the mouth, or uses to sprinkle on and thereby sacralize ritual space.25 Frequently, it is a liquid used in the guest-water offering (argha) for deities, particularly as proffered from a skull-bowl. Often, argha is specied as consisting of either picu or alcohol, or both together.26 It can also serve as an oblation in the re sacrice (homa).27 BraY a xlv, expounding the deity Manth nabhairava and his unusual churning [of the caula dron] ritual, is particularly rich in references to picu as an offering substance. Typical is its instruction, a skull lled with picu, fully empowered by the vidy -mantra and a its ancillarieswith that, the guest-offering should be given to the goddesses and to Bhairava.28 In BraY vi and viii, the term picuvaktra occurs as an iconographical dea scription, probably meaning having picu in the mouth. Related to this, Picuvaktr a is the name of a minor goddess mentioned in the Agnipur na and Tantrasadbh va, a. a while the latter also mentions a Bhairava named Picuvaktra.29 Despite substantial data, the identity of the substance picu is not entirely certain. As a liquid used in ritual, picu must surely be distinguished from the picu-tree and its products. It is in fact almost certainly a bodily substance, and appears primarily to refer to sexual uids. That it is a bodily product is suggested, for instance, by references to mah picu, in which the prex great (mah -) carries its not-so-secret a a code sense of human; cf. mah m msa, human esh. BraY xlix makes reference a a. a to cooking rice for the food offerings (caru) using raja (female sexual/menstrual uid) and picu, or else blood (rakta); paired as it is with female uids, it seems likely that here picu refers to semen.30 In no case where picu is listed alongside other substances
Cf., e.g., lxv.105, quoted below. Cf., e.g., xlv.52cd (picun madirenaiva devdev n sa tarpayet) and xlvi.38ab (vigrhya d payed argham a a a . . . picun suray tha v ). a a a 27 . Cf. BraY xlviii.23cd: picupurnahutim dady t s dhyabjasamanvitam. a a a . 28 BraY xlv.105: a
26 25

kap lam picun purnam vidy ng bhih sumantritam | a . a . . a a . ten rgham tu prad tavyam devn m bhairavasya tu 105 a a a. . . Agnipur na 146.19ab (piac picuvaktr ca lolup aindrsambhav h), Tantrasadbh va 13.83cd (bhairav a. s a a a. a jharjhar caiva picuvaktr tath par[ ]), and Tantrasadbh va 13.41ab (bhairavo jharjhara caiva picuvaktras tath a a a a s a parah). . 30 BraY xlix.10cd12ab: a

245 does seminal uid nd separate mention. But as discussed below, several passages appear to identify picu with the combined male and female sexual uids, although these passages read much else into the term as well. It hence seems possible that picu refers in general to sexual uids, whether male, female, or both.31 Somewhat confusing matters, an unmetrical verse quarter in BraY xlv makes reference to gaa japicu, elephant picu, and elsewhere reference seems to be made to picu of the jackal (kros. uka).32 It appears somewhat improbable that picu refers here to the sexual uid .t of the elephant and jackal; urine or blood seems more likely. This might in fact sugrajapicubhi v raktam tandul ms tena bh vayet 10 a a . .. a. tryaha ca bh vayet pr jah sosayitv sth payet punah | a a . . a a . tat s dhayta carukam picun kros. hukasya tu 11 a a . .t bhaksayed amrtam divyam bhaksayitv sivo bhavet | a . . . . .
10d tandul ms ] em.; tandul Bya 11a tryaha ] em.; tr ha Bya bh vayet ] em.; bh vaye Bya a a a .. a. .. a

Blood, or with rajah and picuone should prepare the rice with that. The wise person . should prepare it for three days; and then, after drying it out, should set it up again. He should prepare that caru with the picu of a jackal. He should consume the divine nectar; after consuming it, he would become Siva. One solution to the metrical problem of 10c could be rajahpicubhi v raktam. a . . Incidentally, one might wonder whether and how Saiva ritualists actually procured sexual uids by the pitcher-full or skull-full. Barring those most zealous in such matters, this seems improbable. Although the texts say little on the subject, one possibility is that the sexual uids used in ritual were diluted in water. Note for example the following instruction in BraY xxiv: a ksobhayitv tato saktim yonim praksalayet tatah 11 a . . . . . ksalanam bh ndake grhya aesam raktasamyutam | a. . s . . . . . purvoktena vidh nena carukam sr vayet tatah 12 a . a . Next, after bringing the sakti to orgasm, one should then wash the female organ. After gathering the rinse-uid into a container, completely, together with the female sexual uid, one should cook the food offerings (caru) [with that], in accordance with the aforementioned procedure. Similarly, references to picumadya, picu-liquor, suggest dilution in liquid. 31 This might explain what the Matas ra means by picukrda, sporting with picu; while not explicitly a . clear, the context suggests that this refers to something done after copulation. Matas ra, chapter 1: a bjam jvamayam sres. ham mrtyujid rasam uttamam | . . .t . . picukrda[m] tatah kury d devy mantrena mantrayet | . . a a . . madyena suray v tha purayec c rghap trakam | a a a a ngmpp b28/16, folio 6r (transcription courtesy of Somadeva Vasudeva). 32 On kros. ukapicu, see BraY xlix.10cd12b, quoted above. Gajapicu is mentioned in BraY xlv.21c: a a .t gajapicukasamyuktam bhmavaktr bhimantritam 21 a . . purayitv kap lam tu v yavy m sth payet priye | a a . a a. a This section, for which Bya is not fully legible, outlines a mandala with an inner circuit of eight skulls .. arrayed in the cardinal and ordinal directions, lled with various nondual liquids. These are empow ered by the mantras of the four devs and four duts who are installed therein. The other liquids include, for instance, human and rabbit blood, blood of the practitioner (yoginasya [!] rakta), alcohol, and human picu (mah picu). a

246 gest that picu is a general term for bodily uids, usually having the specic sense of sexual uid. Kashmiri nondual exegetical literature attests the term picuvaktra (Picu-mouth) as a synonym for yoginvaktra, the mouth of the Yogin.33 In a vision of the scrip tural canon articulated in the Tantr loka and Tantras ra, Abhinavagupta posits Kaula a a scriptures as revelations of a sixth, hidden stream, emerging not from Sad sivas a ve faces but a lower face/orice (adhovaktra, etc.) or netherworld [facing] face (p t lavaktra). These are synonyms of yoginvaktra.34 In the kulay ga (clan rite) exaa a pounded in Tantr loka 29, this yoginvaktrathe ultimate source of the lineage (sama prad ya) and scriptural wisdombecomes the divine homolog of the ritual consorts a sexual organ. Drinking uids from this thus becomes a medium for gnostic experience. The use of picuvaktra as a synonym of yoginvaktra might be based upon the latters identication with the vagina in ritual, locus of the substance picu. However, anatomical conceptions of the yoginvaktra are ambiguous: it in fact appears that to the Kashmiri Saiva authors Jayaratha and Ksemar ja, the adhovaktra or the lower a . mouth refers in microcosmic terms to the root plexus of the human body, associated as much with the rectum as the genitals.35 In the Kaula yoga of the Sats hasrasamhit . . a . a
Jayaratha provides picuvaktra as a synonym for yoginvaktra in commenting on Tantr loka 15.206. a On the concept of the adhovaktra, see T ntrik bhidh nakoa, vol. i, 110, and Marc Dyczkowski, The a a a s a Canon of the Saiv gama and the Kubjik Tantras of the Western Kaula Tradition, 6365. The expression a p t lavaktra is found in Tantr loka 15.206a, adhovaktra in 6.193c, and the mouth of the yogin in 29.40d, aa a 29.124d, and 29.221d. In all likelihood, the idea was already present in Kaula scriptural sources. Dyczkowski, for instance, cites a suggestive passage from the Cicinmatas rasamuccaya which speaks of a a a . the amn yahrdaya (heart of the lineage) as located in the mouth of the Yogin. Canon of the Saiv gama, 168 (n. 49). 35 In the Pratyabhij hrdaya, Ksemar ja speaks of the bodily sakti as the middle n d extending a . a a. . from the brahmarandhra to the lower mouth (adhovaktra) (s . . . a brahmarandhr t adhovaktraparyantam a a . pr naaktibrahm srayamadhyaman drupatay pr dh nyena sthit ). The lower mouth/orice cannot of a. s a a. a a a a course refer here to female genitalia, for the central n d would then not exist in males. Jayaratha, coma. menting after Tantr loka 6.194ab, refers to the adhovaktra as where the ap na vital wind has its resting a a point, the orice which removes the delement of duality. He explicitly identies this with the yoginvaktra. The imagery of waste removal suggests the anus, although perhaps also the urethra:
34 33

a yatra n m p nasya vir ntis tad idam dvaitakalank paham adhovaktram a a a sa . . . sas. hasrotorupam yoginvaktram ity ucyate . . . . .t [The place] where in fact the ap na [vital wind] rests, the lower mouth by which the a delement of duality is removed, is called the yoginvaktra, which takes the form of the sixth [scriptural] stream. The association between the area of the rectum and the ap na v yu, which is responsible for food intake a a

247 as well, the downward netherworld face (p t lavaktra) or picuvaktra is correlated aa a with the yogic bodys adh ra- or root cakra.36 David White has suggested that picuvaktra means cotton mouth, his hermeneutics of literal readings nding in it a synonym of vagina.37 (Cotton, apparently, implies pubic hair.) Extrapolating from this, he renders picu itself as vagina, interpreting the title Picumata to mean Doctrine of the [Nether] Cotton Mouth.38 This seems highly improbable. More sound, but still unlikely, Marc Dyczkowski suggests that the title Picumata means the Doctrine of Picubhairava, the latter deity being taught in BraY lxviii.39 More probably, a Picubhairava is an apotheosis of the title Picumata, in much the same manner as Y malabhairava or Y male vara, a deity expounded in BraY lxxi. a a s a Whatever may be its basic, material referent, the BraY also uses picu as a mula tivalent technical term (samj ) some distance removed from the picu substance. As . a such, the title Picumata invokes a range of embedded homologies, from mixed malefemale sexual uids to the supreme Goddess. Several discussions of the term picu
and the elimination of waste (Svacchandatantra 7.307ab: praveayed annap nam tan malam sr vayed adhah), s a . . a . is made explicit in Svacchandatantra 7.316a: pr nap nam gude dhy yet, one should visualize/meditate a. a . a on the ap na-wind in the anus. Cf., e.g., the Goraksaataka (Briggs edition): a . s hrdi pr no vasen nityam ap no gudamandale | a. a . . .. sam no n bhidee sy d ud nah kan. hamadhyagah 34 a a s a a . .t . Pr na would always remain in the heart, ap na in the area of the rectum; sam na would a. a a be in the navel area, ud na within the throat. a It seems certain that Jayaratha and Ksemar ja identify the adhovaktra as neither the rectum nor gena . itals, but rather the plexus associated with both located at the base of the torso, which they call the janm dh ra, the root of birth. Jayaratha makes this identication explicit commenting on 3.95ab, rea a marking, trikonam ity anena yoginvaktr parapary yajanm dh rarupatvam apy asya sucitam | tata eva hi par a a a a a . saktir udetiiti bh vah | Jayaratha elsewhere refers to the janm dh ra as the place of the arising of a . a a the sakti, commenting after Tantr loka 5.94ab and 15.104ab. He also provides as synonyms the terms a a a mul dh ra, kulamula (e.g. ad 5.94ab), guhya and guhyacakra (plexus of the privies, e.g. ad 29.88), and mulasth na (place of the root, ad 32.35c), etc. Cf. Netratantra 7.3132ab, which provides a list of a synonyms for the base plexus. It is perhaps worth noting that the later Sivasamhit speaks of a yoni (vulva, womb) located between . a the penis and rectum as the locus of the kundalin: .. gud d dvyamgulata cordhvam medhraik ngulatas tv adhah | a s a . . . . ekam c sti samam kandam samant c caturamgulam 77 a . a . . . pacim bhimukh yonir gudamedhr ntar lag | s a a a . a tatra kandam sam khy tam tatr ste kundal sad 78 a a . a a . .. Sivasamhit 5.7778 (Lonavala edition). . a 36 J. A. Schotermann, The Sats hasrasamhit , Chapters 15, 8687. . . a . a 37 Kiss of the Yogin, 101; on literal readings, see ibid., 78. 38 Kiss of the Yogin, 101. 39 a Canon of the Saiv gama, 16870 (n. 56).

248 occur in the BraY s second half (satka), prior to the Addendum Tantras, particularly a . . in chapters lxvi, lxviii, and lxxi. In the last section of lxxi, the Dev asks, Why does this tantra have the designation picu? What is stated by the syllable pi, and the syllable cu, O Mahe vara? [And what] through the conjunction of the two? Tell s [me] everything. Bhairava in answer embeds in this two-syllable word a series of doctrinal and ritual meanings. Within the syllable pi is present the supreme Siva, for the entire universe fell (pap ta), i.e. came to pass, through [his] contact with the sakti.40 a In union with Siva, the supreme Sakti is cu, for she kisses (cumbana) and sucks (cusana). Their orgasmic discharge (ksobha) is the great substance (mah dravya), a . . . the seed [of creation?]. Since the entire world is composed of siva and sakti, the term picu, furthermore, connotes supreme nonduality (par dvaya). A subsequent, a problematic line apparently describes picu as bloodpresumably menstrual blood or female sexual uidas well as cognition (buddhi), and the yoga of meditation upon Siva.41 Picu is nothing other than Siva and Sakti; it is the nectar of immortality, in which all substances have origin. Picu has the nature of gnosis, the ever-exalted Ucchusmabhairava. Pi has the form of the penis, while cu denotes male and female . sexual uids.42 In chapter lxvi, picu is described as the union of Siva and Sakti,
It is possible that the verb pap ta is used in allusion to sukrap ta, seminal emission. a a This linepicu raktam tath buddhi dhy nayogam siv tmakam, lxxi.117abis highly problematic. It a a a . . . is conceivable that buddhi is a corruption of viddhi, the imperative of vid. 42 BraY lxxi.11021ab: a
41 40

devy uv ca a picusamj mah deva tantrasy sya katham bhavet | a a . a . kim v pik ram ity uktam cuk ram v mahevara 110 a a a s . a . ubhayor yogayogena kathayasva samastakam | bhairava uv ca a srnu devi pravaksy mi picupran rtham uttamam 111 . . s a . a pratyaksaravibh gena yath vastham var nane | a a a . . pap ta saktiyogena yasm t sarvam car caram 112 a a a . pik rastham sivam vindy d vy pakam vibhur avyayam | a a a . . . cumbanam cusanam yogam sarvagam sivayojitam 113 . . . . . . sa saktih kurute yasm c chivayog d var nane | a a a . picusamj cal sambhuyukt prokt var nane 114 a a a . a a ubhayor bhogayog c ca ksobho yah param rthatah | a a . . . picuvad bjavaj jeyam mah dravya ca p rvati 115 a a . sivaaktir na c nyam hi sarvatr pi ca drsyate | s a a . . vy pitva-m-aprameyatve picusamj par dvaye 116 a a . a

249 taking the form of the two-syllable mantra hum phe.43 The syllable pi connotes the . male genitals and semen, and cu menstrual/sexual uid and the female hole or vessel (kunda). In addition, picu is the supreme Sakti, Aghore var, from whose s ..
picu raktam tath buddhir dhy nayogam siv tmakam | a a . . a picu saktiiv bhy m tu n ntaram varavarnini 117 s a a. a . . yogatv t sarvabodh c ca sarvasattvavivecan t | a a a nityayuktam picuj nam s rvvannaikopam sthitam 118 a . a a . . amrtam mrtyun sam tu sarvadravyam picudbhavam | a . . . . . picu j nasvabh vam tu ucchusmam satatoditam 119 a a . . . pik ro lingasamsth nam cuk rah sukraonitam | a a . a . s . . picu-r-aksarasamyog suddh suddhavilaksana 120 a a . . . . antahkaranake b hye saik saktih siv nvit | a a a a . . .
111d pra n ] corr.; prasn Bya s a a 114a saktih ] corr.; sakti Bya yasm c ] em.; yasy c a a . Bya 115a bjavaj ] em.; bjava Bya 116a saktir ] em.; sakti Bya 118a yogatv t ] em.; yogatv a a Bya bodh c ] em.; boddh c? Byapc ; ( - - ?) Byaac 118b sattva ] em.; sattvam Bya 119a a a . n sam ] em.; n sam Bya a . a . 119d satato ] em.; satato Bya 120b sonitam ] em.; sronitam . . . Bya 120d vilaksana ] em.; vilaksanam Bya 121b siv nvit ] em.; siv nvit m Bya a a a a. . . . . .

The Goddess spoke: O Mah deva, why does this tantra have the title Picu? What is a stated by the syllable pi, and what by the syllable cu, O Mahe vara? [And what] by s joining them in combination? Tell me everything. [11011ab] Bhairava spoke: Hear, O goddess, the ultimate answer to the question about picu, in accordance with its state with the division of each syllable. Since the entire universe came to pass [lit. fell] through [his] union with the Sakti, the pervasive, immutable lord, Siva, should be known as present in the syllable pi. [111cd13ab] Since, in union with Siva, the Sakti kisses (cumbanam) [and] sucks (cusanam), () [and creates a] union that is . . . . all pervasive, conjoined with Siva (?), the immovable [Sakti] is said to have the title picu when united with Siva, O fair woman. [113cd14] And the orgasm uid (ksobha) from . their combined pleasure, () composed of picu, should in truth be known as containing the seed [for the creation] (?), the Great Substance (mah dravya), O P rvat. And a a everywhere verily are seen siva and sakti, nothing else. Because of being pervasive and immeasurable, the term picu has the sense of supreme nonduality (par dvaya). [11516] a () Picu is [sexual/menstrual] blood and cognition (buddhi); yoga and meditation consist of Siva (?). Picu is none other than siva and sakti, O fair woman. [117] Because of its state of unity, because of omniscience, and because of discriminating knowledge of all beings, () . . . . . . (?). [118] It is the nectar which destroys death; all substances originate from picu. Picu has the nature of wisdom; it is the ever-arisen Ucchusmabhairava. [119] The . syllable pi has the form of the penis; the syllable cu is semen and female sexual/menstrual uid (onita). Possessing the conjoined syllables pi and cu, having both pure and impure s . characteristics, the Sakti is singular, both internally and externally, conjoined with Siva. [12021ab] The text and interpretation offered are provisional. Note for example the problem of what Bya transmits as picuvad bjava in 115cd. This should probably read picuvad bjavaj, the -vat sufx having the sense of possessing, containing. In 116cd and 120c, there appear to be compounds with internal hiatus breakers, vy pitva-m-aprameyatve and picu-r-aksara , respectively. 118 is particularly problematic. For 118d, a . the only conjecture I can offer is sarvam ekopam sthitam, everything, present in a single comparison. a 43 Bya is not legible here, but the form of the mantra is perhaps conrmed by the occurrence of hum . phe phat in xxvi.39ab: humphephadeti raksardham tato dutidvay ntimam. However, there is a possibility a . . . . . . that phe is corrupt for phet, a syllable whose importance is illustrated by the exposition of a deity called of Phetk rabhairava (the bhairava of the syllable Phet ) in the very next chapter, lxvii. The a manuscript evidence is mixed; in support of phe (or pheh, phem?), note for example lxxxiv.57c and 188c . . . in Bya : humpheti dvyaksaram mantra and humpheti dvyaksaram mantram, respectively, the latter probably . . . . . being the correct text for both ( pheti appears to represent phe iti).

250 womb the entire creation arose.44 Hence, the title Picumata appears to foreground the importance of sexual uids and coitus in the ritual of this scripture. The texts exegesis of this term advances a rudimentary theology of sexual ritual, embedding in picus two syllables the polarity of siva and sakti and their earthly embodiment in male and female uids. And consistent with this scriptures emphasis on the female pole of the Godhead, although picus syllables are homologous with siva and sakti,
BraY lxiv. The text of this passage has several problems, in part because the oldest codex, Bya , is a damaged. The following edition reports the readings of Bya , Byb , and Byd : humk re bhairavo devo phek rasth mahevar | a a s . a ubh bhy m yo tra samyogah saktikunde mahevari a a. s . . .. . raktaret tmako hy ekah picurupah sad vyayah | a a . . s saktir devadevasya sivabjasamanvit 7 a a aksar ksarayogena pralayotpattik ranam | a . . a . . vivotpattivibh gena saktirupam mahodayam 8 s a picusamj gatam devi lolbhutam tu sarvagam | . . a . pik ro lingam ity uktam cuk ram kunda ucyate 9 a a . . .. pi sukra cugatam raktam picuyogam bhav tmakam | a . . :::::::::::::::::::: . jvadehavibh gena saktiaktimatam picum 10 a s . pradh nam jathare s tu yayotpannam car caram | a . . a a . drs. am saktimayam devi tena saktir udryate 11 ..t . . cint manisam devi jagaty asminn aghorik | a . a a
k rasth ] Byd ; 6a humk re ] Bya ; humk ro Byb ; humk ra Byd 6b phe ] Bya Byb ; pha Byd a a . a . a . a a Byb a ; srotasamyog c By 6c yo tra samyogah ] conj.; (yo[tr]a[y/s/g]am[y/g][ /o]?) By a a . . . . Byb ; y tra samyogo Byd a 6d sakti ] Bya Byd ; chakti Byb kunde ] Bya Byb ; kunda Byd 7a . .. .. ret ] Byb Byd ; ( - ?)et Bya . a a ekah ] Bya Byd ; eka Byb 7b rupah ] corr.; rupa Bya Byb Byd 7c . s ] Bya Byb ; so Byd a saktir ] em.; sakti Bya Byb Byd 7d samanvit ] em.; samanvit h a a. Bya Byb Byd 9a samj ] Byb Byd ; saj (anusvara perhaps obscured) Bya a 9b lolbhutam . . a tu ] Byb Byd ; lol( - - tam - ?) Bya 9c pi ] Byb Byd ; pa Bya 9d cuk ram ] corr.; cuk ra(n?) a . a . Bya ; cuk ra Byb ; cuk rah Byd a a . 10a pi sukra cu gatam ] em.; pi sukra(m -?)gatam Bya ; pi sukra . . . tu gatam Byb ; pi sukram vugatam Byd raktam ] Byb Byd ( - ktam) Bya 10b picuyogam . . . . . . bhav tmakam ] conj.; picu([p/y][ /o] - )bhav tmak( - ?) Bya ; picuy gam tav tmakam Byb ; a a a a . a picuy gam bhav tmakam Byd 10c jvadeha ] Byb Byd ; j( /?)va( - - - ) Bya 10d picum ] Byb Byd ; a . a a . pi( - )m Bya 11a jathare ] Bya Byb ; jathara Byd tu ] Byb ; tuh Bya Byd 11b yayotpannam ] conj. . . . . . Isaacson; yathotpannam ] Byb Byd ; yathotp( - - )m Bya 12 cint manisam ] Bya Byb ; cint manisamo a a a . . . . Byd 12b aghorik ] Bya ; agh rik Byb ; aghy rik Byd a a a a a k rastho a ::::::::::::::: 44

In the syllable hum is the god Bhairava; Mahe var is present in the syllable phe. () The s . conjunction by both here, in the hole of the consort (?), O Mahe var, as one, has the form s of the everlasting picu, which consists of female uid and semen. [67ab] It is the Sakti of the god of gods, possessing the seed of Siva. Through combination of the respective syllables, it is the cause of the universal creation and destruction. [7cd8ab] Through the divisions of the creation of the universe, it has the glorious form of the Sakti, located in the designation picu, vibrating and omnipresent. [8cd9ab] The syllable pi means penis; the syllable cu means the hole (kunda). pi is semen; female uid is present in cu. () .. The combination pi-cu comprises the universe (bhava) (?). [9cd10ab] Picu is [both] sakti and the possessor of sakti, through the division of the soul (jva) and body of living beings. She is Prakrti (pradh na), by whom was born from the womb the animate and inanimate a . universe. [10cd11ab] Everything seen consists of sakti, O goddess. That is why the Sakti, Aghor, is said to be like a wish-fullling jewel in this world, O goddess. [11cd12ab] The interpretation offered of this somewhat obscure material is again provisional.

251 as an integral unit it seems primarily to embody the supreme Saktishe who gives birth to the universe. Interestingly, the term picu occurs in mantras in several Buddhist yogintantras: the Candamah rosanatantra, Krsnayam ritantra, and Hevajratantra. It features promia . . a .. .. . nently, for instance, in the latter source in the heart mantra (hrdaya) of Hevajra: om . . . . . deva picuvajra hum hum hum phat svaha.45 According to Isaacson, commentators . on the Hevajratantra understand picu to refer to ne cotton, with Hevajras epithet picuvajra indicating his simultaneously soft or grace-bestowing nature and his hard, vajra-nature as destroyer of obstacles and evil.46 Outside of mantras, I am unaware of the occurrence of picu as a technical term in Buddhist tantric sources, wherein it thus appears anomalousperhaps inherited from teachings no longer present in surviving scriptures, or drawn from Saivism, even the tradition of the Picumata itself.

. 5.3 Navaksaravidhana: the Method of Nine Syllables

Among the secondary epithets or titles of the BraY is Nav ksaravidh na, the Proa a . a cedure/Method of the Nine Syllables of the vidy -mantra of Aghore var. This a s title foregrounds the pivotal function of the vidy -mantra, the mantra-form of the a supreme Goddess herself, in cementing this texts doctrinal vision and vast corpus of ritual into a coherent whole. Whether in accounts of the cosmos, mantra, or ritual, the vidy of Aghore var serves as the organizing principle which patterns and a s establishes order among these intersecting domains. As the sonic embodiment of the Goddess, the vidy , moreover, encodes the core pantheon of the Brahmay mala and a a . the mantras for its systems of practice. Its syllabic form is [om] hum cande kapalini . .. . svaha, the nine syllable-deities accordingly being hum-cam-de-ka-pa-li-ni-sva-ha: . . Bhairava; Rakt , Kar l, Candaks, and Mahocchusm the Four Devs or guhyak s; a a a . . . . a and Kar l , Dantur , Bhmavaktr , and Mah bal the Four Duts (Consorts), kiaa a a a a nkars, or anucars. A syllabic series based upon the vowels of the vidy embodies the a
45 46

Hevajratantra I.ii.5. Unpublished annotation to Ratn kara antis Bhramaharan ma Hevajras dhana. a s a a

252 Eight Mothers, while the Six Yogins comprise a set of inected forms of the vidy . a As the nine-syllable whole, the Great Goddess Aghore var or Bhairav subsumes all s (table 5.1). Vidh na and its synonym a vidhi connote in this context the processes of ritual. Procedures for mantra-incantation and consecration are, for instance, called japavidh na and a abhisekavidhi, the subjects of . chapters xviii and xxxiii, respectively. The vidh nas of a
Table 5.1: Deities of the Nav ksar Vidy a . a a

four devs i Rakt (cam) a . Kar l (de) a . Candaks (ka) . . . Mahocchusm (pa) . a six yogins i Krostuk (om hum cande .. .. . . kapalini namah) . Vijay (om . . . svaha) a . . Gajakarna (om . . . hum) . . Mah mukh (om . . . vausat) a . . . Cakraveg (om . . . vasat) a . . . Mah n s (om . . . phat) a a a . .

i four duts Kar l (li) aa Dantur (ni) a Bhmavaktr (sva) a Mah bal (ha) a a eight mothers Mahe var (am) s . Br hm (e) a Vaisnav (a) .. Kaum r (a) a Vaivasvat () i Indr n (i) a. Candik (sva) .. a Aghor (ha)

the BraY are patterned by a mantra-congurations repre-

senting so many inections of the Nine, their uninected

a conguration being the basic/root pantheon or mulay ga, consisting of Bhairava Bhairav, the Four Goddesses (dev), and the Four Consorts (dut). These pattern the entire gamut of ritual, from re sacrice to vratas to yoga. In ny sa, for instance, a installation of the mantra-deities upon the body and other substrates preliminary to almost all ritual, the practitioner or ritual object embodies the mantra-deities in various congurations. Among the fundamental rituals of the system are the vidh nas a a of the navay ga, nine pantheons, these being the mulay ga and eight other congua rations of the pantheon.47 Vratas or observances are also ninefold, or vefold, corresponding to the mantra-deities of all nine syllables or the Four Devs with Bhairava, respectively.48 Other y gas are based upon different and extended congurations of a
The Nine Y gas form the subject of chapter thirteen. The opening verses of chapter 3, the a mah y gapatala, provide moreover names for each of the Nine. a a . 48 The nine and ve vratas form the primary subject of chapter xxi. Among these, the fth of the

253 the vidy s mantra-deities.49 Cakras or mandalas drawn upon a substrate, usually the a .. ground, provide templates for these pantheon congurations and loci for many of the vidhis connected with them. The nav ksar vidy -mantra comprises the subject of the second chapter of the a . a a BraY , which follows the narrative of revelation (BraY i.1119) and a short exposition a a on the supreme Sakti (i.12033).50 The vidy s mantric form and embedded pantheon a were rst identied by Sanderson: . . . the essential components of the mantras of the nine deities who form the core of the greater mandala and are the pantheon of daily worship are .. . the syllables of the mantra of Canda K p lin: (om) hum cande kapalini . .. . . a a . a a svaha (. . . O Canda K p lin . . . !). Thus Kap lsabhairava (hum), his a .. four goddesses (Rakt (cam), Kar l (de), Candaks (ka) and Mahocchusm a aa . . . . . . a (pa)) and their four attendant powers or Duts (Kar l (li), Dantur (ni) a a Bhmavaktr (sva) and Mah bal (ha)) are aspects of a feminine power a a a which transcends the male-female dichotomy which patterns the lower revelations.51 This account of the vidy requires correction in one detail: the second Dev is Kar l, a a while Kar l is rst of the Duts.52 The confusion is understandable, for these deities aa of similar nomenclature are frequently mixed up in the manuscript evidence.53 Besides the core pantheon of the Nine, the vidy is said to contain within itself a pana
Five seems most important: the great Observance, called also the Bhairava-observance and observance of the vidy . This probably corresponds to the ekavravidh na, procedure of the solitary Hero a a a [Bhairava], an important inection of the mulay ga in which the focus is Bhairava alone and not the coupled divinity (y mala). a 49 BraY xxv in particular, the y ganirnayapatala, teaches extended inections of the basic pantheon. a a . . 50 Note that BraY ii has been included in the critical edition forming the latter part of the present a dissertation. 51 Saivism and the Tantric Traditions, 672. Sanderson also provides the vidy on the basis of a a passage from BraY lxxxv (verses 4243ab in his numbering), in History through Textual Criticism, a 4446. 52 Among the numerous lists, see for example a concise one in chapter 4, quoted in full in the annotation ad BraY 1.8cd9; and another passage from the same chapter (verse 262): a rakt y s tu kar l sy t kar ly y s tu dantur | a a aa a a a a a . a bhmavaktr tu candaksy ucchusm y mah bal | a a a a a .. .
kar l ] em.; kar lam Bya aa a .

[The attendant] of Rakt is Kar l , but of Kar l, Dantur ; of Candaks, Bhmavaktr ; of a aa a a a . . . Ucchusm , Mah bal . a a . a Note that the genitive kar ly y h of kar l occurs with great frequency in the Brahmay mala. This formaa a a. a a tion is similar to the locative ujainy y m in i.81a, discussed in the footnote thereon. a a 53 For instance, in the oldest ms, Bya , chapter lxi (13342) lists Kar l as rst of the Duts, while for a the second Dev, it provides kar l y [h], the genitive of kar l . There is no metrical reason why these aa a . aa

candaksy ] em.; candaks Bya . . . a . . .

254 theon augmented by a sextet of yogins and the Eight Mother goddesses (table 5.1). The Six Yogins emerge from the limbs of Aghore var,54 although as the anga s mantras of the vidy ,55 these deities appear distinct from the anga-mantras of the a Goddess.56 Somewhat less natural is the derivation of the Mother Goddesses (m tr) a. i from the vidy : the syllabic series am-e-a-a--i-sva-ha, which maps loosely to the a . vowels of the vidy . a Evidence conrming Sandersons reconstruction of the vidy -mantra is ample. a . The point of least clarity concerns its rst syllable, hum, the seed-mantra (bja) of Kap lsabhairava. On the evidence of chapter two alone, it might appear that the a vidy begins with om, and that this is the seed-mantra of Bhairava; ii.15a states that a . the God [Bhairava] exists in the pranava,57 and the chapter gives no indication of . pranava having a sense different from its normal referent, om. There are, moreover, . . mantras in the BraY which begin, as would the vidy , om cande . . . .58 However, a a . .. . elsewhere the stated form of the vidy clearly begins with hum. Note for exama ple lxxxviii.11315, where we nd a mantra-installation (ny sa) with the sequence a . hum-ca59 -de-ka-pa-li-ni-sva-hathe nine-syllable vidy . It seems in fact likely that a . . pranava in BraY ii.15 refers to hum. This possibility is illustrated by the Svacchandaa . . tantra, which appears also to attest reference to hum, the seed-mantra of Svacchandcould not be emended to kar l and kar ly y [h], respectively, and the confusion is probably scribal. In aa a a a . BraY x, which concerns the mantroddh ra of the goddesses, Bya is inconsistent: three of the occurrences a a of the names in question imply the order suggested by Sanderson (41b, 51a, and 86b), while two follow the pattern I believe to be correct (47d and 81b). 54 BraY ii.14ab. a 55 On the concept of anga or ancillary mantras, see T ntrik bhidh nakoa, vol. I, 9395. a a a s 56 Although in theory the vidy -mantra is Aghore var, an important distinction appears between the a s two as well. The supreme Goddess is both embodied by the vidy and appears as a subset within itas a eighth of the Mother goddesses, the m trpuran (she who makes the Mothers complete; cf. BraY a . . a ii.18b, etc.). The latters root mantra, as given in BraY x, is om ha namah, and her anga- and other a . . ancillary mantras are based upon this, rather than the full vidy (see BraY x.210cd215ab). It hence a a appears that two levels of being are posited for the Goddess, the higher of which comprises the vidy a of nine syllables. 57 pranave tu sthito devo. . 58 Cf. seven mantras provided in the prose following lxxxiv.55, beginning with om cande mohane s . .. . vari ghatani hum phat svaha. . 59 Note that cak ra is provided instead of camk ra, despite the latter occuring in the vidy . This might a a . a not be a corruption; although cam is provided in the samkalanavidhi (quoted below), ca occurs as the bja . . of Rakt used for ny sa in x.24a. a a

255 abhairava, as pranava.60 In the Brahmay mala, furthermore, reference is made to the a . . Bhairava-pranava;61 and this surely refers to hum, for the context is installation of . the mandala deities of the vidy onto the body, beginning with Bhairava. Yet while a .. . the vidy proper hence begins with hum, in many of its applications, om is nonethea . less prexed to the nine syllables. This reects standard mantra formation in this system;62 a mantra not preceded by om is said to be headless.63 . As the seed-mantra of Kap lsabhairava and the initial syllable of the nav ksar a a . a . vidy , hum possesses particular signicance in the Brahmay mala. It is given the a a special designation smarana, said to be formed by conjoining the sixth vowel (u) and . the candrabindu (m) to the hamsa (ha).64 One is to afx the smarana to all mantras, . . .
60 Commenting ad Svacchandatantra 4.203, Ksemar ja glosses pranava as niskalan tha, which as Isaaca a . . . . son points out to me refers to hum, the niskala (undifferentiated, unmanifest) form of Bhairava in this . system (e-mail communication, January 29, 2006). That the term can have multiple applications is evident also in the teaching of paca pranav h or ve pranavas in Svacchandatantra 6, discussed by Ksemar ja a . a. . . ad Svacchandatantra 6.3. It seems that these involve permutations of the ve constituents of om, along . with the hamsa (ha). . 61 BraY xii.60cd: sikh y [m] pranavam nyasya bhairav khyam na samsayah. a a a . a . . . . . 62 Cf., e.g., BraY lxxxv, which opens with the mantra om hum phat vausat kapalabhairavaya a . . . . . namah. . 63 Brahmay mala xi.16cd19ab: a

kotayah sapta mantr nam omk rasya na samsayah 16 a. . . . a . . kimkaratvam prakurvanti evam vai bhairavo bravt | . . . sarvesam mantraj tn m mastako pranavah smrtah 17 a a. . . . . . . yatr dau na bhavaty esa amundo mantra ucyate | a . .. n nena tu vin mantrah sidhyateha kad cana 18 a a a . ten dau pranavah proktah sarvamantresu suvrate | a . . . .
a. 16c mantr nam ] em.; mantr na Bya a. 18c n nena ] em.; n mena Bya a a 18d cana ] em.; canah Bya .

The seventy million mantras are undoubtedly subordinate to om. Thus did speak Bhairava. . The pranava is called the head of all classes of mantra. When it is not present at the be. ginning, the mantra is called headless. A mantra never bears fruit in this world without it; that is why the pranava is taught at the beginning of all mantras, O pious woman. . That the pranava is here om is made explicit in the preceding verses. . . 64 Cf. xi.36cd39ab: hams khyam satatam bjam niskalam tu samuccaret 36 . a . . . . . adyasvaram tath caiva uccaren n tra samsayah | a a . . . ekkrtam mah devi hak rah parikrtitah 37 . . a a . . sas. hasvarena samyuktam kartavyam tu mahevari | s . .t . . . . ardhendubindukalay samyuktam k rayet tatah 38 a . . a . smaranasamj sam khy t sarvamantrevarevarah | a aa s s . . a .
36c hams khyam ] em.; hans khy Bya a a 37a svaram ] corr.; svaram Bya . a . . . hak rah ] em.; hak ra Bya parikrtitah ] em.; parikrttitam Bya 38a svarena ] svarena Bya a . a . . . . smarana ] em.; smaranah Bya akhy t ] em.; akhy t h Bya aa a a. . . . 37d 39a

One would pronounce [the mantra] called hamsa constantly, but [just] the seed-syllable . [h] without a vowel. One would likewise utter the rst vowel [a], no doubt. Made one, O

256 particularly those of the BraY .65 Knowing the syllabic content of the smarana, the a . form of the nav ksar vidy can be conrmed from multiple other passages. In BraY a . a a a . lxxxv, for example, we nd the nav ksar vidy headed by hum and called the Heart a . a a of All Saktis: Possessing the smarana, with the word (pada) cande [afterwards] in the .. . beginning, ending kapalini, and decorated [at the end] by the word svahathis is the Heart of All Saktis, bestower of the fruits of supernatural experience and liberation.66 BraY xxxiv offers further evidence conrming the form of the vidy -mantra.67 a a
Mah dev, this is known as the syllable ha. One should conjoin this with the sixth vowel a [u], O Mahe var, and then make it joined with the crescent moon and dot [i.e. candras bindu]. The term smarana has been taught, the lord of all Mantra-lords. .

BraY xi.75cd76ab: a

y van mantro mah devi asmi[ms] tantre viesatah a a s . . . smaranasamputitam k ryam khecaratvajigsinam | . . . . . . a . As many mantras as there are, particularly in this tantra, those aspiring to become skytravellers must frame with the smarana. . The masculine mantro appears to be a collective singular, in agreement with the neuter putitam k ryam . . a . and y vat. Cf. y van mantragano (em.; ganau Bya ), xxxviii.44c. a a . . 66 BraY lxxxv.42cd-43: a smaranena sam yukta candety dipad nvitam 42 a a . .. a . k p linyantasamyuktam sv h padavibhusitam | a a a a . . hrdayam sarvaaktn m bhuktimuktiphalapradam 43 s a. . .
sam yukta ] em.; sam yukta Bya a a
saktn m a a a. . ] corr.; saktn m By

This passage is cited and discussed by Sanderson, in History through Textual Criticism, 4446. 67 This extensive chapter, entitled Mantrasamkalanavidhi, teaches the prelimary ritual for preparing . or empowering (samkalana) the basic mantras of the system, apparently necessary for entitlement to . a a . ac ry bhiseka, consecration as an ofciant. Its instructions begin thus: atah param pravaksy mi mantrasamkalan vidhim | a . . . a . yena vij tam trena dksam vai kartum arhati 1 a a . . . pranavam coccaren mantr punah smaranam uccaret | . . . . punah smaranam ucc rya punah pranavam uccaret 2 a . . . . sv h k r ntasamyuktam ahutyaik m tato hunet | a a aa a. . punah smarana+m ucc rya+ punah pranavam uccaret a . . . . puna[h] smaranam ucc rya sv h k r ntakam hunet | a a a aa . . .
1b samkalan vidhim ] em.; samkalan vidhih Bya a a . . .

Next, I shall further teach the procedure of preparing the mantras, upon the mere learning of which one is worthy to bestow initiation. A mantrin should rst utter the pranava, . then utter the smarana; then after again uttering the smarana, then again utter the pranava. . . . [After uttering the vidy ] conjoined with svaha at its end, one should offer sacrice of a one oblation. After again uttering the smarana, one should again utter the pranava. After . . again uttering the smarana, one should sacrice, [uttering the vidy ] ending with svaha. a . (It seems that 3b would mean ek m ahutim tato juhuy t. Note the optative verb hunet for juhuy t, a a a a formation, inuenced by Middle Indic verbal systems, based on the non-standard, non-reduplicated

257 References to the smarana, which appears distinctive to and probably has origins . in the system of the BraY , also occur elsewhere in Saiva literature. The Bhairavamaa a ngal refers to the eightfold smarana as originating from the Picutantra.68 In the . Tantr loka, Abhinavagupta too makes reference to the smarana, in a citation from a . the BraY xi on the subject of expiation ritual (pr yacitta).69 a a s Sanderson, more-

over, shows that a reference to the smarana has found its way into the Buddhist . Laghuamvaratantra, in a passage redacted from BraY lxxxv, for which the texts s . a commentators vainly endeavored to offer a plausible interpretation.70 Despite the central importance of the nav ksar vidy and its components, a numa . a a ber of alternate or parallel congurations of mantra-deities exist within the BraY , a which while correlated in various manners with the vidy and its pantheon cana not be said to derive from it. Such for example appears to be true concerning the kula- or khecarcakra, which is connected with the kulavidy (Vidy -mantra of the a a Goddess Clans) rather than the nav ksar vidy .71 A number of the deities taught a . a a in separate chapters (often called kalpas), furthermore, have only loose connections a a a with the nav ksar vidy and its pantheon; these include Mah k la,72 Gart bhairava,73 a . a a
present indicative hunate for juhoti. In the BraY , I note one occurrence of hunate, in xlix.7a; none of a juhoti; nearly two-hundred instances of hunet; and eleven of juhuy t.) In this inimitably tedious style, a which continues for fteen folios, the text then instructs the following sequence: smarana cam . . smarana svaha oblation; smarana cam smarana svaha oblation; smarana de . . . . . . smarana svaha oblation; smarana de smarana svaha oblation. This patterns continues . . . . with the remaining syllables of the vidy , ka-pa-li-ni-sva-ha, and probably implies om hum cande a . . .. kapalini svaha as the full form of the mantra. 68 a Bhairavamangal 235ab: smaranam as. adh j tv picutantr t samudbhavam ( tantr t samudba a . . .t a a a a havam ] em.; tantr samudbhavam cod.). Cf. Bhairavamangal 242: a a avadhut tu s saktir svar khy mayodit | a a a a nir c rah sivo jeyah smaranatve vyavasthitah a a . . . .
saktir ] em.; sakti ms

sivo ] corr.; sivo ms

s a a The Avadhut Sakti is the [same] one I said is called I var . Siva is known as nir c ra, a a existing as the smarana. . a nak 5-687, f. 13v; electronic transcription courtesy of Somadeva Vasudeva. Regarding avadhut and nir c ra, see the annotation on BraY i.3639 and ii.2 in the critical edition. a a a 69 Abhinavaguptas citations of the BraY are discussed in appendix a. a 70 History through Textual Criticism, 4446. 71 The kulacakra or khecarcakra forms the subject of BraY xiv. Its basic form comprises a conguration a of sixteen vowel-goddesses as n ds around Bhairava in the center, in which it resembles the bhautikacakra a. of BraY xix. a 72 Mah k la and his mandala of the Eight Mother goddesses are the subject of BraY liv. a a a .. 73 Chapter lxix teaches this Subterranean Bhairava and his mandala of four goddesses. ..

nir c rah ] corr.; nir c ra ms a a . a a

258 Hairambhabhairava (Gane a or Ganapati as Bhairava),74 and Utphullakabhairava.75 . s .

5.4 Mulatantra: the Root Scripture

In the revelation narrative of chapter one and at several others points in the text, the BraY refers to itself as the mulatantra, the Root Tantra or Root Scripture.76 Yet a although claiming the status of the foundational, originary scripture, it also describes itself as originating from a text of 125,000 verses called the Vimala or Unblemished, the primordial font of scriptural wisdom (j na). These claims are embedded within a a theology of revelation that posits scriptural wisdom as originating at the pinnacle of the hierarchy of tattvas, reality levels or cosmological principles, its descent (avat ra) a into the bounded form of texts marked by ever-increasing brevity and numerical profusion. Perhaps the earliest surviving account of the descent (avat ra) of revelation in a Saiva sources occurs in the Uttarasutra of the Niv satantra. In its simple descrips a tion, scripture (the sastra or sivatantra) emerges from the formless, supreme Siva as subtle sound (n da), which Sad siva sets down into language. This he transmits to a a the gods, whence it reaches in various redactions to the sages and mortals (figure 5.1).77 A demonstrably later scripture, the Svacchandatantra describes a more detailed genesis of scripture. In its conception, from the supreme, formless Siva emerges a
74 Taught in BraY lxxvi, this skull-bearing, elephantine Bhairava has four heads and ten arms, and a stands upon a dead rat (mah musakapreta). His is a ten-syllable vidy , and he possesses a mandala of a . a .. yogins. 75 a The subject of BraY lxxxii. While Utphullakas pantheon mirrors the mulay ga in having a ninea syllable mantra, embodying himself in a mandala of saktis, the mantra-deities are distinct. .. 76 Colophons to chapters ii, xxx, and xxxiv. Text passages describing the BraY as mulatantra are a discussed below. 77 Niv satantra, Uttarasutra 1.22cd25: s a

sivatantrasya cotpattim krtyam nam nibodha me 22 a . . adrs. avigrahe sante sive paramak rane | t a . .. n darupam viniskr ntam sastram paramadurlabham 23 a . . a . . sad sivas tu vett vai sa ca m m prati bodhakah | a a a. . n darupasya sastrasya aham granthanibandhakah 24 a . . anus. upchandabandhena devebhyah pratip ditam | a .t . rsibhya ca puna c ham tebhyo martyesu santatih 25 s s a . .. . . Text as constituted in the working draft edition of Goodall, et al.

259 s sonic resonance (tad dhvanirupam), which Sad siva transmits to I vara in the form a s of countless scriptures suitable to the variety of beings deserving of grace. I vara then teaches the tantras to the Mantra-lords and rudras of the upper reaches of the .. universe, as well as to Srkantha, via whom the teachings descend unto the various .. lower rudras. Bhairava too takes initiation from Srkantha and learns the scripture, and teaches the Goddess, who transmits tantras to the ganas, gods, Mothers, n gas, a . yaksas, etc., whence it nally reaches the sages and humankind (figure 5.2).78 While . both of these sources attest the notion of a subtle, sonic original wisdom, neither however applies to this or its rst textual evolutes the term mulatantra. The BraY s conception of the genesis of scripture is considerably more complex a (figure 5.3). It in fact posits a hierarchy of levels of the originary scriptural wisdom: that of an undifferentiated totality, that of a scripture called the Vimala, and that of a Root Tantra. In some accounts, these appear to be distinct levels or stages of the primordial wisdom in its descent into the texts of the canon, while other passages appear to conate one or more of them. Chapter one describes the supreme

Svacchandatantra8.28cd39 (ksts edition):

. tad ev pararupena sivena param tman 28 a a a mantrasimh sanasthena pacamantramah tman | a a . a purusartham vic ry su s dhan ni prthak prthak 29 a a a a . . . . laukik diiv nt ni par paravibhutaye | a s a a a tadanugrahayogy n m sve sve visayagocare 30 a a. . anus. upchandas baddham kotyarbudasahasradh | a a .t . . guruisyapade sthitv svayam devah sad sivah 31 s. a a . . . a purvottarapadair v kyais tantram adh rabhedatah | a . taj j nam svare d t tad svarena sivecchay 32 a a a . vidy y h kathitam purvam vidyeebhyas tath dar t | a a. s a a . . m y niyatiparyantais tasm d rudrair av pi tat 33 a a a a srkan. henevar t pr ptam j nam paramadurlabham | .t s a a . a . ten pi tadadhah proktam rudr nam svarecchay 34 a a . a . . pradh n c chatarudr ntam dksayitv vidh natah | a a a a a . . . mam pi ca pur dksa tath caiv bhisecanam 35 a a . a a . srkan. hena pur dattam tantram sarv rthas dhakam | .t a a a . . may pi tava devei s dhik ram samarpitam 36 a s a a . tvam api skandarudrebhyo dadasva vidhipurvakam | brahmavisvindradev n m vasum trdiv krt m 37 a a. a. a . a . loke samgrhya n g n m yaksanam paramevari | a a a. s . . .. . kathayasva rsnam ca rsibhyo manujesv api 38 ... . .. . evam tantravaram divyam siddharatnakarandakam | . . . .. tvay guptataram k ryam na deyam yasya kasyacit 39 a . a . . A Nepalese codex, nak 1-224, in 37ab reads tva c pi ganarudr nam dadasva siddhipurvakam; and in 38ab, a a. . . lokap l grah m n g m yaksanam paramevari. aa a. a a. s .. .

Figure 5.1: The Origin of Scripture (Sivatantrotpatti) according Niv satantra, Uttarasutra 1.2325 s a


n"dar#pa$ %"stra$


The Gods The Ten !ivatantras The Eighteen Rudratantras



Tantras by the thousands

Sakti, Sivas volition, as awakening bindu, the primordial substance of cosmic sound, from which emerges the undifferentiated totality or ood of scriptural wisdom (j naugham niskalam). This takes the form of a body of mantra (mantravigraha) a . . . situated at the level of the sad sivatattva in the pure universe.79 At the time of crea ation, Sad siva causes the Vimala to emerge forth from this, therefore representing a the rst descent of the primordial wisdom to the level of language, taking shape as a text of 125,000 verses in the anus. hubh or sloka meter.80 This primordial text, .t the font of all scriptureeven of the universe, in some descriptions81 also exists at the level of the sad sivatattva. Thus, just as does the supreme deity, the mass or toa tality of scriptural wisdom (j naugha) possesses a higher, undifferentiated (niskala) a . and a differentiated (sakala), manifest form, the latter being the Vimala. It is from
79 80

BraY i.35. a BraY i.3738. a 81 Cf. BraY i.44ab and xxxii.329ab, quoted below. a


Figure 5.2: The Descent of Scripture (astr vat ra) according to Svacchandatantra 8 s a a


Apara!iva/ Sad"!iva

dhvanir!pa" [#$stra"]

#!vara Laukik$di

vidye#varas of the vidy$tattva



rudras of m$y$tattva to niyatitattva rudras of prak%titattva down to the Hundred Rudras


The Goddess (Bhairav%, Aghore!%)

Skanda, rudras, the gods, n$gas, and yak&as



262 the Vimala that the Root Tantra emerges. Description of the Root Tantra as a text of twelve-thousand verses identies this as the BraY , one of the epithets of which is a The [Tantra of] Twelve-Thousand [Verses] (dv daas hasraka). a s a Mirroring and in some accounts interwoven with the cosmogenesis, scripture hence descends into the world from the apex of the universe through the levels of reality (tattva) into the materiality of language. In this conception, the scriptural wisdom rst takes on xed, linguistic form as the Vimala of one and a quarter hundred .. thousand verses, learnt by Srkantha, who transmits it to humankind in countless abbreviated redactions.82 This is not a scheme which automatically privileges the BraY , a problem apparently addressed by providing it the status of Root Tantra. a Chapter thirty-two, teaching initiation (dksa), contains a passage contrasting the . j naugha, the font of both scripture and the universe, with the Four-fold Bhairaa vatantra (catusp. ham bhairavam) which emerges from it. In contrast with the account . t . of chapter one, the j naugha itself is identied as the primordial Root Tantra, a rather than the catusp. ham bhairavam. In this account, no mention is made of the . t . Vimala, nor is the BraY specically identied with the catusp. ham bhairavam:83 a . t .
BraY i.3442. a Several verses preceding those quoted below, though relevant to the subject at hand, present yet unsolved textual problems and have unfortunately been omitted. BraY xxxii.323cd31: a
83 82

sad sivena devena ubhayam rg nuvartin 323 a a a a tat sarvam nirgatam devi j naugh sm n na samsayah | a a a . . . . suddham rge tath suddhe suddh suddhe tathaiva hi 324 a a a suddh ny eva hi prokt ni auddh ny api k ni ca | a a s a a suddh suddhau tu c ny ni ebhir bhedaih sthit ni tu 325 a a a a . prthakprth ni mantr ni ubhayam rgopasevin m | a a. a a . . sad sivena prokt ni bahubhir v pravistaraih 326 a a a . . ayam tu j nasandoham svarupenaiva samsthitam | a . . . na dvityam yata c sm t ten dvaitam param smrtam 327 . s a a a . . . suddh suddhavimiram sy t trividhasth py asamsayah | a s . a a . . yath vasthitakam proktam j naugham paramam sthitam 328 a . . a . . asm j j n d yato mantr s trailokyam sacar caram | a a a a a . aesani tu tantr ni bhedayitv sthit ni tu 329 s . a. a a ten dyam tu sam khy tam mulatantram na samsayah | a a a . . . . . asm d vinirgatam devi catusp. ham tu bhairavam 330 a . . t . . tato smin dksito yas tu sarvatantresu dksitah | . . . . suddh suddhavimiresu n tra k rya vic ranat 331 a s . a a a .
323c devena ] conj.; deve e Bya s 324b j naugh sm n ] em.; j nogh sm Bya a a a a a a 327b samsthitam ] em.; samsthit m Bya 327c yata c sm t ] em.; yato casm Bya 328a sy t ] em.; sy a. s a a a a a . . Bya 328b py asamsayah ] conj.; pi samsayah Bya 328c vasthitakam ] em.; vasthitak m a. . . . . . Bya 328d j naugham ] corr.; j nogham Bya 329a asm j j n d ] em.; asm j n Bya 330d a a a a a a a a . .


Figure 5.3: The Descent of Scripture (astr vat ra) according to Brahmay mala i and xxxviii s a a a






The Vimalatantra of 125,000 verses !r$ka%&ha

The Right Stream The Middle Stream The Left Stream The Goddess


The Root Tantra


The Bhairavatantras of Four P!"has Mudr"p$&ha Ma%'alap$&ha Mantrap$&ha Vidy"p$&ha

Ten scriptures of the #ivabheda

The V$matantras Eighteen scriptures of the rudrabheda

264 Through the Lord Sad siva, who follows both paths [viz., the pure and a impure], all this [scripture] emerged from this mass of scriptural wisdom, undoubtedly. [323cd24ab] For the pure way, the impure, and likewise the [mixed] pure-impure, were taught pure [tantras], some that are impure, and others that are pure-impure. Having these divisions, Sad siva taught a a variety of separate mantras with numerous elaborations for the followers of both [the pure and impure] ways. [324cd26] But this [scripture] is the mass of scriptural wisdom, existing as [its] very true nature. Since it is second to none, it is therefore known as supremely nondual.84 [32627] Pure, impure, and mixed comprise the threefold states, undoubtedly. The supreme mass of scriptural wisdom remains taught in accordance to the state.85 [328] Since the mantras [and] the triple universe, with its animate and inanimate [beings], came from this, [and] the tantras, without remainder, have divided from this, it is therefore called the Primordial Root Tantra ( dyam a . mulatantram), without doubt. From this emerged, O goddess, the Bhairava[tantra] with four divisions (catusp. ham bhairavam). [32930] Therefore . t . one initated into this is initiated into all the tantras, pure, impure, and mixed. About this, there is no need for deliberation. [331] Here too the boundary between the primordial scriptural wisdom and the text one actually reads is obfuscated: the proximate pronoun (ayam) describes the j naugha, a . as though it is the present text, while the catusp. ham bhairavami.e. the BraY a . t . is referred to as another scripture altogether. Through such rhetorical claims, the BraY continually inscribes upon itself the transcendental identity of the primordial a scriptural wisdom. Further evidence on this question emerges in chapter xxxviii, which, entitled A Denitive Judgment on the Streams of Revelation (srotanirnayapatala), is devoted . . to dening the Saiva canon. Its account articulates more clearly the nature of the Root Tantra in relation to the j naugha, the two of which are not in this case mua tually identied. From the j naugha emerge the tantras in three streams, which a
catus ] corr.; catuh Bya . .

Bya in 324b transmits j nogh sm , which I understand to mean asm j j naugh t; cf. 329a. The text a a a a a a should probably be corrected to j naugh sm n, with metri causa contraction of j naugh t. On elision a a a a a of the nal -t of the ablative in the BraY , see the annotation ad i.5 in the critical edition. a 84 There is here a play on the words dvaya (two, and dual) and dvitya (second). 85 The implication here is that scripture is redacted from its primordial source in accordance with the differences in the speakers and audiencei.e. prcchak srayabhedena, as stated in xxxviii.14a (quoted a . a below). Cf. Svacchandatantra 8.32b, referring to adh rabheda, difference in the locus [of revelation].

265 .. Sad siva teaches to Srkantha. These correspond to the pure, impure, and mixed a ways, and to the bhairavatantras of the Right stream, siddh ntatantras of the Middle a stream, and v matantras of the Left stream, respectively. At the root of the Bhairava a stream lies the mulatantra, the Bhairavatantra of Four Divisions (catusp. ham . . . tantram . t . . bhairavasamjitam, 19cd):86 . Proclamation of the tantras is done according to differences in [their] interlocutors and recipients. Through three streams [of revelation] for the pure, impure, and mixed, respectively, the tantras emerged from the mass of scriptural wisdom, [and] were spoken. [1415ab] The wisdom is not in all respects undivided; rather the tantras exist with differences in ritual for the pure, impure, and mixed, O fair woman. [15cd16ab] The supreme mass of scriptural wisdom emerges forth in three streams, situated on the left, right, and middle with manifold divisions. Seventy-million mantras emerge from the Right stream, with a profusion of manifold tantras based upon it, having differences in ritual. Even within this [stream], the recipients vary, being pure, impure, and mixed. [16cd18] Via the way of the Right stream arose the tantra called the Bhairava, possessing the four

BraY xxxviii 1426ab: a prcchak srayabhedena tantr nam krtanam krtam | a a. . . . . suddh suddhavimisrebhyas tribhih srotair yath kramam 14 a a . vinirgat ni tantr ni j naugh d bh sit ni tu | a a. a a a. a n bhinnam sarvath j nam kriy bhedaih sthit ni tu 15 a a a . a a . . suddh suddhavimiresu tantr ni varavarnini | a s . a. . j naughah paramo yas tu tribhih srotair vinirgatah 16 a . . . v madaksinamadhyastho n n bhedavyavasthitah | a a a . . . sapta kotyas tu mantr nam daksinay vinirgat h 17 a. . a. . . . a tam asritya kriy bhedair n n tantrapravistaraih | a a a . tatr pi asrayo bhinnah suddh suddhavimiritah 18 a a s . . daksinasrotam rgena kriy bhedavyavasthitam | a . a . . catusp. ham samutpannam tantram bhairavasamjitam 19 . t . . . . daksinasrotasambhutam sarvam asm d vinirgatam | a . . . .
14b tantr nam ] em.; tantr n m Bya a. . a a. 14c vimi rebhyas ] corr.; vimisrebhyah Bya s 14d . a tribhih ] corr.; trbhih By srotair ] em.; srotrai Bya 15b j naugh d ] corr.; j nogh d Bya 15 a a a a . . . sarvath ] em.; sarvvarth Bya 15d bhedaih ] conj.; bhede Bya 16c j naughah ] corr.; j noghah a a a a . . . Bya 16d tribhih ] corr.; trbhih Bya srotair ] em.; srotrair Bya 17b vyavasthitah ] em.; vyavasthit h a. . . . . . Bya 17c kotyas ] em.; koty s Bya 18a asritya ] corr.; asrtya Bya bhedair ] conj.; bhede 19a . . a srota ] em.; srotra Bya 19b vyavasthitam ] em.; vyavasthitah Bya 19c catus ] corr.; catuh . . . Bya 20a srota ] em.; srotra Bya 20b sarvam ] corr.; sarvvamm Bya .

In referring to the right stream of scriptural revelation, the BraY alternates between the masculine a daksina and feminine daksina. In 15d, Bya transmits kriy bhede, which appears to require emendation to a . . . . a. a. a the instrumental bhedaih. Cf. xxxviii.33ab, vinirgat ni tantr ni kriy bhedena caiva hi, where however the . instrumental is singular. Cf. also kriy bhedavibh gaah, BraY i.41b. In 15c, n bhinnam with its double a a s . a a . negative is somewhat suspect; one might consider na bhinnam. It is possible that the line intended . something along the lines of scriptural wisdom [or doctrine] is singular, but the tantras are divided according to differences in ritual. In 18a, tam probably refers to the daksina-stream, but one might also consider t n [mantr n]. a a . .

266 p. has [or Catusp. habhairavatantra], and arranged according to divisions t . t in ritual. Everything arisen from the Right stream emerged from this. [1920ab] The passage immediately following narrates the genesis of the Left and Middle streams. Although not specically designated as the mulatantra, the text BraY xxxviii dea scribes as catusp. ham bhairavam, identied elsewhere with the BraY itself, functions a . t . . as the Root Scripture, insofar as it is the source of the scriptures of the Bhairava Stream in its four divisionsthe mantra-, mandala-, mudr -, and vidy p. haswhich it a a t .. alone contains within itself.87 As such, it occupies a position in the canon above all but the transcendental j naugha or Vimala. In the conception of this chapter, after a .. Sad siva transmits the three scriptural streams to Srkantha, the latter transmits the a canon to the world in ten streams in the ten directions, with tantras by the millions (table 5.2).88 In this ten-stream model, the three primary streams remain largely the

BraY xxxviii 29cd33ab: a daksinena tu vaktrena daksinasrotasambhavam 29 . . . . . catusp. haprabhedena suddh suddhavibheditam | a . t prcchak srayabhedena bahudh samvyavasthitam 30 a a . . vidy srit ni y ni syur vidy p. ham var nane | a a a a t . a mantr srt ni y ni syur mantrap. ham tathaiva ca 31 a . a a t . mudr srt ni y ni syur mudr p. ham tu suvrate | a . a a a t . mandal p. hak ni syur mandalam p. ham ucyate 32 . . a t a .. . t vinirgat ni tantr ni kriy bhedena caiva hi | a a a
29d srotasambhavam ] em.; srotrasambhavah Bya 30a catus ] corr.; catu Bya 31a syur ] em.; syu . . . Bya 31c syur ] em.; syu Bya 31d ptham ] em.; pth Bya . . . a 32a syur ] em.; syu Bya 32b ptham ] em.; ptham Bya . . . . 32c syur ] em.; syu Bya 32d ptham ] em.; ptha Bya 33a tantr . . a ni ] corr.; tantr ni Bya a .

Through [Sad sivas] right face arises the daksinasrotas [of the bhairavatantras], divided a . . into the pure and impure, with division of Four Pthas, existing manifoldly because of . differences among the interlocutors and audiences. Those based upon vidy -mantras coma prise the Vidy ptha, O fair woman, and likewise, those based upon mantras comprise a . the Mantraptha. Those based upon mudr s comprise the Mudr ptha, O pious woman. . a a . The ones belonging to the Seat of Mandalas are called the Mandalaptha. And [thus] did . .. .. emerge the tantras, with differences in ritual (kriy bhedena). a In 32c, mandalap. hak ni appears to exhibit metrical lengthening, thus avoiding the fault of short sylla. . t a bles in both the second and third positions. 88 BraY xxxviii.8486: a adimo j nasandohas tribhih srotair vinirgatah | a . . sad sivena devena srkan. h ya prabh sitah 84 a .t a a. . sap daj nasandohah srkan. hena mah yae | a a a s . .t daasrotavibh gena bh sitah s dhakecchay 85 s a a. . a a

267 same. However, the siddh ntatantras become distributed among ve streams, with a the upward face of Sad siva revealing siva-division tantras, and four streams in the a intermediate directions revealing rudra-division tantras. The eastern face of Sad siva a is said to teach the Vedic revelation, while the downward stream is also of considerable interest, for here the BraY places the texts it considers least worthy of a veneration: treatises on alchemy, magical herbs, exorcism, and snakebite cure; last, and perhaps least, are the tantras of the Vaisnavas (see tables 4.24). .. Table 5.2: The Ten Streams of Revelation according to BraY xxxviii a s face of sada iva Southern Northern Upward Western Eastern Eastern Southern Western Northern n/a? presiding sakti Daksina . . V m a a Madhyam a V m and Madhyam a a a V m and Madhyam a a a direction [S] [N] [Up] [W] [E] E, SE S, SW W, NW N, [NE?] Down scriptural category bhairavatantras v matantras a sivabheda [rudrabheda?] vedas, etc. (ved dni) a rudrabheda rudrabheda rudrabheda rudrabheda rudrabheda

This three- and ten-stream model is unusual, and probably archaic; what actually becomes normative is a ve-stream model of Saiva revelation, in which the primary three streams are augmented by the lowly bhutatantras, in the west, and garudatantras .
koti koti pravist raih kalpakalpasahasrakaih | a . . . . pracodito mah devi tadvidair bahudh punah a a .


84a sandohas ] em.; sandehas 84b srotair ] em.; srotrair 84d prabh sitah ] em.; prabh sitam a. . a. . Bya 85b mah ya e ] corr.; mah yase Bya a s a 85a sandohah ] em.; sandehah Bya 85c . . srota ] em.; srotra Bya 85d bh sitah ] em.; bh sitam Bya a. . a. .

The primordial mass of Wisdom emerged forth in three streams. The Lord Sad siva a .. .. taught it to Srkantha. Srkantha, O woman of great renown, taught the mass of scriptural wisdom () having [one hundred] and a quarter [thousand verses] (?), through divisions of ten streams, according with the wishes of s dhakas. Those who learned these further a revealed [the wisdom] manifoldy, with millions and millions of elaborations, () and kalpatexts by the thousands (?). The interpretation of sap daj nasandohah (85a) is uncertain. I have taken this as an abbreviation of a a . sap dalaksaj nasandoha (j nasandoha meaning j naugha), i.e. the mass of scriptural wisdom measura a a . a ing 125,000 verses; however, it is conceivable that sap da here means having verse-quarters, i.e. actual a text in verse. Of linguistic interest in this passage is the thematized a-stem tatvida, for tadvid.

268 in the east; these have as their respective concerns exorcism and the magical cure of snakebite. And although the Root Scripture, the Bhairava Tantra of Four Divisions, has here articulated for it a compelling position within the canon, never in this chapter is a relationship of identity with the BraY stated directly. In fact, the BraY appears a a to have a humble position in this account of the canon, being mentioned only as the third of eight y malatantras in the Vidy ptha-division of the bhairavatantras. a a . The closing verses of the BraY propose what seems to be another variation on a the conception of its descent. In agreement with chapter one, the twelve-thousand verse BraY emerges from the Tantra of One and a Quarter Hundred-Thousand. Howa ever, the BraY is also said to be the ultimate essence of a tantra of one billion a verses.89 This might imply that the billion-verse text forms the source of the 125,000verse scripture, this in turn being the source of the BraY ; elsewhere, however, the a laksap d dhika tantra is said to emerge from the undifferentiated totality of scriptural . a a wisdom (niskalaj naugha). Both schemes of course agree in positing three fundaa . mental levels of originary scripture: gargantuan, 125,000 verses, and 12,000 verses. Invoking the authority of an intangible Ur-wisdom and claiming for itself the status of mulatantra, the BraY positions itself at the apex of actual and possible a texts within the parameters of existing models of the canon. Scriptural wisdom (j na) is a transcendental essence which descends in streams from a primordial a source, taking on concrete form according to the capacities of its various redactors and audiences. These linguistic manifestations of scripture are provisional, and the canon fundamentally an open one. Possibilities for new revealed texts are endless, each containing within itself the essence of what precedes. This model of revelation dictates that a new scripture claim for itself a privileged position on a hierarchical scale of texts, a process which involves, if necessary, reconguration and extension of models of the canon and their implicit hierarchies. As the Root Tantra, the BraY claims to be uniquely reective of the primordial wisa

See chapter 4, n. 27.

269 dom. Being the source and essence of the Vidy ptha, it places itself at the head of a . the four-fold (catusp. ha) canon of bhairavatantras, which in turn it places as highest . t of the three streams of Saiva revelation. And in broader sectarian terms, its additional taxonomy of ten scriptural streams positions the BraY above the orthodox a Veda, the Vaisnava Pacar tra, and so forth. Most distinctive about its articulation a .. of a place in the canon, however, is the layering of the BraY s identity. Drawing a upon the notion of scripture as an essence existing on a scale of texts, the BraY a continually blurs the boundaries between itself and higher levels of the scriptural wisdom. This begins with chapter ones revelation narrative: the text opens with the Goddess stating that, having learnt the mulatantra, she now desires to hear the Vimala from which it emerged. Bhairavas narration moreover is concerned primarily with the descent of the Vimala, and thisnot the twelve-thousand verse BraY is in fact a what he promises to teach her.90 Throughout the text, references to the BraY as a the Root Tantra, the Tantra of One and a Quarter Hundred-Thousand, and the Vimala, the boundaries of which are never completely clear, serve to articulate multiple levels of identity: it is but one of several y malatantras, but also the Root Scripture of the entire a Bhairava stream, having moreover as its highest existence the primordial source of scripture itself. BraY is by no means unique in utilizing the concept of mulatantra. The idea a of a vast scripture as the source of texts on a smaller, more human scale is attested in both Saiva and Buddhist tantric sources, and similar conceptions are found in Sanskrit texts of a variety of genres. Ideas of a divine-scale source text abound in the pur na corpus, while the Mah bh rata itself is sometimes said to exist among a. a a the gods in a version of millions of verses.91 Note also, for instance, in the medCf. BraY i.4244, especially. a On the notion of an Ur-pur na, both in the pur nas themselves and colonial-period Indology, see a. a. Ludo Rocher, The Pur nas, 4147. It seems probable to me that the pur nic notion of a mulasamhit has a. a. . a inuenced the notion mulatantras. As for the Mah bh rata, note for example two verses found in some a a manuscripts of Svarg rohanaparvan 5 (quoted on p. 29 in the apparatus of the critical edition), which a . speak of the Mah bh rata as having an Ur-text of six million verses, a three-million verse recension in a a the world of gods (devaloka), versions of one and a half and 1.4 million verses among the Ancestors (pitr) and the n gas and yaksas, respectively, as well as the 100,000-verse text known to men. a . .
91 90

270 ical literature, that the Surutasamhit claims origins from a text of 100,000 verses s . a composed by Brahm himself, which was divided and abridged for the good of a a short-lived and dull-witted humanity.92 In the Buddhist tradition, the exegetical tradition of the Cakra amvara cycle of yogintantras conceives of a Root Scripture of ones . hundred thousand verses as the source of the Laghuamvaratantra or Heruk bhidh na.93 s . a a Commentatorial literature of the Hevajra cycle also invokes a fabulously large Root Tantra,94 while the K lacakra tradition purports to have the lost Param dibuddha as its a a original scripture; the extant K lacakratantra is but an abridged tantra (the Laghua k lacakratantra), much as is the Laghucakraamvaratantra.95 a s . Among early Saiva scriptures, the Svacchandatantra professes descent from a version of a billion verses, which posed understandable obstacles for short-lived mor tals. Like the mulatantra the BraY describes, the original Svacchandatantra contained a within itself all four divisions (p. ha) of the Bhairava canon.96 The conception of Root t
Surutasamhit 1.1.6. This case was brought to my attention by Isaacson. s . a Cf. Bhavabhattas remarks on Laghuamvaratantra 1.1, commenting on its opening phrase, s . .. ath to rahasyam vaksye: bhagav n srcakrasamvar dhimukt n m arth ya laksaparim nan mulatantr t tad a a a a a a a. a . . . . .. [laghusamvaratantram] akrsya deayate . . . mulatantradean y anantaram mulatantram ev karkrtya rahasyam s s a a a . . . . vaksye iti sambandhah. Sarn th edition, p. 3. Similarly, Vajrap ni makes numerous references in a a. . . the Laghutantratk to the Root Tantra as Laksabhidh na, the One-Hundred Thousand Verse [Heruka]a . a . Abhidh na Tantra. a 94 Although conceptions of the Hevajra vistaratantra or mulatantra are diverse, perhaps most common is that of a Hevajratantra ve-hundred thousand verses in extent, from which the received text was drawn. Isaacson, lecture handout, The Problem of the Lost Root-Tantra of the Hevajratantra, University of Pennsylvania, February 2001. 95 On the subject of the Param dibuddha, the Root Scripture of the K lacakra cycle, see John Newman, a a The Outer Wheel of Time: Vajray na Buddhist Cosmology in the K lacakra Tantra, 118; and especially a a Francesco Sferra, Constructing the Wheel of Time. Strategies for Constructing a Tradition, 255273. Sferra also mentions the case of the Yoginsac ratantra and its references to a mulatantra; ibid., 268. a 96 Svacchandatantra 1.4cd7 (ksts edition):
93 92

yat tvay kathitam mahyam svacchandam paramevara 4 a s . . . satakotipravistrnam bhed nantyavisarpitam | . . a . catusp. ham mah tantram catus. ayaphalodayam 5 a . t . . .t na saknuvanti manuj alpavryapar kram h | a a a. alp yuso lpavitt s ca alpasattv s ca samkara 6 a . a a . tadartham samgraham tasya svalpaastr rthavistaram | s a . . . bhuktimuktiprad t ram kathayasva pras datah 7 aa . a . O supreme lord, the great Svacchandatantra which you had taught me, possessing [all] four divisions (catusp. ha) [of the bhairavatantras], bestowing the four fruits [of k ma, artha, a . t dharma, and moksa], ten-million verses in extent, and spread forth with an innitude of . divisions, humans are incapable [of comprehening], being of little wealth, little spirit, little vital energy and courage. For this reason, please teach an abridged version (samgraha) of . it with very little elaboration on the scriptures meaning, which bestows supernatural experience and liberation.

271 Tantras continued to have currency in Kaula sources as well; note for example that the Urmikaul rnava or Nratantra refers to itself as extracted from the Tantra of a Hundreda . Thousand (laksap doddhrta).97 Both the Manth nabhairavatantra and Sarvavratantra rea . a . portedly attribute their origins to massive mulatantras,98 as also do Vaisnava tantras .. such as the Jay khyasamhit and P dmasamhit .99 a a . a . a Not all Root Tantras were pious ctions, moreover, for the epithet is also applied to real texts deemed fundamental to particular traditions. As is the case with the BraY itself, being or claiming to be an authorative source-text justies the designaa tion mulatantra. The Trika Tantrasadbh va, for example, refers to the Siddhayogevara s mata as mulatantra, although in a version one billion verses in extent.100 This might be linked to the Siddhayogevarmatas own claim of being the source of the sixty-four s tantras, including the BraY .101 Reinforcing this status, the M linvijayottaratantra a a claims origin from a version of the Siddhayogevarmata ninety-million [verses] in s extent, via the intermediary stages of a M linvijayottaratantra version of twelvea
A Nepalese codex, ngmpp reel b28-18, offers as substantive variants ananta for anantya in 5b, alpacitt s in 6c, and ato rthasamgraham samyak for 7a. However, in the latter case the Bodleian Librarys a . . Nepalese codex of the Svacchandatantra agrees with the ksts edition, according to Trzsks transcription. Doctrine of Magic Female Spirits, 198. 97 Thus the colophon to Urmikaul rnava 1: iti nratantre srmadurmikaul rnnave mah sastre laksap doa . a .. a . a ddhrte paramarahasye srsrbhogahastakram y te srkaulagirip. havinirgate srmnap d vat rite sadsat dhike a a t a a a . . . a sate mamtroddh rarahasyaguruisyaparksavic ro n ma prathama patalah. nak 5-5207 (ngmpp reel b115/9). a s. . a a . . . 98 a Dyczkowski, Canon of the Saiv gama, 99, 11011. 99 Sferra, Constructing the Wheel of Time, 268. 100 Tantrasadbh va 1.1314ab: a siddhayogevartantre satakotipravistare | s . mulatantre mah sutre sutradvayavinirgatam 13 a tantraikam tu may j tam yony rnavasamudbhavam | a a . a . . Text as constituted by Dyczkowski; read however yonyarnava . . 101 Siddhayogevarmata 29.19: s evam dy s tu ye tantr s catuhsas. ivibhedit h | a a a a. . . .t . nirgat iha candaksi siddhayogevarmate 19 a s .. Note also that Siddhayogevarmata 31.5 seems to refer to the bhairavasrotas as possessing one hundreds thousand verses: srnusvaikaman bhadre yad vaksy mi sam satah | . . . a a . a . yena puryanti k ry ni kal bhir bhairavasya tu | a a. a srotasya tu samagrasya laksap dayutasya ca 5 . a Trzsk understands laksap dayutasya to refer to multiple lacs of verses, which is also possible. Doc. a trine of Magic Female Spirits, 182.

272 thousand verses, and the thirty-million verse Ur-M linvijayottaratantra.102 While it a does not apply the term mulatantra, the function is analogous. At least two later sources support the BraY s claim for the status of Root Tantra. a a In the Bhairavamangal , the Goddess states, you have indicated the eightfold smarana. mantra [taught] in the mulatantra.103 Conrming that mulatantra refers to the BraY , a several verses later the text refers to the eight-fold smarana-mantra originating in . the Picutantra, i.e. BraY .104 The Jayadrathay mala, which conceives of multiple Root a a Scriptures, lists ve y malatantras beginning with the BraY among the mulatantras a a of the Vidy ptha. This account of the Vidy ptha and y malas gives pride of place a . a . a to the BraY , and indeed many lists of y malatantras place the BraY at their head. a a a In support of its claim to be a Root Scripture, there are moreover indications that the BraY served as the primary authorizing scripture for a body of practice and a exegesis. While the record is more fragmentary than in the case of the Trika, several a surviving sources, such as the Pingal mata (a pratis. h tantra, i.e. concerned with rites .t a of empowerment for images and so forth) and the Matas ra claim to belong to the a tradition of the BraY . a

M linvijayottaratantra 1.812: a svasth nastham um dev pranipatyedam abravt | a a . siddhayogevartantram navakotipravistaram 8 s . . yat tvay kathitam purvam bhedatrayavisarpitam | a . . m linvijaye tantre kotitritayalaksite 9 a . . yogam rgas tvay proktah suvistrno mahevara | a a . s . bhuyas tasyopasamh rah prokto dv daabhis tath 10 a . a s a . sahasraih so pi vistrno grhyate n lpabuddhibhih | . . a . . atas tam upasamhrtya sam s d alpadhhitam 11 aa . . sarvasiddhikaram bruhi pras d t paramevara | a a s . The goddess Um prostrated to [Siva] in his own abode, and said this: the Siddhayoa gevarmata which you had previously taught me extended ninety-million verses, spread s forth with the three divisions [of the sakti?]. [And] the path of yoga which you had taught in the M linvijayottaratantra possessing thirty-million verses was extremely lengthy, O a Mahe vara. And furthermore, the abridgment of this you had taught with twelves thousand verses was also lengthy, not understood by those of little intellect. Hence, for the good of those of little intelligence, please further abridge this and teach [a tantra] which produces all the siddhis, O supreme lord.

103 104

a a. Bhairavamangal 230ab: as. adh smaranam deva mulatantre tu sucit m. .t a . . a Bhairavamangal 235ab: smaranam as. adh j tv picutantr samudbhavam. a . . .t a a a



s Dvada asahasraka: the Tantra of Twelve-Thousand Verses

A number of Indic texts are known by their verse count. Famous examples include the Sattasa compiled by H la, and the Durg saptaat or Devm h tmya, both ostensi a a s a a bly compositions of seven-hundred verses. There exists also the sataka or century genre, comprising texts of approximately one-hundred verses such as the Amaru sataka, and the Candsataka of B na. A number of Saiva scriptures too make promi a. . . nent reference to their length in colophons or the text, sometimes even in their titles; note for example the S rdhatriatik lottara (The Three-hundred and Fifty-Verse a s a Addendum-tantra of the K latantra). The BraY for its part calls itself Dv daaa a a s s hasraka, the [Tantra of] Twelve Thousand [Verses]. Twelve-thousand is moreover a not an exaggeration, as the text in fact consists of roughly 12,800 verses.105 According to the revelation narrative (astr vatarana) of chapter one, the Tantra s a . of Twelve Thousand is but one of many redactions of various lengths from the Vimala or Tantra of One and a Quarter Hundred-thousand. In this narrative, the numerous tantras mentioned are almost all referred to by length alone. The Dv daas hasraka dea s a scends at the advent of the Kaliyuga, revealed to the residents of the Isle of Maidens. Taught the Tantra of One and a Quarter Hundred-thousand by the supreme Goddess herself, a guru named Svacchandabhairava contracts this to the size of twelve-thousand verses, and then teaches the scripture to Visnubhairava in the village of Kal pa. Visa .. . nubhairava then reveals the text to the residents of the Isle of Maidens (kum rdvpa), a . among whom the scripture attains paramount popularity. In the last portion of the Kaliyuga, however, the scripture and lineage disappear altogether, snatched by the yogins, not to reappear until the commencement of the next Kaliyuga.106 The epithet Dv daas hasraka hence places the BraY within a canon of myriad aca s a a tual, possible, and mythical texts, structurally embedded in the cosmos and its cycles
105 Approximated on the basis of an average of eighteen verses on each of seven-hundred and thirteen folio sides. There is a margin of error of perhaps three-hundred verses, for it is possible that I have overlooked errors in the folio numbering. The typical range of verses per folio is 17.518.5. 106 BraY i.102cd105. a

274 of time. In the course of cosmic cycles, the lineages of gurus abbreviate the scriptural wisdom and redact it into countless scriptures of various lengths for the good of a variously-abled humankind. Its ultimate condensation consists of the vidy -mantra a alone.107 At the other extreme lies the Vimala of 125,000 verses, from which directly emerges the mulatantra of twelve-thousand versesthe BraY . In these idealized a terms, the BraY is thus a redaction of merely medium length. Its size, however, a makes it almost uniquely voluminous among surviving tantras, surpassed only, I believe, by the Jayadrathay mala.108 Other large and early surviving scriptures include a the Tantrasadbh va, and there may have existed several more extremely voluminous a tantras. BraY s revelation narrative speaks of a tantra in twenty-four thousand verses, a and the implication appears to be that anything larger than that is beyond the capacity of mere humans.109 Within the wider Saiva tradition, the epithet dv daas hasra does not appear to a s a have been associated exclusively with the BraY . I am not in fact aware of extera nal references to the BraY by this epithet, and Abhinavagupta refers in Tantr loka a a 15 to another scripture, the Anandevaratantra, which his commentator describes s as dv daas hasra.110 As mentioned earlier, the M linvijayottaratantra as we have it a s a a claims to be an abridgement of a version of twelve-thousand verses, although it .t seems doubtful that such a text existed. The Srkan. hyasamhit also refers to the . a Saukrasamhit , a lost v matantra, as containing twelve-thousand verses.111 Among a . a Tantric Buddhist sources, the Vimalaprabh commentary on the K lacakratantra is a a dv daas hasra, as is its purported Root Tantra, the Param dibuddha.112 a s a a
BraY i.67. a Sanderson reports that the Jayadrathay mala consists of twenty-four thousand verses. Saivism and a the Tantric Traditions, 674. According to Dyczkowski, the Manth nabhairavatantra calls itself a tantra of a a twenty-four thousand verses as well, its actual verse-count being closer to 17,000. Canon of the Saiv gama, 97. 109 i.49ab. 110 Jayaratha, commenting on Tantr loka 15.281. As mentioned previously (chapter 4, section 3, n. 89), a the Tantrasadbh va appears to mention a dv daas hasra scripture, although the grounds for linking this a a s a to the BraY appear weak. a 111 .t The Srkan. hyasamhit describes the Saukrasamhit as dv daas hasr, containing twelve-thousand. a s a . a . a Verse 258 in the transcription of Hanneder, Abhinavaguptas Philosophy of Revelation. 112 The Laghutantratk in fact quotes from the Param dibuddha, referring to it as dv daas hasrika (chapa a s a . a
108 107



Vimala and the Ucchusmatantra .

Although there are but a handful of references to the BraY as the Vimalatantra and a the Ucchusmatantra, these may have disproportionate signicance. What sets apart . these particular epithets is that they might possibly refer to independent scriptures upon whose tradition and authority the BraY draws, and with which the text also a occasionally identies itself. These text titles hence have a special status in the selfpresentation of the Brahmay mala. Moreover, the BraY intrinsically connects the a a Vimalatantra with the deity Ucchusmabhairava, suggesting a relationship between . the Vimala and Ucchusmatantra. . As discussed previously, the BraY s revelation narrative provides Vimala (Una blemished) as the name for the j naugha in its manifest form as a text of 125,000 a verses, the source of all scripture. In several instances the BraY refers to itself by the a title Vimala, thus claiming as its highest identity the originary scriptural wisdom.113 While the notion of 125,000 verses is surely mythic, a Saiva scripture called the Vimala might nonetheless have existed. No old tantra by this title appears to survive;114 how ever, the Vimala has a place in some accounts of the fundamental Saiva canon, and may hence have been an early Siddh ntatantra. Among the extant early Saiddh ntika a a sources attesting the ten-plus-eighteen model of the canonthat of ten scriptures of the siva-division (ivabheda) and eighteen of the rudra-division (rudrabheda)several, s although not perhaps the earliest, include the Vimala among the rudrabheda tantras.115 The colophon to BraY xxxv provides Ucchusmatantra as yet another title for the a .
ter xiv, p. 124). 113 Viz., divy divy ms tath hy ete tantre smim vimal hvaye (iv.8cd.); saktayas tu tav khy t [s] tantre smin a a. a a a aa vimal hvaye (lxxxviii.149cd); and et s tu tithayah prokt[ h] tantre vimalasamjake (xxiv.103cd). a a a. . . 114 There does survive some material attributed to a Vimal gama in composite South Indian a manuscripts, but these Saiddh ntika ritual tracts give no suggestion of antiquity. Cf, e.g., Institut a Franais de Pondichry manuscript T. 71, which contains a tract on temple mantras (pr s damantra) aa drawn from the Vimal gama. (This is a transcription of Government Oriental Manuscripts Library, a Madras, ms no. r14398.) There also exists a Saiva paddhati (ritual manual) called the Vimal vat, but this a is certainly unrelated. It is preserved in a Nepalese palm-leaf ms (nak 1-1536, ngmpp reel b28/7), and apparently two paper manuscripts as well (nak 3-586, ngmpp reel a187/1; and nak 1-131, ngmpp reel a186/10). 115 See Goodall, Bhat. a R makan. has Commentary, appendix iii. .t a .t

276 text,116 while xlv cites a text called the Ucchusmottarabhairavatantra as the source of . the extended version of a ritual it teaches.117 We do have some knowledge concerning a text entitled Ucchusmatantra or Ucchusmabhairavatantra, although nothing . . by the name appears to have survived. Abhinavagupta cites a scripture bearing this title in the Tantras ra and Par trimsik vivarana, while Ksemar ja quotes from it in his a a . a a . . commentaries on the Sivasutra, Netratantra, and Svacchandatantra. One of the verses Ksemar ja adduces from the Ucchusmatantra is also quoted by Jayadratha, without a . . attribution, in the Tantr lokaviveka. From most of these short quoted passages it is difa cult to ascertain the character of the text. Ksemar ja, for example, quotes a passage a . which describes the distinct mantra-functions of om, namah, svaha, phat, vausat, . . . . . . and hum.118 More illuminating, in the Tantras ra, Abhinavagupta appears to refer a to the Ucchusmatantra as representative of the radical far left of Saiva scripture, . paired opposite the Saiddh ntika P ramevara.119 This is corroborated to an extent by a a s a passage Ksemar ja quotes several times in support for the nondualist position of a . there being no such thing as impurity.120 Offering further conrmation of the texts
116 117

ity ucchusmatantre picumate n dsac rapatalah sattrmsatimah. a. a . . . . . .. . BraY xlv.12425: a purvam eva may khy t m ucchusmottarabhairave | a aa . mah manth navidh nam vistarena yaasvini 124 a a a . s . samksiptam tantrasadbh vam sarahasyasamuccayam | a . . . . dadhn d ghrtam ivodhrtya s r t s rataram subham 125 a aa a . . .
124d ya asvini ] corr.; yasasvini Bya s 125c dadhn d ghrtam ] conj.; dadhn ghrtav Bya a a . .

I had earlier taught the Great Rite of Churning at length in the Ucchusmottarabhairava. tantra, O woman of renown. The essence of [that] tantra, along with its collection of secret teachings, has been extracted like ghee from yoghurt and abbreviatedthe auspicious essence of the essence. Underlying what Bya transmits as dadhn I conjecture to be an ablative, dadhn d, formed from the a a thematized stem dadhna, for dadhi (yoghurt). 118 Commenting ad Netratantra 19.8. This passage is quoted by Peter Bisschop and Arlo Grifths, The Practice Involving the Ucchusmas (Atharvavedapari ista 36) (forthcoming), 34. s .. . 119 Tantras ra 4.45: tasm t vaidik t prabhrti p ramevarasiddh ntatantrakulocchusm diastrokto pi yo niyamo a a a s a . a . a s vidhih v nisedho v so tra y vad akimcitkara eva iti siddham | (Therefore, starting with Vedic revelation, a . a a . . any regulation, whether a rule or prohibition, stated even in the scripturesthe P ramevara Siddh ntaa s a tantra, the Ucchusmatantra of the Kula, in this case inoperative. So it is established). . 120 Quoted in the Svacchandoddyota ad Svacchandatantra 8.249 and 11.927, and in the Sivasutravimarin, s ksts edition, p. 8: y van na vedak ete t vad vedy h katham priye | a a a a. . vedakam vedyam ekam tu tattvam n sty aucis tatah s . . . . a As long as they are not knowers, how can there be objects of knowledge, my dear? But the knower and object of knowledge are one; therefore impurity is not something real

277 orientation, the Buddhist author Advayavajra quotes a verse from the Ucchusmatantra . which speaks of the bliss of the union of Siva and Sakti as the supreme nonduality (param dvaya).121 It would however be a matter of considerable surprise if a bhairaa vatantra antedating the BraY exhibited developed nondualism of the gnostic variety, a to the extent of being singled out for this by Abhinavagupta. This raises the possibility that the Ucchusmatantra referred to in the BraY is distinct from the text of this a . name cited by later authors. As for the title, the term ucchusma occurs in the Taittiryasamhit ,122 and is pre . . a sumably related to susma or susman, words connected with heat and light. One of . . the pariis. a texts of the Atharvaveda bears the name Ucchusmakalpa, on which Peter s .t . Bisschop and Arlo Grifths have just contributed a study.123 This text opens with an invocation of the ucchusmas, here apparently a class of minor male spirits.124 The . Tantrasadbh va, moreover, describes a class of female spirits called ucchusmik s.125 a a . Buddhist tantric sources attest a deity called Ucchusmajambhala, a form of the yaksa. .
(tattva). Cf. Bisschop and Grifths, Practice Involving the Ucchusmas, 4, to which my interpretation is in. debted. See also the entry ucchusma in T ntrik bhidh nakoa, vol. I, 22526. a a a s . 121 Sekanirdea, verse 10: s sivaaktisam yog t satsukham param dvayam | s a a a . na sivo n pi sakti ca ratn ntargatasamsthitam a s a . Advayavajrasamgraha, part iv, p. 50. Remarking upon the verse, the (anonymous) editors identify the . a Ucchusmatantra with the BraY , on the authority of Goudriaan and Gupta, Hindu Tantric and S kta a . Literature, 42. I am grateful to Isaacson for providing this reference. 122 Taittiryasamhit; see Bisschop and Grifths, Practice Involving the Ucchusmas, 2. . a . 123 Bisschop and Grifths, The Practice Involving the Ucchusmas (Atharvavedapari ista 36) (forths .. . coming). 124 om nama ucchusmebhyah, Pariis. a 36. s .t . . . 125 The Tantrasadbh va uses both the terms ucchusm and ucchusmik for this variety of female spirit. a a . a . In a passage Ksemar ja quotes in commenting on Netratantra 19.55, the ucchusmik is described thus: a a . . a r trau bhutv vivastr y mutrayitv pradaksinam | a a a a . . krtv tu pr sayed raktam muktake tu karsayet a a s . . . ucchusmik tu s jey s dhakair vran yik | a a a a a a . A woman who at night becomes naked, urinates, then circumambulates, and would consume bloodand then with hair unbound, would subdueshe s dhakas should know a to be an ucchusmik , a heroine. a . This corresponds to 16.18788ab in Dyzckowskis collation/draft edition. Aside from several probable corruptions, the Nepalese mss collated attest two substantive variants: pr sayate for pr sayed, and a a vravatsal for vran yik . The syntax is ambiguous, for either pr sayed or karsayet lacks an object. Cf. the a a a a . interpretation of Bisschop and Grifths, Practice Involving Ucchusmas, 5. .

278 lord Jambhala,126 while in Saivism, a rudra by the name Ucchusma is known from . early Saiddh ntika sources such as the Kiranatantra.127 Ucchusma the rudra appears a . . to have developed into a bhairava and acquired his own cult, for the Ucchusmatantra . or Ucchusmabhairavatantra must have expounded a pantheon headed by Ucchusma. . bhairava. Similarly, the BraY s Kap lsa appears also to have begun his career as a a a rudra, heading the list of one-hundred rudras in the Niv saguhya before achieving his s a place among the eight bhairavas in the mandala of Svacchanda. .. The BraY associates Ucchusma with the groups of goddesses forming the maa . ndala of Kap lsabhairava, which suggests the possibility that the BraY s Kap lsa a a a .. has inherited his pantheon.128 The association seems particularly strong between Ucchusma and the Mothers, who are referred to as the Mothers arising from the . Ucchusma[tantra/bhairava?], or simply, the Ucchusma Mothers.129 A female coun. . terpart to this deity, Ucchusm or Mahocchusm , has an important position within . a . a the mandala of Kap lsa and Aghore var, as the lastbut perhaps highest130 of a s .. the Four Duts. Ucchusmabhairava himself makes an appearance only in the man. .
Bisschop and Grifths, Practice Involving the Ucchusmas, 56. . In the cosmology of the Matangap ramevara (12.34d), Ucchusma is one of the rudras presiding over a s . k latattva, while in the Kiranatantra, Ucchusma is listed among the rudras at the tattvas k la and niyati a a . . (8.129a). 128 Note in particular the following passages from BraY iv: a
127 126

rakt kar l candakhy m mahocchusm tathaiva ca 252 a a .. a. . a ucchusmatantre n m ni guhyak n m na samsayah | a a a a. . . . And, rakt y s tu kar l sy t kar ly y s tu dantur | a a aa a a a a a bhmavaktr tu candaksy ucchusm y mah bal 262 a a a .. . a . a a guhyak nucar hy et h kinkaryocchusmasambhav h | a a a. a. .
kar l ] em.; kar lam Bya candaksy ] conj.; candaks Bya aa a . . . . a . . . sambhav ] em.; kinkary cchusmasambhav Bya a a a . et h ] corr.; et Bya a. a kinkaryocchusma.

See also the text and annotation to the translation of BraY i.89, and BraY lix.77: a a ucchusmasambhavam tantram devya cocchusmasambhav h | s a. . . . . a tvayoktam tu mah deva sutr dibhi mahevara 77 a s .
devya ] em.; devy Bya s a

O Mah deva, with sutras, you have taught the tantra arising from Ucchusma and the a . goddesses arising from Ucchusma, O Mahe vara. s . BraY i.133b and lxxxviii.244d. a The high status of Mahoccusm is suggested by her position in cosmology: in descriptions of the . a pervasion (vy pti) of the tattvas by the mantra-deities, Mahocchusm is placed at the level of Sad siva, a a . a above the other goddesses (table 4.7ce).
130 129

279 dala of Picubhairava, in a conguration of male counterparts to the eight goddesses . of the mandala of Kap lsa.131 On several occasions, ucchusma is however used as a a .. . synonym of the supreme Bhairava.132 The BraY s appeal to the authority of the Vimala establishes its link with the a earliest canonical scriptures of the Mantram rga. Whether or not its contents bore a a relationship is another matter. A distinct possibility exists that the Vimala, assuming the text in fact existed beyond lists of the canon,133 had been lost by the time of the composition of the BraY , continuing nonetheless to command considerable prestige. a Its actual contents largely forgotten, it remained nonetheless an empty signier of canonical authority on which to inscribe the pedigree of the BraY .134 In contrast, a the Ucchusmabhairavatantra is more likely to have been a text contemporaneous to . the BraY , perhaps lying in its immediate cultic background. Its mandala appears a .. to have consisted of primarily female deities, headed by Ucchusma and perhaps his . consort Mahocchusm . Furthermore, the BraY posits a close connection between the a . a Vimala and Ucchusmabhairava, describing the Vimala as the tantra originating from . Ucchusma.135 There could conceivably be substance to this link, given Ucchusmas . . early history as a rudra and the Vimalas presence in some lists of the rudra-division scriptures. According to the Mrgendr gama, one of the sources listing the Vimala in its a . account of the canon,136 Siva only indirectly teaches the rudrabheda-scriptures; these
BraY lxviii, especially verse 21. a BraY lxxxvii.96cd: nir c ra[h] sa ev tra ucchusmam parikrtitam (That very [state of the supreme a a a . a . . Bhairava known as] beyond conduct (nir c ra) is here called ucchusma). a a . 133 It might also be possible that the notion of a mythic Vimala contributed to the inclusion of a text by this name in lists of the canon. 134 This perhaps bears comparison with the phenomenon of replacing forgotten but authorative scriptures, passing off the new as the ancient so as to authorize contemporary practice. Such is best attested in the case of the South Indian Saivasiddh nta, in which case texts were duly provided to ll a in gaps between extant scripture and authorative accounts of the canon. See Goodall, Bhat. a R makan.t a . t a . has Commentary, xxxvixlvii. As mentioned in the introduction, the BraY itself appears to have been updated in later times in both east India and Nepal, besides South India. 135 BraY i.4b. The opening of chapter xxviii, moreover, announces it will teach the Vimala-division a (vimalam vibh gam) brought forth by the lotus [mouth?] of Ucchusma, revealed by Ucchusma: a . . . .
132 131

atah param pravaksy mi vibh gam vimalam dhruvam | a . . . . a . ucchusmakamalodgrnam ucchusmen vat ritam 1 . . a a . .

Mrgendr gama 1.46. a .

280 are brought forth by rudras who have been possessed/entered into by Sivanot [brought forth] from their own intellects.137 It might be possible that the historical Vimala had as its speaker the rudra Ucchusma, in which case its link with the . Ucchusmatantra would be based upon the shared gure of Ucchusma, as a rudra and . . bhairava, respectively. What can be asserted with more condence is that through its appropriation of the gure of Ucchusma, the BraY projects a pedigree rooted a . in earlier layers of the tradition: both the early canon of Siddh ntatantras, through a a Vimalatantra purportedly spoken by the rudra Ucchusma, and the bhairavatantras, . through the cult and scripture of Ucchusmabhairava. . Conspicuous by their absence are the Vimala and Ucchusmatantra in the extended . account of the Saiva canon in BraY xxxviii, despite their importance in the revelation a narrative (astr vat rana). Understandably so given its status as the mass/totality of s a a . scriptural wisdom (j naugha), the Vimala nds no place in the list of twenty-eight a Saiddh ntika scriptures; but neither does Ucchusma gure among the eight bhairavaa . tantras.138 The possibility seems substantial that the BraY has in the form of the Vimalaa tantra and/or Ucchusmatantra woven historical texts into its genealogy and identity, . although our knowledge of these sources remains vague. That this scenario is plausible nds support in the authorizing strategies of later scriptures, which offer insight into the processes of legitimization the BraY itself might have employed in a relation to older literature. Adding new layers to the palimpsest of revelation, subsequent literature claiming to belong to the tradition of the BraY continues the praca
Mrgendr gama 3.43ab: raudr rudraih siv vis. air udgrna na svabuddhitah. a a . . . a .t . The eight Bhairavas, who correspond to the eight bhairavatantras of the Vidy ptha, are Svacchanda, a . a Krodha, Unmatta, Ugra, Kap l, Jhank ra, Sekhara, and Vijaya. BraY xxxviii.33cd35ab: a a
138 137

svacchandabhairavam devi krodhabhairavam eva ca 33 . unmattabhairavam devi tath caivograbhairavah | a . . a kap lbhairavam caiva tath jhank rabhairavah 34 a a . . sekhara ca tath caiva vijayabhairavam eva ca a .t Perhaps closest to this list of eight bhairavas is that of the Srkan. hya, quoted by Jayaratha commenting on Tantr loka 1.18: Svacchanda, Canda, Krodha, Unmatta, Asit nga, Mahocchusma, and Kap l. See a a a .. . also Svacchandatantra 2.117cd19. Note in the passage quoted above free alternation of the nominative and accusative cases, the sense being nominative.

281 tice of self-identication with authorizing sources. Thus the nal colophon of the Matas ra, Essence of the [Picu]mata:139 Thus ends the thirty-third chapter within a the Matas ra in the Vidy p. h , in the Tantra of Twelve-Thousand, within the Great Scripa a t a ture numbering 125,000 verses.140 Here the ambiguity of the locatives is meaningful: the passage may in part be read as a hierarchy of texts within texts, from the originary wisdom of 125,000 verses down to the Essence of the [Picu]mata, or it may be read as a series of identications: the Matas ra which is the BraY which is the Tantra a a of One and a Quarter Hundred-Thousand. And so continues the perpetual descent of the scriptural wisdom into the multiplicity of texts.

139 While the text refers to itself simply as the Matas ra, Sanderson suggests that this means a Picumatas ra. Personal communication, May, 2004. Given the texts close relationship with and appeal a to the authority of the BraY , this seems quite plausible. a 140 iti laksap d dhike mah samhit y m dv daas hasre vidy p. he matas re trayovimsatimah patalah. nak a . a a. a s a a t a . a a . . . . 3-379, f. 161v. Transcription courtesy of Somadeva Vasudeva.

Chapter 6

A signicant concern of the present thesis has been to position the BraY relative to a rst-millennium textual and other sources concerned with yogins. With this aim in mind, chapters two and three have surveyed early evidence for the cult of yogins, attempting to establish a chronology of sources and important developments within which to situate the BraY . It has been shown that, although a Saiva cult of yogins a must have developed prior to the eighth century, it is in this period that yogins begin to come into prominence, a trend which continues with the emergence of a public temple cult in the tenth century. I have argued that the BraY in all likelihood a belongs to an early stratum of evidence: being mentioned in the old Skandapur na, a. predating the Buddhist Laghuamvara and (most probably) the Saiva Tantrasadbh va, s . a making no obvious reference to Kaula literature, showing no awareness of traditions of sixty-four yogins, having an archaic model of the Saiva canon, and being a well established authority by the mid-tenth centuryat the very latestthe BraY shows a notable signs of archaism. It has not yet, however, been possible to date the text with any precision. Likely as the BraY is to be one of the oldest surviving bhairavatantrasand pera haps among the older extant works of tantric Saiva literaturethe text nonetheless presupposes a wide variety of Saiva cults and tantric scriptures. With roots in the cults of Rudra/Bhairava, Mother goddesses, the Sisters of Tumburu, dakins, and . 282

283 bhutas, the BraY and other Vidy ptha sources synthesize a range of esoteric tradia a . tions and sacred gures, our knowledge of which has signicant gaps. Furthermore, while the BraY represents, in ritual terms, a radical and highly esoteric tradition, a its roots in aspects of earlier Saivism are notable, as suggested by comparison with the Niv satattvasamhit in particulara text the BraY appears in one case to draw s a a . a from. In the gures of Svacchandabhairava, Kap lsa the rudra, Ucchusma, and the a . Sisters of Tumburu, traces in the BraY of past cults point toward it being the product a of complex historical layering. This picture becomes all the more complex when the BraY s own textual development is queried; it has been suggested that the text as we a have it possesses two or more strata, and contains chapters potentially drawn from or otherwise connected with independent Saiva works. Layering is a theme in the BraY s reexive vision as well, for in articulating a model of scriptural revelation, the a text positions itself at multiple levels: it is but one of several y malatantras, yet also a the Root Tantra of the entire bhairava-stream, itself a contracted form of the Tantra of One and a Quarter Hundred-thousand Versesscripture in its primordial linguistic form. A key development in the BraY and Vidy ptha sources lies in the gure of the a a . yogin: a malleable goddess typology which comes to encompass cult deities, ev ery manner of female spirit and demi-goddess, and even female tantric adepts, all of whom become linked in a hierarchical matrix of clans (kula) emanating from the Supreme Sakti herself.1 While it has been shown that Saiva conceptions of yogins have roots in Mother goddesses and gures such as the vidy dhar, aspects of the fora mation of this category of sacred gure in Saiva tantric literature remain murky. This undoubtedly has much to do with chronic losses in tantric Saiva literature; indeed, the BraY makes reference to numerous other Vidy ptha sources, few of which are a a . extant. On the other hand, it appears possible to map the emergence of the yogin
BraY lv.319 describes a hierarchy of nine principal goddess clans, correlated with nine tattvas: a m trs, yogins, duts, rudradakins, damars, davs, siv s, bhagins, devs, and their genetrix (prasuti), the a. . a . . supreme Sakti, which exists at the level of the sivatattva. Human practitioners enter into the hierarchy of clans through initiatory kinship with the Mothers, goddesses who exist at the level of the prakrtitattva. .

284 or dakin in Buddhist tantric literature, on which subject chapter three has attempted . to contribute. The relationship between these two textual corpusesand more importantly, the religious traditions which produced themremains a signicant area for future research. Concerning the case of the BraY , evidence has been adduced in a support of Alexis Sandersons contention that it is a source for material redacted into a Buddhist yogintantrathe Laghucakraamvaratantra. s . Despite the efforts documented in the present thesis, study of the BraY remains a at an early stage. The endeavor to critically edit the texta prerequisite to deep understandingremains a project of signicant proportions. It is expected that data gleaned from the close study entailed by further critical editing will lead to revision of some claims I have made, while opening up new avenues of investigation as well.

Part II The Brahmayamala: A Critical Edition and Translation of Chapters i, ii, lv, lxxiii, and xcix



a The late T. Ganapati S str, the distinguished rst editor of, among a great many . Sanskrit texts, the Majurmulakalpa, offered in his preface the following apology for s reproducing the text precisely as found in the only manuscript available to him: This is a holy work of the Buddhists and deserves to be placed along with the Vedas. As the non-observance of the rules of Vy kara[n]a in regard to a . gender, number and case, found throughout this work is becoming of its sacred character, and as no second manuscript has been obtained, the text in this edition is adopted exactly as it is found in the original manuscript.1 a It is not clear which of the two points weighed more heavily in S strs decision: the texts holiness, or the paucity of manuscript evidence. In either event, the choice spared him considerable trouble, for cases such as the BraY or Majurmulakalpa, in a s which the language is non-classical and manuscript evidence very limited, present special problems. The approach of the present edition might be faulted as representing the opposite extreme, for the critically constituted text of the BraY introduces numerous emena dations and conjectures. But to avoid the risk of emendation would, in this case, be to embrace unintelligibility, for the readings of the oldest codexfrom which, I contend, the other extant mss are descendedare with vexing frequency incomprehensible. Furthermore, in the absence of corroborating evidence, I have erred in favor of normalizing potential linguistic irregularities. Considerations of sense, coherence, and context have served as the principal bases for emendation, with attention to the

a Ganapati S str, preface to Aryamajurmulakalpa, vol. 1, 2. s .


287 paleographic plausibility of scribal error as well as intra- and intertextual parallels (as adduced in the annotation to the translation). The present edition is a work in progress; although it seems unlikely that valuable new manuscript evidence will surface, I presume that further reection, the identication of additional parallels, and the insights of other readers will make possible considerable improvement in the constitution and interpretation of the text, prior to its eventual publication.

the manuscripts
nak 3-370 (reported in the critical edition as A, and as Bya in part i). ngmpp microlm reel no. a42/6. Palm-leaf, 358 folios. Dated Sunday ( dityadina), the a 8th of M gha (waxing fortnight), Nepal samvat 172i.e. Sunday, 12 January, a . 1052 c.e. Copied by Jay karajva, a resident of the Pa upatin tha temple area a s a of Kathmandu.2 This codex was cataloged more than a century ago by Harapras d S str, who notes: the MS. is marked with letter numerals on the left a a and with Newri gures on the right. They agree up to the 129th leaf, and from the 130th the Newri gure make a mistake of 10, and the mistakes on the right side continued to be added to and subtracted from till last leaf (358) becomes 364 in the Newri side.3 The text-nal and scribal colophons are as follows: iti bhairavasrotasi mah tantre vidy p. he brahmay male nav ksaravidh ne picumate a a t a a . a dv daas hasrake ekottaraatimah patalah sam ptah a s a s a . . . . samvat a cu 2 m ghaukl sa s a.

t a. aa a s s a s a . amy m adityadine r j dhir japaramevararbaladevar jye | srpaupativ stavya | srjay karajvena brahmay malam n ma sastram likhitam. a a . a . . nak 5-1929 (reported as B). ngmpp microlm reel no. a165/14. Paper, Nepalese
Cf. Luciano Petechs discussion of this colophon in Mediaeval History of Nepal (circa 7501482), 2nd ed., 44. 3 S str, A Catalogue of Palm-leaf and Selected Paper Manuscripts Belonging to the Durbar Library Nepal, a a vol. ii, reprinted in Reinhold Grnendahl, A Concordance of H. P. S stris Catalogue of the Durbar Library and the Microlms of the Nepal-German Manuscript Preservation Project, 6061.

288 N gar script; undated and unsigned. Complete in 353 folios. The rst portion a of this ms appears to have been copied from the same highly corrupt exemplar as C (see below). However, the latter sections of the text appear copied from a different, superior exemplar, on the basis of which corrections are introduced into the earlier portion of the text as well. This second exemplar appears to be A itself. However, the scribe frequently emends the text, correcting some of its most glaring linguistic aberrations. Thus far no evidence suggests that its departures from A reect anything more than the editorial activity of the scribe. Although hence of little value in constituting the text of the BraY , the a readings of this ms reect the interpretive activity of an educated reader, and are reported in all chapters of the critical edition. nak 1-143 (reported as C). ngmpp microlm reel no. a166/1. Paper, Nepalese N gar script; undated and unsigned. Complete in 440 folios. Descended from a A, but hopelessly corrupt, this ms offers little to reward the labor of its transcription. Its readings nonetheless are recorded in the apparatus of chapters i and lv in the critical edition. nak 1-286 (reported as D). ngmpp microlm reel nos. a165/13 and a1178/1 (lmed twice). Paper, Newari script; undated (the word samvat appears in the .

colophon, but with no number) and unsigned. Complete in 248 folios. Ff. 210233 are written in Nepalese N gar, perhaps replacing damaged or missa ing leaves. This ms is closely related to E, and it seems likely to descend from A, for in the sections collated, none of its variant readings appear to have independent value. Its readings are reported in all chapters except lv of the critical edition. nak 6-2608 (reported as E). ngmpp microlm reel no. a1319/4. Paper, Nepalese N gar script; undated and unsigned. Incomplete, with 136 folios. The readings a of this ms, written in a Newari-inuenced N gar, are very close to those of D. a

289 The two might share a (lost) examplar descended from A. Es readings are reported for chapter i alone. As alluded to, it is the working hypothesis of the present editor that codices D and E, and to a lesser degree B and C, are closely related, all furthermore being descended from A. While further collation is necessary to rule out other possibilities, I believe that the manuscript evidence for BraY i, collated fully in the critical edition, a is sufcient to sustain the premise that neither B, C, D, nor E provide variants not attributable to scribal error or, in the case of B, editorial activity. I will argue this in greater detail with the publication of the critical edition. Although not utilized in the present edition, there exists another codex of Nepalese provenance transmitting BraY ivvii, housed in the Sanskrit manuscript cola lection of Vi vabh rat University, West Bengal.4 Being written on paper in N gar s a a characters, this is of no particular antiquity. Further collation of the ms is required to determine if it offers readings independent of A; the sample studied so far suggests otherwise. In addition, S. N. Ghoshal Sastri has published chapter xlii of the BraY based upon an untraced ms transcribed by the late Harad s Mitra.5 The text a a transmitted is tantalizing, for it departs from A more signicantly than the other extant mss. However, there is no conclusive evidence that it represents an independent transmission.6 Whether or not Ghoshal Sastris ms descends from A, it appears to
4 Sanskrit Manuscript Section, Vidy -Bhavana, Vi vabh rat University. Old collection; no accesa s a sion number. 5 See chapter 1, section 2. 6 Although containing almost identical material, the order of verses in Sastris ms differs signicantly from A. Its readings also vary frequently, containing a comparatively large number of obvious corruptions. However, variants are also plentiful. Some of these are synonyms, giving no indication of the relationship between the mss. Many of its readings which differ from As appear to be products . of secondary editing. Note for example the crude attempt to correct As kanis. h n mikobhau tu angus.t a a t . hau parisamsthitau, which contains what appears to be a non-standard contraction of kanis. h n mikau . .t a a .t ubhau. Sastris ms reads kanis. h n mike dve tu angus. hopari samsthitau, which provides the correct fem.t a a . .t inine dual in the rst p da; but this still agrees with the masculine samsthitau. However, angus. hopari a . appears likely to be the original reading, in whatever manner it was arrived at. Note also the p da a transmitted by A as ubhayo hastayo []nguly (31c in A, 37a in Sastris ms). Here As text arrives at the a correct meter by non-standard sandhi, viz. hastayoh + ang yo ng (rather than yor ang ). In the . ms used by Sastri, there appears the grammatical but hypermetrical hastayor angulyah, with correct . sandhi, as well as the correct plural angulyah. While the BraY allows this sort of hypermetricism (cf. a .

290 represent a complex transmission. The degree of corruption and apparent reorganization of the verses suggest the possibility that the textperhaps contained in an anthologypassed through numerous copies. In the edition presented below, the critical apparatus is positive: the readings of all the mss used are reported, including their lacunae. Several normalizations are however silent: those of anusv ra (ivan, for instance, when it occurs before a dental a s consonant, is normalized to sivam), and cases of degemination (e.g. tatva for tattva) . and gemination (e.g. purvva for purva). However, when the mss are reported, their readings are reproduced exactly. The most doubtful readings appearing in the critically constituted text, whether or not emendations, have been identied through wavy underlining. In the apparatus, parentheses enclose poorly legible syllables, with those wholly illegible represented by hyphens. Syllables which have been corrected or marked as erroneous by the scribe appear in parentheses, followed by the superscript corr. and err., respectively. Short interlinear and marginal insertions are enclosed within a pair of plus signs. The upadhm nya, which occurs numerous a times in A, is rendered as f. Asterisks mark folio changes in A, the numbers for which are reported in the rst level of the critical apparatus, along with occasional notes on lacunae, marginal insertions, and so forth. Among the mss, only As readings are reported as unmetrical, when this is the case.

the annotation on BraY i.20), there are several other cases where Sastris ms violates meter in favor of a grammarquite the reverse of the language of the BraY , in which meter overrides grammar. Note for a example the unmetrical correction of bandhayet to badhny t (44d in Sastri, 47b in A). In some cases a Sastris ms contains better readings which could but might not reect later editorial activity. Note e.g. kim cid unnamitau karau (19b) for As unn matau (19d), and saramudr (37c) for As suramudr (32a). a a a . The two mss do however share obvious corruptions, an indication of possible afnity: note in particular s . tarjanmkuarupinm (38d in A) or tarjanmkuarupinm (Sastri 29d) for, presumably, tarjany ankuarupin. . s .. s


om namah sivadibhyo gurubhyo yogsvarnam i i. . . yat tattvam mantragarbham sakala ivamayam hetu nirv nabimbam s a. . . . .
:::::::::::: :::::::::::

a. dutn m padmasande samasukhavilasallingarupam bibharti | . . .. n n bhog dhiv sair vividhalayapadaih aktir varddhak nde a a a a a a. . . s tat tattvam vi vagarbham bhavanagadalanam bhairavam vah pun tu a . s . . . . srutv sastram pur dev mulatantram mahodayam | a a . . mudr mandalamantraugham vidy pthopalaksitam a a . .. . . sahasr ni da a dve ca catusptham tu bhairavam | a. s . . .
Codices: ABCDE A: f. 1v

Mangalam: om namah siv dibhyo gurubhyo yoge varnam ] em.; . . . svar(/ )n( ) A (aksara tops misss . a . a . . a . ing; bottoms consistent with D and E from guru ; nal anusv ra possibly lost); om namah siv ya B; erasure a . . a Cpc (before correction, perhaps om srgane aya namah); om namah siv dibhyo gurubhya yogsvarna DE . . . s . . . a 1a tattvam ] ABCE; tatva D garbham ] BC; garbha ] ADE mayam ] BCDE; maya . . . a. a A bimbam ] C; visv(am?) A; bimb d B; vi vam DE a s . 1b dutn m ] ABDE; dutn (- ?) C . . padmasande ] AB; padmakhande DE; (- ?)admasarnde Cac ; +padmasande+ Cpc sukha ] ABC; .. . .. . .. . .. (sra va?) D; sukha E s rupam ] AD; rupa E; k yam BC a . bibharti ] B; bibharttih AC; . . vibhurttih DE 1c bhog ] ABCpc DE a adhiv sair ] BCpc E; adhiv s(air?) A; adhiv sai a a a . D vividha ] ABC; dvividha DE laya ] conj. (Sanderson); naya ABCDE 1d bhairavam ] AC; bhairavo B; bhairavas DE 2a srutv ] BDE; srutv AC a a sastram pur ] B; sastra a . . pur AC (anusv ra possibly lost in A); mantraparo DE a a dev ] BC; dev(/ /o) A; devo DE a 2b mulatantram mahodayam ] em.; (m/s)u - - - - - - - A; p rvatim pratyuv ca ha B; p rvvatim pratyu+v ca a a a a . . . ha+ C; mulatantramahodayah D; mulatantramah ( - )yah E a 2c mantraugham ] em.; mantr(- gh . . . -) A; mamtr ni BC; mantroghe DE 2d pthopalaksitam ] ABC; pthe ya laksitah DE . . 3a . a. . . . da a ] ABCD; dda a E s s 3b catus ] corr.; catuh ABCDE bhairavam ] ABC; bhairavah DE . . .


292 vimal n nirgatam yat tad aghor bhmavikram | a a . pratyuv ca mah devam bhairavam mantravigraham a a . . yat pur sucitam deva tantram ucchusmasambhavam | a . . vimal hveti yat proktam j naugham saktipurvakam a . a . 4 3

yasmims tu samsthitam hy etac catusptham tu bhairavam | . . . . . . yasma hi nirgatam sarvam mantraptham mahodayam . . . .

catuspthasya sambandhe yat tvay coditam mama | a . . . . tantr vat rasamyuktam adyam yat siddhik ranam a a a . . . sarahasyam mam caksva j naughocc rapujitam | a a a . . ::::::: siva aktivibheda ca bindubhedam tathaiva ca s . nava aktivibheda ca srstibhedam suvistaram | s . .. . ek kin yath saktir navabhedair vyavasthit a a a 8

guhyak kimkarbhi ca kimkaryocchusmasambhav h | a . s a. . . yoginyo l kinn m tu bahubhedair vyavasthit h a a. a.

A: f. 1v 3c vimal n ] em.; vimal ABCDE a a nirgatam ] ABCD; nirggata E yat ] ACD; ye DE 3d . aghor ] ADE; ar+(- ?)+ B; a(- rgho?)r Cac ; a+gh +r Cpc a 3e mah devam ] ADE; mah deva a a . Bac C; mah devo Bpc a 3f vigraham ] ABCpc ; vigrahahm DE 4b tantram ucchusma ] Bpc ; . . . tantr m ucchusma A; tamtr(- n- ?)susma C; mantr(omre?)cchusma D; mantro(m -?)cchusma E a . . . . . . sambhavam ] ABC; sambhavah DE 4c vimal hveti ] ABCE; vis(a?)l hveti D a a proktam ] ADE; . . pro+kta+m Bpc ; proksa(- ?) Bac C 4d j naugham ] em.; j nogham A; (- n gham?) B; sten dyam a a a a . . . . . . D; sten (gh)am E; j(- n th- m?) C a a 5a yasmims ] A; yasmin BCE; (y- sm- n?) D hy . . . etac ] ACD; hatac D; heta E 5b catus ] corr.; catuh ABCDE 5c yasm hi nirgatam ] D; a . . . yasm (- - - - ) A; yasm t ptham tu tat B; yasm t ptham tu +tat+ C; yasm di nirggatam E a a . . a . . a 5d p . tham ] ABDE; ptha C . mahodayam ] ABDE; mahodayahm C 6a catus ] corr.; catuh ABCDE . . . . . . sambandhe ] ABC; samvatta D; samvartta E 6b tvay ] ABC; tvayo DE a coditam ] em.; codito . . ABCDE mama ] Bac ; mamah ABac CDE 6c tantr vat ra ] ADE; tamtr vat ram BCpc ; tamtroa a . . a a . . vat ram Cac a . samyuktam ] A; sampreks(- )m BC; samyukt m DE a 6d adyam ] ABDE; . . . . . . adya C 7a sa ] ABCE; sa(m?) D mam caksva ] A; sam caksva Bpc CED; sam caksve a a a . . . . ac B 7b j naughocc ra ] em.; j noghoc ra A; j n dy c ra BC; sten dyoc ra D; sten ghoc ra a a a a a a a a a a a a E 7c sakti ] BCDE; sakti A vibheda ] ADE; vibheda BC s 7d bhedam ] ADpc ; . ac (err.) pc bhedas BC; bhen D E 8a vibheda ] ADE; vibheda B C; vibhedada Bac s s 8c ek kin ] corr.; ek kin ABCDE a a . 8d bhedair ] ABC; bhedai DE vyavasthit ] Bpc ; vyavasthit h a a. ABac CDE 9a guhyak ] ABCpc ; (- - ?)k Cac ; suhyak DE a a a 9b kimkaryo BCDE; kim(. . yy ?) A a sambhav h ] em.; sambhavam ABCDE a. 9c l kinn m ] ABCE; l kinm m D a a. a a. 9d bhedair ] ABC; bhedai DE

293 ek eva mah vry vy pin sakti cottam | a a a a a tasy y gam a esam tu kriyate surapujita | a a s . . yath tath mah deva yoginsiddhik nksinam a a a a . . guru u rusanirate v mam rg nuvartine | s s . a a a advaitabh van vasthe nirvikalpe mane sthite a a 11 10

siddhir yath bhaved deva tadvidh n m vada prabho | a a a. yoginyah svalpabuddhyas tu alpacitt lpas ttvik h a a a. . bhartuh su rusanapar gurubhaktisamanvit h | a a. . s . . t s m siddhir yath deva bhavate ca sam satah a a. a a . svara uv ca a s dhu s dhu mah devi yat tvay ham procoditam | a a a a . nikhilam tat pravaksy mi sarvasamdohalaksanam . . . a . . 14 13 12

yath ca tantrasadbh vam bahvartham gudhavikramam | a a . . . sarahasyam mah bh ge srnusvek gram nas a a . . . a a a .


pur -d- akasm d deve i krdam nasya svasthitau | a a s . a

A: f. 1v
10a ek ] ABCD; eko E a vry ] ABD; vryyo CE a 10b vy pin ] BC; vy pin ADE a a . sakti cottam ] ACDE; saktir uttam Bpc ; sakti cottam Bac a a s a 10c tasy ] em.; tasya AD; tan me BC; a tamsya E y gam ] em.; yogam ABCDE a 10d surapujita ] em.; surapujitah ABCDE 10f . . . k nksinam ] BCDE; k ksinam A a . . a . . . 11a su rusa ] corr.; susrusa A; su una Bpc ; su osa Bac ; s . s . s . su rosa C; sa(nto?)sa D; mantrosa E s . nirate ] ABpc E; nirato Bac Ccorr.? ; nirat(- ?) D 11b . . m rg nuvartine ] ABC; m rgan tu varttine DE a a a 11c advaita ] AB; advaitad C; a(rdvai?)a D; arddheta E bh van ] ACDE; ( - ) van B a a a a 11d nirvikalpe ] em.; nirvikalpair ABCDE 12a bhaved deva ] em.; bhaved eva ABCDE 12b vidh n m ] ADE; vidh nam BC a a. a . 12c yoginyah ] Bpc ; yoginyo ABac CDE svalpa ] ABC; alpa DE buddhyas ] em.; buddh s a . AB; buddh ms C; buddh n DE a. a 12d citt lpa ] ACDE; cint +lpa+ B a a s ttvik h ] ABCD; a a. sotvik h E a. 13a su rusanapar ] corr.; su rusanapar A; su rusanapar h BC; santosanaparo Dpc E; s . . a s . . a s . . a. . . santrosanaparo Dac 13d sam satah ] A; namo namah BC; sam natah DE a a 14b yat tvay ] B; a . . . . . yatvay ADE; yatvayo C a pracoditam ] em.; pracoditah ABCDE 14c nikhilam ] ABC; ni ilat s . . DE pravaksy mi ] ABC; pravaksami DE a 15a tantra ] ACD; tatra DE sadbh vam ] ABCD; a . . . satbh vam E a . 15b bahvartham ] BC; bahv rtham A; bahv rtha DE a a vikramam ] B; vikram m a. . . . A; vikram h C; vikram DE a. a 15c bh ge ] ABC; se(no?) D; se(ro?) E a 15d ek gram a a nas ] corr.; ek gram nas h ACDE; aik gram nas B a a a a. a a a 16a pur -d- akasm d ] conj.; ( - rodakas)m a a a A; purodaka(- ?) B; pudodakamny C; purodakasm DE a 16b krdam nasya ] ADE; kridam nasya . a a . a . Berr. C

294 yadi y gavij t sy m icch y van mamotthit a aa a a a a icchay preritenaiva srkantho bhaktavatsalah | a .. . divyam varsasahasram tu ijy jalipurahsaram | a . . . . a ar dhito may devi bhakty vistena cetas a a .. a 17 16


srkanthena tato mahyam par karunay mahat | .. a . . a j naughas tu sam khy tah padabandhakramena tu a a a . . srkanthasya pras dena* sarvo yam parinato mama | .. a . . matsampark t tvay caiva a esa c vadh ritah a a s . s a a . tatas tvay hit rth ya ade ena vin priye | a a a s a parijanasya sam khy tum pr rabdham bhaktihrstay a a . a . ... a




viplapyam nam tam drstv mah tantram may punah | a . a a . ... a . . krodh vistena sapt si j nam te n sitam yay a .. a a . a . a
A: ff. 1v*2r 16c y gavij t sy m ] conj.; y gam vij n sy A; yogam vij n (ny ) BC; yogam vijen nyo DE a aa a a . a a a a a a a 16d . . icch y van ] conj.; icchay ca ABCDE a a a mamotthit ] em.; mamotthit h A; sam sthit BC; sam a a. a a a sthit h DE a. 17b srkantho ] ABC; srkanth( /o?) DE .. .. a bhakta ] em.; bhakti ADpc E; laksmi BC; . bhaktibha Dac 17c divyam ] ABCD; divya E varsa ] ABDE; varse C sahasram . . . . tu ] ABCpc DE 17d ijy ] BC; jy A; icch l DE a a a purahsaram ] B; purassaram A; . . a a purahsarahm Cpc ; puramsara(h?) Cac ; puratsarah DE 17e ar dhito ] ACDE; ar dhit B a . . . . . . may ] ABC; nay DE a a 17f bhakty vistena ] ABC; bhakt visnena DE a .. a .. cetas ] ABCDEpc ; a cetas hCac a. 18a srkanthena ] BCDE; srka(n- n- ?) A .. tato ] ACpc DE; tamo B mah. pc yam ] ABC ; seham DE 18b karunay ] em.; k runay AB; k r+u+nay C; k runayo DE a a . a a a a . . . . . mahat ] ABCpc ; sahat DE 18c j naughas ] A; j n khyam BC; steno(gha)n D; stena(ghe)n E a a a . sam khy tah ] ABD; sam (khy )(err.) tah E; mam khy tah C a a . a a a a . 18d padabandha ] ACpc ; padavedha . ac pc B; padeva C ; yadavatta DE 19a srkanthasya ] ABC DE; srkanthasye Cac .. .. pras dena ] ABE; a pram dena C; pras dana D a a 19b sarvo yam ] ABCpc DE parinato ] ADE; parinato B(err. under na) ; . . pari+na+to Cpc ; pa(di?)to Cac 19c matsampark t ] ABpc ; matsampat Bac ; (matsampa)corr. +rkk + a a . . C; satsamparkk t DE a tvay ] ABC; tvayo DE a 19d a esa c ] ADE; a esatv B; a esa+ s . s a s . a s . c + Cpc ; a esa(- a?) Cac a s . 20a tatas tvay ] A; tatas +tvay + B; tasta+s tva+y Cpc ; tata(s tam?)yo a a a . ac A ; tatas tvayo DE 20b priye ] ABCD; priya C 20c parijanasya ] BCDE; parij nasya A a sam khy tum ] em.; sam khy tam ABCE; samokhy tam D a a . a a . a . 21a vipl pyam nam ] em.; vipl pyaa a . a .. .. m na ABCDE a 21b mah tantram ] B; mah tantra ACDE a a 21c avistena ] ABC; avisnena DE . sapt si ] corr.; sapt si ADE; sapt (ni)err. B; (mam?)tr ni C a a a . a 21d j nam te n sitam ] corr.; j nam te a . a . a . . n sitam A; j nan ten nitam Bpc ; j nan tu n nitam Bac C; stenattenoditam DE a a a a a yay ] ABCDpc E; a . . . yath Dac a


295 tatas tvay mah bh ge trastay kampam nay | a a a a a a s srulocanay caiva bhumy m gatv tha dandavat a a a. a .. kar jaliputam krtv bhtay j naviplave | a . . . a a a vijapto ham mah devi sok dhisthitay punah a a a . . .. 23 22

tatas tv m vihval m drstv grhtah karunay hy aham | a. a . ... a . . . a evam ukt si k runy n mah manyubhrtena tu a a . a a . 24

bhurlokam gaccha deve e avat ram kurusva tha | s a . . .


br hmanasya grhe deham aparam grhna suvrate a . . . .



tatrasth y s tatas tubhyam bhakty ham sampracoditah | a a a . . . . anugraham karisy mi tav ham sakti-r- jay | a . a a . . a may s rdham punas tv aikyam tat sarvam pr psyasi priye a a a . . . tato vatrna madv ky t pray gasya sampatah | . a a a .
A: f. 2r 22a tatas ] ABDE; tata(- ?) C tvay ] ABCD; tvayo E a mah bh ge ] AB; mah bh ga a a a a C; mah bh ro DE a a 22b trastay ] A; trastayo+h+ B; trastayo CDE a kampam nay ] A; a a . kampam nayo(h?) B; kampam nayo CDE a a 22c s sru ] corr.; s sru A; n sru B; n sru CD; a a a a . n kra E a locanay ] AB; locanayo CDE a 22d gatv tha dandavat ] conj.; gatv (pa?)dama a .. . davat A; gatv pradamttavat B; gatv padam(bh/na?)vat C; padambhavat DE a a 23a kar jalipua . . . . tam ] corr.; kar jaliputam AD; kar jalipu(- )erasure m Bpc ; kar ksalip(ana?)m Bac ; kar ksali(pana?)m C; a a a . a . . . . . . . . . . kar jaliput(- ?)m E a 23b bhtay ] ABpc ; bhtayo Bac CDE a j naviplave ] A; (- ?)naviplave Bpc ; a . . ksanaviplave Bac ; ks(a/u?)naviplava C; stenavisnave DE 23c vijapto ham ] BC; vij ptoham A; a . . . .. vist(- ?)pto ha D; vi(- )space pto E mah devi ] ABCpc DE; mahodevi Cac a 23d sok dhisthitay a a .. punah ] corr.; sok dhisthitay punah A; lok dhisthitap vanah B; (- )dhisthitapovanah Cac ; +lok +dhisa a a a a . . . . .. .. . .. . thitapovanah Cpc ; sok dhisthitayo punah DE a 24a tatas ] ABDE; tata C tv m ] AB; tva C; tvom a. . . . .. DE vihval m ] B; vihval n A; vihval CDE a. a a 24b grhtah ] em.; grhtam ABCDE karu. . . . nay ] A; karu+na+y B; karuyo C; karunayo DE a hy aham ] AB; hrdam CDE 24c ukt si ] A; a . a . . . . u(kt )corr. ni B; a(- ?) ni C; rek ni DE a a a k runy n ] em.; k runy ACDE; k runy +( ?)+ B a . a a . a a . a 25a bhurlokam ] AB; bhurlloka C; bhurlloka DE 25b kurusvatha ] ABCE; kurusva(me)err. tha D 25c . . . grhe ] ABCpc DE; grha Cac 25d aparam ] ABDE; apara C grhna ] ABpc ; grhna Bac DE; grh(n?)a C . . . . . . . suvrate ] ABC; te vrate DE 26a tatrasth y s ] em.; tatrasth y ABpc DE; tatracch y Bac ; tatrasth yo a a a a a a a pc ac C tatas ] B ; tatos AB C; taton DE tubhyam ] ABCD; tu tvam E 26b bhakty ham ] ABC; a . . . bhakt (d)am DE a coditah ] B; coditam ACDE 26e may s rdham punas ] A; nay norddhvam a a a . . . . punas B; nay norddham punas C; sayonorddhasyanam DE a tv aikyam ] AB; tv aikyat Cpc D; tv ekyat . E 26f pr psyasi ] A; pr psyati B; pr (psya)corr. ti C; propsy(ani?) D; pr ps ne E a a a a a priye ] ABDE; priy(a?) C 27a madv ky t ] em.; madv kyo A; madv ky Bpc ; madb hyo Bac C; sadv kso DE a a a a a a a . 27b pray gasya ] A; prayogasya BCDE a sampatah ] A; nam vatah Bac C; nenvatah Bpc ; nemvatah a . . . . D; nemvratah E .


296 kanavre mah gr me meghadattagrhe subhe a a . . chandogasya mah devi utpann laksananvit | a a a . . sattik tatra samj t tava n mam na samsayah a a . aa . . .
::::::: ::::::::::::



tato mah tvay bhakty buddhisampannay hy aham | a a a a . a a ar dhito mah devi satatam lingapujay a . 29

tatra trayoda e varse siddh tvam saktyanugrah t | s a a . . khecaratvam av pnosi sampr pt ca mam ntikam a a . a a . s saktis tvam mah bh ge y sapt viplave krte | a a a a a . . pur may smar tm nam aghor n ma te dhun a a a a a a mady tvam mah saktih sarv nugrahak rik | a a a a a . . tat pravaksy mi te j nam yad bhrastam viplave krte a . . . a .. . a srkanthena mah n proktam bhutv sad sivat pad t | .. a a a .
::::::: :::::::::




sarahasyam mah devi srnusvek gram nas a . . . a a a .


acintyasya par saktih sivasya param tmanah | a a . .

A: f. 2r 27c kanavre ] ACD; karavra Bpc ; kanavra Bac E 27d meghadatta ] A; medhyadatta B; . . madhyadatta C; me(gh?)e datta DE 28a chandogasya ] A; cchandogamye BDE; cchandogamya C mah devi ] Eac ; mah dev ABCDEpc a a 28c sattik ] BDE; sa(n/t)tik A; sa(tvi)corr. k C a a a samj t ] ABC; samy ta DE aa a 28d n mam ] AC; n ma BDE a a 29a mah tvay ] C; mah (n/t)vay a a a a . . . A; mah t tvay B; mah nvayo DE a a a bhakty ] ABC; bhakt DE a a 29b hy aham ] ABpc D; hrdah C; . . . a hrtam E 29d pujay ] ADE; pujay(- ?) Bpc ; pujay(et?) Bac ; pujaye Cpc ; pujay(an/et?) Cac 30a . . tatra ] ABCD; tahtra E 30b siddh tvam ] Bpc ; siddh s tvam ABac DE; siddh stra C a a a saktyanu . . . grah t ] Bpc ; saktinugrah ACDE a a 30c khe ] ABCDpc E; kheh Dac av pnosi ] corr.; av pnosi AC; a a . . a+v +pnosi B; av pnoti DE a a 30d sampr pt ] B; sampr pto ACD; sa pr pto E a mam ntikam ] A; a . a a . a mam ntikahm B; mam ntikah C; samottikah DE a a 31a s ] ABC; so DE a tvam ] ADE; ca BC . . . . . mah bh ge ] ABDE; mah bh ga C a a a a 31b y sapt ] corr.; y sapt A; y (sa)err. pt B; y (m?) apt C; a a a a a a a a yonepto DE viplave krte ] ABpc CD; vipnave krte Bac ; viplavaksatam E 31c pur may ] ABC; a a . . . . a a a. a a a a puro mayo D; puro meyo E atm nam ] em.; atm n m A; atm n m B; atm n C; onm n m a a. DE 31d aghor ] ADE; ak r BC a n ma ] ABDE; n n ma C a a a 32a mady ] ABC; sadyo a DE 32b k rik ] ADE; k rik(oh?) B; k rik h C a a a a a. 32c tat ] ADE; tam BC te ] AD; t+e+ . . B; ta(n?) C; ta E j nam ] AB; j na C; stenam DE a . a 32d viplave krte ] ABC; viplavaksate . . . D; viplavakrte E 33a mah n ] em.; mah ABCDE a a proktam ] BCDE; prokta( ?) A 33b . . sad siv t ] conj.; sad siva+h+ AB; sad siva C; sad sivah DE a a a a a 33d ek gram nas ] corr.; ek graa a a a . . m nas h ACDE; aik gram nase Bpc ; aik gram nas Bac a a. a a a a a 34a acintyasya ] ABDE; acintanya C par saktih ] ADE; par sakti+h+ B; par sakti(h?) C a a a . . .

297 icch n mena samj t tay binduh prabodhitah a a a . aa . . 34

prabuddhasya tato bindor j naugham niskalam tatah | a . . . . abhivyakto mah devi akasm n mantravigrahah | a a . j nasampurnadehas tu sad sivapade sthitah a a . . . tasm t sad siv nuj tatah srstir abhut punah | a a a a . ... . huhuk nt vadhutasth tattvam l svabh vatah a a a aa a . lokasya hitak my y -m- amrt khyena suvrate | a a a .a nibaddham tu sam sena j naugham vimal tmakam a a a . . :::::::::::: anusthupchandabandhena sap dena mah tmane | a a .. laksasamkhyena samksep n mantraj nakriy tmakam a a . . . a . a nir c rapade bhutv puna cobhayadar an t | a a s s a par parena devena srkanth ya prabh sitam a .. a a. .
A: ff. 2r*2v 34c n mena ] ABCD; n ma na E a a samj t ] em.; samj tah A; samj ta+s+ B; samj ta CDE 34d . aa . a . . a . a binduh ] corr.; bindu ABCDE tay ] A; tayor B; tayo CDE a 35a prabuddhasya ] ABE; pratattasya . C; pra(bu - ?)sya D bindor ] em.; bindo ACDE; bindau B 35b j naugham ] ADpc ; j n nya a a a . corr. mark? C; (sta) naugham D; stanaugham E 35c abhivyakto ] ABC; abhivyakt DE a 35d . . .. . mantra ] ABDE; yantra C 35e j nasampurnadehas ] A; j nena purnadehas BCpc ; j nene pura a a . . .. nadehas C; stenasampurnnadehan DE 36a tasm t ] A; tasy BC; tasm n DE a a a sad siv nuj ] A; a a a . . sad siv (d ra)corr. ks+ + B; mah siv raksa C; sad siv trasta DE a a a a a a 36b tatah ] ABDE; tata C srs. .. . a . tir ] A; srstir BCDE . .. 36c huhuk ] A; huhuk BC; hrhuk D; hrhrak E a a a a nt vadhutaa . . . sth ] em.; nt vadhutasth m ACDE; nt m vadhutasth m B a a a. a. a. 36d m l ] em.; m l m A; s l m aa aa. aa. pc BC; sol m DE a. 37a k my y -m ] A; k my y B ; k my yo Bac DE; k sy yo C a a a a a a a a a a 37b amrt khyena ] A; na mrt khy na BC; na mrt ( ye)err. na D; na mrt khyena E a a a as a suvrate ] A; ced . . . . vrate B; ced vrata C; ce vrate D; suvrate E 37c nibaddham ] conj.; nibandhas A; nibandha+s+ B; . niba(ntam)err. C; nivattan DE 37d j naugham vimal tmakam ] corr.; j nogham vimal tmakah A; a a a a . . . . j naugham visal tmakam Bpc ; j naugham vimal tmakam Bac ; j n mjanasal tmakah C; stenoghyem a a a a a a. a . . . vimal tmakah D; stonoghyem vimal tmakamh E a a 38a anusthupchanda ] em.; anasthucchanda . . . .. .. AD; anusthuccha+nda+ B; anustucchanda DE; anasthaccha C bandhena ] A; vamcena B; . .. .. .. vacana C; vatvena D 38b mah tmane ] em.; mah tman ABDE; mah (tman )corr. C a a a a a 38c samkhyena ] ABC; samkhy na DE a samksep n ] A; samkhyay BC; samkseyo DE a 38d . . . a . . . mantra ] ADE; netracorr. B; (nn/tr?)etra C j na ] ABC; stena DE a kriy tmakam ] kria a y tmakah ABDE; kri+y +nmakah C a a 39a pade ] ACDE; +pade+ B bhutv ] ABCE; (kr)err. tv a . . . D 39b puna cobhaya ] A; puna( c bhaya)corr B; puna+h+ sadbhiya C; punar ca bhaya D; s s a s . punar c bhaya E s a 39c devena ] ABC; devena DE .






298 asm j j n n mah devi srkanthena hit ya vai | a a a a .. a


kotikotipravist rair lok n m hitak myay a a a. a a . .


prcchak srayabhedena kriy bhedavibh ga ah | a a a s . . suddh suddhena m rgena asattvena ca suvrate | a a . vist rit ni tantr ni j tv sad sivat pad t a a a. a a a a


a ayam tu j nasandoham svarup vasthitam priye | a . . . sap dalaksasamkhy tam may j tam yath rthatah a a . a a . a . . . tav pi j nabhrastay h sampravaksy mi s mpratam | a a a . .. a . . a


sap dalaksabhedena slok n m samsthitam tu yat a a a. . . . asm d vinirgatam sarvam trailokyam sacar caram | a a . . . tvay pi kathanyam hi lok n m hitak myay a . a a. a a divy divyasvabh vena sthitay saktyanujay | a a a a krodhabhairavadevasya siddhasyaiva sivecchay | a sap dalaksasamkhy tam evam vai bhairavo bravt a a . . . sap da caiva laksa ca krodhabhairavasamjak t | a a . .
A: f. 2v




40a asm j j n n mah devi ] em.; ( - m ?) j n( ) mah dev(i) A; +tasm j n mah devi+ Bpc ; srkaa a a a a a a a a a a nthena sam tm nam Bac C; asm steno mah devi DE a a . a a 40b srkanthena ] ABC; srkanth+e+na .. .. .. D; srkanthasa E .. vai ] ABpc DE; ca Bac C 40c kotikoti ] BCDE; kothikothi A prav. . . . ist rair ] corr.; pravist r(e?)r A; pravist ram B; pravist ra C; pravist rai DE a a a . a a 41b kriy ] ABC; a kriyo DE vibh ga ah ] ABC; vibh ga( )h D; vibh gatah E a s . a a 41c asuddhena ] ABDE; . . asuddha na C m rgena ] BC; m rgena A; morrona DE a a 41d suvrate ] ABC; sevrate D; . s(- ?)vrate E 41e tantr ni ] ABC; tantr ni DE a. a 41f j tv ] ABC; statv DE a a a sad siv t a a pad t ] conj.; sad sivah pad t ABCDE a a a 42a j na ] ABC; stena DE a sandoham ] ABDE; . . a mando ham C 42b svarup ] ABpc ; svak yo Bac C; svak y DE a a a priye ] ABDE; priya C . 42d may ] ABC; sa yo DE a j tam ] A; j +tam+ BC; stenam DE a . a arthatah ] ABDE; athata . . . C 43a tav pi j na ] em.; tay pi j na A; tay vij na BC; tay pi stena DE a a a a a a a bhras. tay h ] BC; bhrastay s A; bhrastay m DE 43d slok n m ] BCDE; slok n m A a a. a a. 44a . a . .. a .. a . asm d ] ADE; asm +d+ B; asm r C a a a 44b trailokyam ] ABDE; trailokam C sacar caram ] ABDE; a . . me car caram C a 44c tvay pi ] AB; tvayopi CDE a 45a svabh vena ] ABC; tvabh vena a a . DE 45b sthitay saktyanujay ] AD; sthitay saktyamnujay B; sthitay saktanujay C; sthia a a a a a . tap saksyanujay E a . a 45d siddhasyaiva ] ABDE; siddhasyeva C 45e sap da ] ABDE; may da a a corr. samkhy tam ] ADE; samkhy ta(m) a . a B; samkhy ta( - ?) C a 45f evam ] ABC; sevam DE C . . . . . 46a sap da ] A; sap da BCDE a a s laksa ] AB; laksa CDE . . s

299 kap labhairavasyaiva kathayisyasi suvrate a . 46

br hmanasya kuruksetre utpannasya mah mate | a a . . srdharetyabhidh nasya adhik rasthitasya vai a a 47

sakty dhisthitacittasya asiddhasya na samsayah | a . . .. kap labhairavo devi laksa caiva sap dakam a a . 48

caturvimsatibhi caiva sahasraih samgharisyati | s . . . ::: . . mulatantravidh nam tu svarupena vyavasthitam a . lok n m alpacitt n m catuspth divarjitam | a a a a. . . a asminn eva hy asau tantre siddhim pr psyati n nyath 50 a a a .
::::::::::::::::::::::::::::: ::::::::::


kap labhairav t siddh d asiddhasyaiva vaksyasi | a a a .


padmabhairavasamjasya evam vai bhairavo bravt . . odrade e tu j tasya devadattasya samjay | s a a . . s carana bahvrc sy tha ade ena na samsayah . . a a . . 52


asiddhas tv eva deve i padmabhairavasamjakah | s . .

A: f. 2v 46d kathayisyasi ] A; kathayisy mi BC; kathayisyeni DE suvrate ] ABC; sevrate D; s(e/u)vrate E . . a . 47a ksetre ] ABD; ksatre CE 47b utpannasya ] ABC; utpannasye DE 47c srdharety ] AB; . . srdharaty CD; srdharabhy E abhidh nasya ] ABC; abhidh nasye DE a a 47d adhik rasthia tasya ] ACDE; adhik ra(dhi)err. +sthi+tasya B a 48a sakty ] AD; sa(jy )err. +(- y?) + B; sajy C; a a a a saksy E . a 48b samsayah ] ABCD; ( a)err. msayah E s 48d laksa ] A; laksa BCDE sap daa . . . . . . s kam ] A; sap dakah BDE; na p dakah C a a 49a vimsatibhi ] BE; vinsatibhis A; vi atibhi D; s s s . . . vidr bhi C a s 49b sahasraih ] corr.; sahasrais A; sahasrai+h+ B; sahasrai C; sahasrain D; sahasain E . . samgharisyati ] A; sam(- ?)(ri)err. +ghi+syati B; ma(thya?)risyati C; sasyarisyati DE 49d svarupe. . . . . . . na ] AE; (svarupe)corr. na B; sva(- ?)pena D; cakr yana C a . vyavasthitam ] em.; vyavasthitah ABCDE . . . . 50a lok n m ] BE; lok n mm ACD a a a a. 50b catus ] corr.; catuh ABCDE pth di ] ABCE; . a . . err. pc ac pth(o) di D . varjitam ] AB ; vartitah B ; va(t?)ita( -?) C; varjjitah DE 50c asminn ] B; . . asmimn ACDE eva ] conj.; eka ABCDE hy asau ] A; (kra)err. +hy a+sau B; kramau C; hy . amau DE tantre ] AB; tatra C; tantra DE 50d siddhim ] em.; siddha ACDE; siddh+(i?)+m . . B pr psyati ] conj.; pr psyasi ABDE; pr psya i C a a a s 51a bhairav t ] A; bhairav BC; bhairavot a a D; bhairavon E siddh d ] conj.; siddho ] ADE; siddh BC a a 51b asiddhasyaiva ] ADE; asi(ai)erasure +ddha+syaiva B; atisaumyaiva C 52a odra ] A; udra B; uddra C; utra DE tu . . j tasya ] ADE; tu j+ +tasya B; bhuje tasya C a a 52b dattasya ] ADE; dantasya BC 52c bahvrc ] em.; (v)ahvaj sy tha A; va( - ?)many tha B; vahvaj ny(o) tha DE; vakramany tha C a a a a a 52d . a s s ade ena na ] ABC; ade e na DE 53a asiddhas tv eva ] ADE; asiddhas (tv)corr. eva B; aniddhastha ca C 53b samjak h ] ADE; samjay B; samjay h C a. a a. . . .

300 caturvimsatis hasram grantham dv da abhih punah a a s . . . . . samgh ram tu sahasrais tu karisyati sivecchay | a a . :::: . . anenaiva tu tantrena tatah siddhim pray syasi a . . .
::::::: ::: :::



etat tantram asiddhasya sak sat tava eva hi | a srunvisyanti mah bh ge sisy s caiva caturda ah . . a a . a s . raktabhairavako n mn jv l bhairavako parah | a a aa . hel bhairavaka caiva trayo py ete mah ya e | a s a s madhyade asamutpann s caranatharvanam tath s a a . . . v mabhairavako devi vijayabhairavako parah | a . saur str y m samutpannau sudrau j ty prakrtitau a. . a a . a a bhbhatsabhairavo devi gajakarnas tu bhairavah | . . candabhairavaka caiva sindhuvisayasambhav h s a. .. .
A: f. 2v
53c vimsati ] BDE; vinsati A; vi(m?) ati C s hasram ] ACDE; s+ +hasram B a a 53d . . . s . err. grantham ] ACDE; (guccham) +grantham+ B 54a samgh ram tu ] A; samkhy catu+h+ B; a . . . . a . . . samkhy catu C; sampyarin tu DE a sahasrais ] ABC; sahasran DE 54b sive ] ABC; sive . . DE 54c anenaiva tu ] ABDE; anena vartta C tantrena ] conj.; mantrena ABCDE 54d . . siddhim ] B; siddhih ACDE pray syasi ] ADE; praya(ccha)err. +sya+si B; pray syati C a a 55a . . etat ] ABC; etan DE tantram ] B; tatram AC; tatra DE asiddhasya ] ABC; sasiddhasya DE 55b sak sat tava ] corr.; sak s t tava A; nek n nte ca BC; nek nottava DE a a a a a a 55c srunvisyanti ] ADE; . . ( ruci)err. + ro+syanti +ca+ B; srucisyanti C s s mah bh ge ] AB; sad bhoga C; mah bh ro DE a a a a a 55d . . sisy s ] corr.; sisy s A; sikhy s BC; sisyo D; sisyo E . a . a a . s . 56a rakta ] A; ra(kta?) Bpc ; r( - - ?) Bac ; ruksa C; raksa DE bhairavako ACDE; bhaira+va+ko B n mn ] ABC; n mno DE a a a 56c . . pc bhairavaka ] B ; bhairavak(o ?) Bac ; bhairavako C; bhairavak s ADE s s s a 56d trayo py ete ] A; trayo (py? e)corr. te B; tray preta C; trayo py eta DE a mah ya e ] corr.; mah yase A; mahopame a s a Bpc ; mahopam(- h?) Bac ; mahoyamah C; mah yame DE a 56e madhyade a ] A; madhyade e BD; s s . . madhyadese C; madhy(e?) dese E samutpann s ] corr.; samu(tp)ann A; (samupte?)corr. nn B; a a a mah netr C; sam(re ca)nn DE a a s a 56f caranatharvanam ] conj.; caranam parvanam A; caranam . . . . . . . . . va(rddhan?)corr. am B; caranam varddhanam C; caranam sarvvanam DE 57a bhairavako ] ABDE; . . . . . . . . . . bhairavak C a 57b vijayabhairavako ] ADE; vijayabhairavo B; vijayam bhairav C a 57c saur sa. . tr y m ] ADE; saur s(tr - - ?)corr. B; saur str ya C a. . a. . a samutpannau ] em.; samutpanna ABCDE . a a. 57d sudrau ] em.; sudro ADE; sudr BC a j ty ] A; (j ?)corr. ty B; ( - tpa?)corr. C; jatyo DE a a a a prakrtitau ] em.; prakrttit(ai?)h A; prakrttitah BCDE 58a bhbhasta ] ADE; bbhatsa Bpc ; ( . . - - - ?) Bac ; srpadma C bhairavo ] ABDE; bhairav C a 58b karnas ] ABpc ; karnam Bac CDE . . . bhairavah ABpc DE; bhairavam Bac ; bhairava(mh?) C 58c bhairavaka ] Bpc ; bhairavak s ABac CDE s a . . . . 58d sindhu ] A; si(ndu)err. +ndhu+ B; mindu C; siksya DE sambhav h ] em.; sambhavah a. . . ABCDE





301 ksatriyau r japutrau tu candabhairavakah punah | a .. . . . br hmano tharvano devi caranena na samsayah a . . . . . yajasomasuto bhavyo brhodarvi abdite | s . gr me j to mah de*vi n tra k rya vic ranat a a a a a a . gr mab hye tu deve i tatra dev brhodar | a a s . tasy n mena sa gr mo brhodar prakrtit a a a a .




a ar dhayitv sau vipras tato devm brhodarm | a . . vidy m tram tu sampr pya japam tatraiva suvrate a a . . a . karisyati mah sattvas tatas tasya bhavisyati | a . . s s a ade a c sya sastrasya sravanaya na samsayah . . . srunvisyati mah devi padmabhairavap r vatah | . . a as . tatra siddhas tv asau vipras tantrakart bhavisyati a . g layitv imam c rtham satair astada air mitaih | a a . a . . .. s




A: ff. 2v*3r
59a ksatriyau ] em.; ks(e)triyo A; ksatriy BC; ksatriyo DE a putrau ] ABD; putrau(+h+?) C; . . . . . putr(o) E 59b canda ] ACDE; vanda B bhairavakah ] em.; bhairavak ABCDE a 59c .. .. . br hmano tharvano ] ABDE; br hmanatharvvana C a a 59d caranena ] ABpc DE; caranana Bac C . . . . . . 60a yajasomasuto bhavyo ] A; ya(ksa)err. +ja+(so)corr. (mah )err. +( - - ?)+bhavyo B; yaksa n ma mah a a . . a bhavy C; yajas sa(s/m)rato bhavy D; yajas samrato bhavy E a a a a a 60b brhodar ] ABpc D; . vrkodar Bac C; brhodar(i)err. E vi abdite ] corr.; visabdite ABDE; visabdita C s 60c . . corr. err. gr me j to ] AE; gr m(e) (ksa) +j +t(o)corr.? B; gr maksat m C; gr maj to D a a a a a a a 60d k rya vic raa a . . a. nat ] ABac CDE; k ryy vic rana Bpc a a a . 61a b hye tu ] ABpc DE; ( - - ?) Bac ; v hrta C a a . de. err. ve i ] CDE; devesi A; deve(si) B s 61b brhodar ] ADE; vrkodar BC 61c tasy ] em.; tasya a . . ABCDE n mena sa gr mo ] Bpc ; n mena so gr mo A; n m(a?)na( - ) gr m(o) Bac ; n manam gr m a a a a a a a a a a C; n manamogr ( - ?) D; n manamogr pi E a a a a 61d brhodar ] ABpc DE; vr( - )odar Bac ; vrkodar C . . . a a prakrtit ] Bpc ; prakrttit h ABac CDE a a. 62a ar dhayitv ] ACDE; ar dhayi(tya?)corr. B a vipras ] A; viprah BC; vipran DE 62b devm brhodarm ] em.; dev brhodar ABpc DE; dev vr( - )odar Bac ; . . . . . . dev vrkodar C . 62d japam ] A; japet BCDE suvrate ] AB; suvrata C; ca vrate DE 63a . karisyati ] em.; karisy mi ABCDE a mah sattvas ] ABpc DE; mah satv( - ) Bac ; mah satvah C a a a 63b . . . s s s tasya ] ABCD; tasyai E 63c ade a ] em.; ade o ABCDE sastrasya ] AD; (namta?)say B; g trasya a . . C; sastre sya E 64a srunvisyati ] em.; srunvisyanti ADE; (bhavi)err. syanti +( - ?)+ B; bhavisyanti . . . . . . C 64b p r vatah ] Bpc ; p rsvatah A; ( - - )tah Bac ; pujitah C; yottatah DE as a 64c tatra ] B; . . . . . tatrah ACD; tatah E tv asau ] ABDE; n(v)asau C 64d tantra ] A; +ta+(tu?) B; rtu C; . . . ac pc tatra DE 65a imam ] AB CDE; im m B a. c rtham ] conj.; c rth (n/t) A; v rth B; c rth C; a a a a a a . c rth( /?)t D; c rtht E a a a 65b satair ] B; satair ADE; matair C astada air ] corr.; astadasair AD; .. s .. asta(m)da(sai)err. r B; astodanai C; astadasai E mitaih ] A; mmitai BC; smitaih DE . . .. . .. ..

302 samharisyati tattvajas tath caiva trayoda aih a s . . . saptabhi ca tath caiva samgharisyati suvrate | s a . .


lok n ca hit rth ya n tra k rya vic ranat a a a a a a a .



n dhik r d yatah krtv j napr ptir na j yate | a aa a a . . a a vidy m travidh nam tu samksepena hit ya vai | a a a . a . . . sr vayisyati lok n m tatra tritayakena tu a a a. . kum rabhairavo devi astada a atam tath | a a . .. s s candabhairavakasy tha srutv vist rayisyati a a a .. . krodhabhairavako devi trayoda a atam tath | s s a . candabhairavak c caiva j tv vist rayisyati a a a a .. . 69 68 67

tejabhairavan m nas tath sapta atam punah | a a a s . . candabhairavak c caiva srutv sau vistarisyati | a a .. . avat re tu samproktam sisy nam tritayam tath a a . . . a. . . bhavisyakrtit hy atra ye caturda amadhyatah | a s . .
A: f. 3r 65c samharisyati ] A; samharisyanti BC; samdarisyati DE tattvajas ] corr.; tatvaj( ?)s A; tatvaj a a . . . . . . Bpc ; tadvaja Bac C; tatvajan DE 65d tath ] B; tath s ACDE a a trayoda aih ] Bpc ; trayodasaih s . . ADE; trayoda( - - ) Bac ; trayodanah C 66a saptabhi ] Bpc ; sapt bhi ABac CDE s a s ca tath ] Bpc ; cas a . tath A; cat tath Bac C; can tath DE a a a 66b samgharisyati ] A; samcarisyati BC; sampyarisyati DE . . . . . . suvrate ] ABCE; s(re)vrate D 66c lok n m ] BCE; lok na AD a a. a. 66d vic ranat ] Bac CDE; a . vic rana(t?) A; vic rana Bpc a . a . 67a n dhik r d ] conj.; (na?)dhik r A; adhik r BC; nadhik r DE a aa aa aa aa 67b j na ] ABC; stena DE a pr ptir na j yate ] Bpc ; pr ptin na j yate ADE; pr pti(nn arja?)yate a a a a a Bac ; pr ptinn arjayate C a 67d samksepena ] ADE; samksepana BC vai ] ADE; ca BC 67e . . . . . . lok n m ] ABpc DE; lok n mn Bac C a a. a a. 67f tritayakena ] corr.; trtayakena ABpc CDE; tatayakena Bac . 68a bhairavo devi ] em.; bhairavam devi ABCD; bhairavavande E 68b satam ] corr.; satan . . A; matan BCDE 68c bhairavakasy tha ] A; bhairavakany tha BC; bhairavakas(y?) tha D; a a a bhairavakasy rtha E a 68d srutv ] BCDE; srutv A a a 69a bhairavako ] ABpc DE; bhairavak a Bac C 69c bhairavak c ] em.; bhairavak s A; bhairavaka Bpc DE; bhairamvak s Bac C a a s a 69d . j tv ] ABC; statv DE a a a 70a n m nas ] em.; n m nam ABDE; n m na C a a a a . a a 70c can. da ] ACDE; cam( - ?) B bhairavak c ] em.; bhairavak s ACDE; bhairava+ka+ B a a s 70d . . srutv ] BCDE; srutv A a a vistarisyati ] ABCE; vistari(sya?)ti D 70e avat re ] A; avat ram a a . . . BCDE samproktam ] Bpc ; s(a)prokt m Bac ; samprokt m AD; sa prokt m C; samprokt E a. a. a. a 70f . . . . tritayam ] corr.; trtayan ABpc CDE; t(- ?)tayan Bac 71a bhavisya ] ACDE; bhavisya+(t?)+ B . . . . krtit hy atra ] ABpc DE; krttit d yatra Bac C a a 71b ye ] AB; ya CDE catur ] ABDE; catu( ?) C


303 kar labhairavo n ma tath ucchusmabhairavah | a a a . . m tangaj tisambhutau padmabhairava isyakau a a s. . yamabhairavaka c nyah k smre sambhavisyati | s a . a . . chandogo br hmano devi tath anyo bhavisyati a a . . visnubhairavan m no lamp y m visaye tath | a a a a. a .. . v jimadhyamdino vipro bhavisyati tath parah a a . . . 73 72 71

daksinabhairavah k sy m utpanno br hmanas tath | a a . a a . . . bahvrca c parah sisyo bhavisyati na samsayah . s a . . . . . oddiy ne mah devi tath sekharabhairavah | a a .. a . br hmanas taittirka ca apastambho bhavisyati a s . . caturda a sam khy t h padmabhairava isyak h | s a a a. s. a. j tv dv da as hasram siddhim pr psyanti suvrate a a a s a a . . 76 75 74

vy khy m caiva karisyanti sisy nam siddhik nksinam | a a. . a. . a . . . sakty dhisthitacitt n m caturda a tu samjak h a a a. s a. . ..
A: f. 3r 71d ucchusma ] ABpc DE; ucch(va?)sma Bac ; ucchvasma C 71e sambhutau ] em.; sambhuto . . . . . pc ac ABCDE 71f sisyakau ] em.; sisyagau AB ; sisya( - )au B ; sisyasau CDE . . . . 72a bhairavaka ] em.; bhairavak s ABDE; bhairavak m C s a a. c nyah ] em.; c ny h ABCDE a a a. 72b k smre ] B; a . k smre AD; k smra C; k smire E a a a sambhavisyati ] ADE; (te bhav?)corr. isyamti B; nan( - ?)risyati C . . . . . 72c chandogo ] ABD; chamdog C; cchanda o E a s br hmano ] ABpc E; br hman( - ) Bac ; br hmana a a a . . . . CD 73a n m no ] ADE; n m n( ?)corr. Bpc ; n m na Bac C a a a a a a a 73b lamp y m ] AD; lasy y m a a. a a Bpc ; vasy y m Bac Cpc ; vasy yom Cac ; lamy y m E a a a a a 73c madhyamdino ] em.; madhyamdine ABC; . . madhyadine DE vipro ] ABDE; vipr C a 74a bhairavah ] BC; bhairavaih AD; bhairavih . . . E k sy m ] corr.; k sy m ABpc Cpc ; k syom Bac Cac ; k ty m DE a a a a a a a 74b utpanno ] Bpc ; utpannau ADE; (u?)tpannau Bac ; anyatrau C br hmanas ] Bpc ; br hmanan ACDE; br hmana( -?) Bac a a a 74c . . . bahvrca ] em.; bahvayo ADE; bahava Bpc ; ba( - - ?) Bac ; bahvay C s s a c parah ] em.; c par s A; a a a . . c par BC; c par n DE a a a a sisyo ] AE; sisy BC; si(syo?) D . . a . 74d bhavisyati ] AC; bhavisyanti BE; . . bhavi( - ?)ti D 75a oddiy ne ] A; uddiy ne Bpc ; uddiy na Bac CDE 75b sekharabhairavah ] A; .. a .. a .. a . (se)corr. kharabhairava B; satvabhairavamh C; mekharabhairavah DE 75c br hmanas ] Bpc ; br hmaa a . . . . no ADE; br hman( -?) Bac ; br hmana C a a taittirka ] B; tettirka A; tattirka C; tentirka DE s s s s . . . 75d apastambho ] A; (apast)corr. ambo B; ( ru?)pantast C; ayastambho DE s a 76a sam khy t h ] Bpc ; a a a. sam khy t ABac C; samo khy to D; sam khy to E a aa a a a 76d siddhim ] ABpc DE; siddhi Bac C su. vrate ] ABCE; nuvrate D 77a vy khy m ] BE; vy khy ACD a a. a a karisyanti ] ABCE; kari(sya?)nti D . . . . 77b sisy nam ] corr.; sisy nam ACE; (si)err. sy nam B; (si - anam?) D . a. . 77c sakty ] ABpc ; sa( - ?) a . a. . . a. . ac B ; sajy C; saksy DE a . a 77d caturda a ] B; caturdda am (unmetrical) ACDE s s . samjak h ] Apc ; a. . samjak Aac BCDE a .


304 padmabhairavakam caiva sastham vai sth nam asritah | a . . . .. .


svacchandabhairavah srutv sak sat krodhabhairavah a a . . atha dv da as hasram sahasrair da abhih punah | a s a s . . . samharisyati deve e sakty dhisthitacetas s a a . . . .::::: ::::: 79


tantr vat ravicchinnam yoginn m prabh vatah | a a a. a . . kathayisyati lok n m da as hasrakam* priye a a. s a . . 80

ujjainy y m tu samj to viprajo ukaputrakah | a a. . a . deik tasya vai m t bahugarbhapras rit a aa a a 81 sn t c mati m t nam puratah putrak nksin | aa a a r. . a . . . . japtavidyo mah vryah samayalanghaprabh vatah a a . . ksipisyanti hy asiddhatv n m tar h sakticodit h | a a a. a. . . tasy garbhe mah bh ge amantrn makas tath a a a a a tatas tasya mah devi t s m caiva prabh vatah | a a a. a .
A: ff. 3r*3v 78a bhairavakam caiva ] conj.; bhairavak s caiva ABCDE a 78b vai sth nam ] ABCD; vai(sa . . . . th )err. vam E a asritah ] Bpc ; asrtah ACDE; as( - ?)tah Bac 78c bhairavah ] Bpc ; bhairav Bac C; a . . . . bhairavo ADE srutv ] corr.; srutv ACDE; (sru)err. tv B a a a 78d sak sat krodha ] A; sakop t a a krodha B; sakop (kr?)odha C; sak sa(ttro?)dha D; sak sa krodha E a a a 79b sahasrair ] ABC; sahasrai DE da abhih ] Bpc Cpc ; dda abhi (unmetrical) ABac Cac E; da abhi D s s s 79c deve e ] ADE; deve i s s . BC 79d sakty ] AE; sa(jj )err. +kty + B; sajj C; sa( - ) D a a a a dhisthita ] conj.; bhutas tu AE; .. bhutan tuBCD 80b yoginn m ] ABDE; yogin n m C a. a a. 80c kathayisyati ] ABC; kathayisyanti . . DE 80d priye ] ABDE; priya C 81a ujjainy y m ] A; ujja(yny ?)corr. n B; ujjany y n CDE a a. a a a samj to ] ADE; samj t BC 81b viprajo ] ABpc D; vipraj( /o) Bac ; vipraj C; piprajo E a a 81c . a . aa deik ] A; dak BCD; dak m E a a a. vai ] ABCD; cai E m t ] Apc ; m tr Aac BDE; m tro C aa a a a 81d pras rit ] em.; pras rit h ADE; pras dit h B; pram dit h C a a a a. a a. a a. 82a sn t ] ABCE; s(n ?)t D aa a a . c mati ] em.; c manti AB; v santi C; c mantri DE a a a a m t nam ] AD; m t( )corr. nam B; sotrnam C; a r. . a r . . . .. . m ttrnam E a .. . 82c vidyo ABpc ] ; vidy DE; vidy( /o?) Bac C a a vryah ] corr.; vrya ADE; . vdya+h+ B; vdya C 82d samaya ] A; s(ai?)ma+ya+ B; masalam C; samaye DE la. . ngha ] DE; lamghe A; lam( - ?) B; sam C 83a ksipisyanti ] corr.; ksipi yanti AD; (- ?)idhinyatti s . . . . . . Bpc ; ksidhin rtti Bac ; ksimdhin rtti C; ksipa yanti E a a hy asiddhatv n ] D; hy asiddhatv (t/n) A; a a . . . . s ha siddhatv n BC; hy a+si+ddhatv n E a a 83b m tar h ] corr.; m tar AD; m tarah B; m taro CE a a. a a a a . 83c tasy ] em.; tasya ABCDE a garbhe ] AB; garbha CDE mah bh ge ] AB; mah bh r C; a a a aa mah bh ro DE a a 83d n makas ] ABpc ; n makan Bac CDE a a 84a tasya ] A(ta+sya+); tatra BC; te s DE a 84b t s m ] A; (t ?)s m B; bh sam C; t s DE a a. a a. a. . a a prabh vatah ] ABDE; prabh vata C a a .



305 vidy m pr pya japam krtv tatah sastram sa vetsyati a. a . . a . .



a tato nibaddhagrantha ca divyasang nubh vatah | s a .


da as hasraken rtham a esam kathayisyati s a a s . . .


tatas tenaiva j nena pa c t siddhim sa lapsyati | a s a . candabhairavan m nah sahasraih saptabhih punah a a . .. . . . tad eva da as hasram karisyati mah dhipe | s a a . . dv da aiva sahasr ni kartuv ch bhavisyati a s a. a a .



na c rtham divya isy nam samhartum sah karisyati | a s . a. . . . . . . saptabhi ca sahasrais tu vighnam tasya bhavisyati s . . anay v chay devi bindubhairavasamjakah | a a a . . saptabhi ca sahasrais tu tasya vighnam bhavisyati s . anayaiva mah devi v chay da asamjake | a a a s .



m y bhairavan m no na ca siddhim pray syati a a a a a .


saptabhi ca sahasrais tu vighnam tasya bhavisyati | s . . anantabhairava caiva vistaram kartuv chay s a a .
A: f. 3v 90a91b omitted in E


84c vidy m ] B; vidy ACDE a. a japam ] ADE; jayam B; (ja)corr. yam C 84d sa vetsyati ] conj. . . . (Isaacson); bhavetsyati A; bhavisyati BC; bhavatsyati DE 85a grantha ca ] conj.; granthasya s . a ABCDE 85b sang nubh vatah ] ADE; sag trabh vatah Bpc ; sadbh vabh vatah Bac C a a a a a 85c . . . s hasrakenartham ] corr.; s hasraken rtham ] AB; s hasraken tham DE; s hasrakam n thamm a a a a a a . . . . a . C 86b lapsyati ] BC; lapsati ADE 86c n m nah ] corr.; n m na ADE; n m nam BC a a . a a a a . 86d sahasraih ] ABC; sahasrai DE saptabhih ] ABC; saptabhi DE punah ] ABCD; pu(. . . na)err. h E 87a eva ] ADE; (evam?) Bpc ; arddham Bac C 87b karisyati ] ABCD; karisyamti E . . . . . . . mah dhipe ] ABDE; mah dhipa C a a 87c dv da aiva ] DE; dv da( ai?)va A; dv da(sai)err. (va)corr. B; a s a s a dv dasam ca C a 87d kartu ] Bpc ; kartt ADE; kant Bac C a a 88a na c rtham ] em.; na c rtha a a . . ADE; na c rthe B; tatr rtha C a a sisy nam ] BC; sisy nam ADE . a. . a. . 88b samhartum ] conj.; sa . . . marttum ABpc ; na narttum Bac C; mamantram DE sah karisyati ] em.; so karisyati A; n karisyati a . . . . . . . Bac DE; no karisyati B; n katisyati C a 88d vighnam ] ADpc DE; vi( - ?) Bac ; visy C 89b . . . . a bindu ] ABCD; bindra E 89d tasya vighnam ] A; (bhavisyati)err. +tasya vighnam+ B; bhavisyatih . . . . C; tasya vighna DE 90a anayaiva ] ABD; anenaiva C 90b v chay ] BC; v cchay AD a a a a samjake ] conj.; samjakah ABCD 90c n m no ] AD; n m nam BC a a a a . 90d na ca ] BC; . . . nava AD siddhim ] C; siddhi AD; siddhi+h+ B pray syati ] ABC; pray myati D a a 91a . . sahasrais ] ABC; sahasr( - ?)s D 91b vighnam ] AD; ( - na?)n Bpc ; vipran Bac C .

306 sahasraih saptabhi caiva vighnam tasy pi suvrate | s a . . bhavisyati na sandeho evam vai bhairavo bravt . . sad sivena devena dv pare bh sitam mah n | a a a. a . tato divyena m nena tasm t saptatime yuge | a a tava devi may khy tam tantram bhairavapujitam a a . . kalau yuge na sandehah srkanthasy jay tath | a a a . .. asm d vai saptame caiva kap lsasya suvrate a a 94 93 92

tvam vaksyasi mah devi tret y m bhairavo bravt | a a a. . . ::::::::::::::: dv pare kalisandhau tu padmabhairavasamjakah a . . sap dalaksasamkhy tam samgharisyati n nyath | a a . a a . . . . kalau caturthap de tu tath svacchandabhairavah a a . samgharisyati deve i evam vai bhairavo bravt | s . . . candabhairavaka caiva tath ca vibhubhairavah s a .. . m y bhairavaka caiva vist ram kartuv chay | a a s a . a a kalau caturthap d nte bhavisyanti var nane a a a .
A: f. 3v 92a sahasraih ] Bpc ; (sahasrais?)bottoms damaged A; sahasrai Bac CDE 92b vighnam ] ABpc DE; vi( . . ac ?)n B ; vi(ppra?)n C tasy pi ] ABC; tasyo pi DE a suvrate ] AB; suvrata C; nuvrate DE 92c sandeho ] ADE; sandeha BC 92d evam vai ] A; evai vai Bpc ; evai ca Bac C; eva( cai?) s s s . pc DE 93b bh sitam ] ACD; bh sito B ; bh sit( - ?) Bac ; bh sitamn E a. a. a. a. mah n ] ABpc DE; mahat a . . Bac C 93c divyena m nena ] AB; divyanam nena C; divyanamonena DE a a 93d tasm t ] ADE; a tasm n BC a saptatime yuge ] A; (me bruhi?) sat pur B; me (bru?)hi sat pur C; saptatima a a pur DE a 93e tava ] ADE; tadva BC akhy tam ] ABC; a(gy ?)err. +khy +tat D; akhy ta a . a a a E 93f tantram ] ABC; tatra D; tat tatra E pujitam ] ABpc ; pujita( - ?) Bac ; pujitah . . . CDE 94a yuge ] ABpc ; pur Bac CDE a sandehah ] DE; sandeheh A; samdeha BC 94c . . . saptame ] AB; saptama C; saptase DE 94c suvrate ] ABCE; nuvrate D 95a tvam ] AD; tvad . BC; tv(am?) E vaksyasi ] em.; vaksy mi ABCDE a 95b tret y m ] ABCD; tre(tt ?)y m E a a. a a. 95c . . sandhau ] corr.; samdhaus ADE; samkhaus B; sakhyaus C 95d padmabhairava ] ABCD; yad a . . bhairava E samjakah ] em.; samjakau ABCDE 96b samgharisyati ] A; sam(ka?)corr.? risyati . . . . . . . B; samparisyati C; sampyarisyati D; sapyarisyati E 96c p de ] ABDE; p dan C a a 96d . . . . . bhairavah ] BC; bhairava ADE 97a samgharisyati ] ABpc ; sampyarisyati Bac CDE de. . . . . ve i ] CDE; devesi A; deve(si)err. B s 97b evam vai ] ABC; evams cai DE bhairavo ] ABDE; . . bhairav C a 97c bhairavaka ] ADE; bhairavaka+ + B; bhairavakam C s s 97d vibhu ] ABDE; . bindu C 98b vist ram ] ABDE; vist ra C a . a v chay ] ABDE; v chay C a a a a





307 anantabhairava caiva kalp nte vistaram sad | s a a . na saknosyati vai kartum bhakty dhisthitacetasah a . . . ..


caturvimsatis hasram m nayisyati suvrate | a . . a . candabhairavako devi vibhubhairavam eva ca .. m y bhairavaka caiva tath c nantam eva ca | a a s a a etad apa cim vr s tantram j tv na samsayah s a a . a a . .



siddhim pr psyanti deve i kalp nte bhairavo bravt | a s a . etac ch stram kalau c nte yo*ginyah sakticodit h a a a. . . apahrtya pray syanti samprad ya ca suvrate | a a . saktyantam n tra sandeha evam vai bhairavo bravt . a .



krte yuge mah devi tret y m dv pare tath | a a a. a a .: n vat ro sya sastrasya sucito bhairavena tu a a . 104

kalau yuge puna caiva evam eva mah dhipe | s a avat ro sya sastrasya karisyasi na samsayah a . . . dv da aiva sahasr ni n dhik ni man g api | a s a. a a a
A: ff. 3v*4r 99b kalp nte ] ABDE; kalp nta C a a 99c saknosyati ] conj.; sakyosyati A; sakty syati Bpc ; sakty a. a . . ac syabhih B C; sa(sko?)syati DE . 99d bhakty ] ABC; bhaksy DE a cetasah ] ADE; cetanah . . . . . . a BC 100a vimsati ] BC; vinsati A; vinsati DE 100b suvrate ] ABE; suvrata C; nuvrate D . 100d vibhu ] ADE; bindu BC 101a bhairavaka ] ABC; bhairavaka D; bhaira+va+ka E s s 101b c nantam ] ABC; c ntam DE a a 101c etad ] ADE; eta( - ?)corr. B; etat C apa cim ] A; (s a ?)pa cim B; tapa cim C; apa ci E s a s a s vr s ] ABpc DE; vr Bac C a a 101d tantram ] ABCD; tantra E . pc ac samsayah ] AB DE; samsayan B C 102a deve i ] BCDE; devesi A s 102b kalp nte ] ABDE; a . . . kalp nta C a 102c ch stram ] BC; ch stra ADE a a c nte ] ADE; c (nte)corr. B; n ntu C a a a 102d . yoginyah ] ADE; (v?)yoginyah Bpc ; ( - - - ?) Bac ; vy senah C a 103a apahrtya ] ABCD; apahrta . . . . . E 103b samprad ya ] em.; samprad ya ADE; samprad +ya+ B; samprad s C a a s a s a 103c . . saktya ] em.; sakty ABpc ; sa(- ?) Bac ; sa(r - ?) C; saksy DE a a a . a sandeha ] em.; sandeho A; sandeho BCDE 104a krte ] Bpc ; krtau ADE; krtv Bac C yuge ] ABpc ; pur Bac CDE a 104c . . . a n vat ro sya ] ABpc D; n vat r nya Bac C; n vat rasya E a a a aa a a sastrasya ] BC; s strasya ADE a 104d sucito ] ABCD; sucito E bhairavena ] ABCD; bhairavena E 105a yuge puna ] ABpc ; pur s a . ac (nata?) B ; pur +(na?)+ta C; pur puna DE s a s a s 105b adhipe ] ABpc DE; adhipa Bac C 105c avat ro sya ] ABpc ; avat r nya Bac C; avat r sya DE a aa aa 105d karisyasi ] A(darkened) ; karisyati BCD; . . pc karisyamti E 106b n dhik ni ] ADE; n dhik (n?)i B ; n dhik ri Bac C a a a a a a man g api ] ABC; a . . man sapi DE a


308 kathayisyasi deve i lok n m hitak myay s a a. a a . 106

kum rdvpav stavy ye lok h samsthit h priye | a a a a. a. . tesam sastrasya n nyasya pr pti caiva bhavisyati a a s . . . kal pagr make devi tatah sah samgharisyati | a a . . . .


kaliyugasya adau tu avat ram karisyati | a . . dv da aiva sahasr ni n tra k rya vic ranat a s a. a a a . sap dalaksam deve i kathayisyasi suvrate | a s . . . svacchandabhairavasyaiva evam vai bhairavo bravt . dv da aiva sahasr ni samhrt ni mah dhipe | a s a. a . .a yena tasya mah bh ge n tra k rya vic ranat a a a a a .




da abhis tu sahasrais tu -m- uktv vracatustayam | s a .. tato sau codito devi tvayaiva varavarnini . 111

kal pagr make sthitv visnubhairavakasya tu | a a a .. asiddha caiva deve i sastram dv da asammitam s s a s . . sahasrair n tra sandehah kathayisyati suvrate | a . .
A: f. 4r 106c kathayisyasi ] ADE; kathayisya+ti+ B; kathayirsya+mi+ C deve i ] BCDE; devesi A s 107a . . . dvpav stavy ] ADE; (- pav sta)corr. vy B; ddh vaco mrsy C a a a a 107b ye ] ABC; ya DE .. a lok h ] corr.; l(o?)k A; lok BC; lauk DE a. a a a samsthit h ] corr.; samsthit ABCDE a. a 107c . . sastrasya ] A; sastra+sya+ B; sastra C; sastranya DE 108a kal pa ] A; kal (pa)corr. B; kal va a a a C; kal pe DE a 108b tatah sah ] A; ta(ta sah)corr. B; tat tasyah C; tatasmah DE samgharisyati ] A; . . . . . . . corr. sam(ha) risyati B; sambharisyati C; sampyarisyati DE 108c kali ] ABDE; (ka)corr. li C 108f . . . . . . k rya ] ABac DE; k ryy Bpc ; k yya C a a a a vic ranat ] ACDE; vic rana(t)err. B a . a . 109a sap da ] ACDE; a sa(p ?)da B a 109b kathayisyasi ] ABpc ; kathayisy mi Bac C; kathayisye i DE 109c bhairavas. . a . s pc ac yaiva ] ADEB ; bhairava caiva B C s 109d evam vai ] AD; evam ca BC; eva cai E s bhai. . ravo ] ABpc DE; bhairav Bac C a 110b samhrt ni ] ABCD; samhrt ni E a a. mah dhipe ] ADE; a . . . . mah dhip(e)corr.? B; mah dhipa C a a 110c yena ] ADE; ye+na+ B; pura C tasya ] em.; tasy AC; a tasm +t+ B; tasyo DE a mah bh ge ] AB; mah bh ga C; mah bh ro DE a a a a a a 110d k rya ] ABac CDE; a k ryy Bpc a a vic ranat ] ACDE; vic rana(t)err.. B a . a . 111b muktv ] ABpc ; (vakt ?) Bac C; nukt D; a a a nuktv E a catustayam ] A; catustaya( - ?) B; catustayah CDE 111c codito ] ABDE; codit C a . .. .. .. 111d varnini ] ADE; varnin BC 112a kal pa ] ABC; kal ya DE a a 112b visnu ] ABDE; vis. . .. . na C bhairavakasya tu ] ABDpc E; ( - - - -)corr. tu Dac ; bhairavakampatu C 112c deve i ] BCDE; s . devesi A 112d sastram ] B; sastra ACDE sammitam ] A; sammita(m)corr. B; sammitah DE; . . . samsmitah C 113a sahasrair ] ABDE; sahasrai C sandehah ] ADE; sandeha+h+ B; samdeha . . . . . C 113b suvrate ] ABCE; nuvrate D


309 visnubhairavako devi kum rdvpav sin m a a a .. 113

kathayisyati lok n m sakty dhisthitacetas m | a a. a a . .. sahasr ni da a dve ca evam vai bhairavo bravt a. s . grhe grhe mah devi yath sapta at ni ca | a a s a . . tath dv da as hasro bhavisyati na samsayah a a s a . . . sakty dhisthitacitt n m n tra k rya vic ranat | a a a. a a a . .. grhe grhe mah devi ye pums h siddhibh jan h a a a. . . . a. striyo v siddhibh ginyas tesam api grhesv atha | a a . . .




pracarisyati deve i evam vai bhairavo bravt s . .


asiddhibh jan ye tu puruso tha striyo tha v | a a a . vidy m tram api caiva na pr psyanti mah dhipe | a a a a sarahasyam mah devi j syante siddhibh jan h 118 a a a a. . etat tantr vat ram tu srkanthena yath sthitam | a a . .. a kathitam mama deve i tath pi kathitam may s a a . . 119

s mpratam sarahasyam tu sarvasandohalaksanam | a . . . . mah bhairavan m nam srnusvek gram nas a a a . . . . a a a

A: f. 4r 113c bhairavako ] ABpc DE; bhairavak Bac C a 113d dvpav sin m ] ABpc ; ( - ?)pav ( - ?)n m a a a a. ac B ; ddhya yogin m C; dvpav sin h DE a. a a. 114b sakty ] AE; sa( - y ?)corr. B; sajy C; saksy a a a . a D 114d vai ] ADE; ca BC 115b sapta ] ABDE; masta C 115c s hasro ] AB; s hasr a a a .. CDE 115d bhavisyati ] ABCD; bhavisyamti E samsayah ] BCDE; samsaya(h?) A 116a . . . . . . . sakty ] ADE; sa(kty ?) Bpc ; sarpy Bac C a a a 116b k rya ] ABac CDE; k ryy Bpc a a a vic raa. a nat ] ACDE; vic rana(t)err. B a . 116d ye pums h ] corr.; ye puns A; (ma)err. ye pumsah Bpc ; ma a. . . . . yat praj Bac C; ya puns D; * E a a 117a bh ginyas ] ABpc D; bh giny(a - ?) Bac ; bh minyamn a a a . . C 117b tesam ] C; tesamm AD; tesa(m?) B api grhesu ] em. (Isaacson); adhigrhesv A; . . . . . . . adhi(grhesv)corr. B; adhimrdusv C; adhigrhesv D 117c pracarisyati ] ABpc D; pracarisyamti Bac C . . . . . . . . 118a asiddhi ] em.; asiddha ABCD bh jan ] ABpc D; bh van Bac C a a a a ye tu ] AD; +ye+ tu B; tu C 118b puruso tha striyo ] AD; purusarth(e?) striy B; purusarthaistriyoh C a 118c . . . . api ] ABD; ayi C 118d mah dhipe ] ABD; mah dhipa C a a 118f j syante ] A; j n nte BD; a a a j nan tu C a 119a tantr ] BC; tatr AD a a 119b sthitam ] A; sthita(m?)corr. B; sthitah . CD 119c kathitam mama ] ABC; kathitasme a D s 119d kathitam may ] ABC; kathitay D a a . . 120a s mpratam ] ABC; s npratam D a a sa ] ABpc D; (me?) Bac ; me C 120b sandoha ] AD; . . samdeha BC 120d srnusvek gra ] AD; srnusvaik gra Bpc ; srnusvak gra Bac C . . . a . . . a . . . a .


310 y s saktih pur khy t nant dy nantasambhav | a a a aa a a a . tasy bhedam mah bh ge kathay mi yath kramam a a a a a . . yo sau acintyam ity ahuh si*vah paramak ranah | a . . . nihsamjo nirvik ra ca vy p santas tathaiva ca a s a . . nihsvabh vo mah devi kriy k ranavarjitah | a a a a . . . niskalo nirvikalpas tu arupo gunavarjitah . . .




nirmamo nirahamk ra advaitapadasamsthitah | . a . . yogin m dhy nagamyo sau j narupo mah ya e a. a a a s 124

nir c rapad vasthah samj m trah prabhuh parah | a a a . . a a . . . tasy par jyotirupam sarv nugrahak rakah a a a a . .


vy p hy avyaktarup ca amanasko mah tmanah | a a . tasya saktir mah devi svabh votth akrtrim a a a . a 126

. jyotsn rup svarupena sph tikasyeva ra mayah | a a a. s .

A: ff. 4r*4v 121a saktih ] corr.; sakti AD; (sa)err. kti+h+ B; sakti C 121b anant dy nanta ] AD; anant s a a a . . pc ac c nanta BC a 121c mah bh ge ] AB ; mah bh ga B C; mah bh ro D a a a a a a 121d kramam ] AD; . krama( - ?) B; kramah C 122a acintyam ] ABD; acityamm C ahuh ] ABpc D; ahu Bac C . . pc ac 122b k ranah ] AB D; k ( - nah) B ; k ( - ?)nah C a . . a a 122c nihsamj ] ABpc D; nihsamsk . . . . . . . . a Bac ; nihsamsk C 122d vy p santas ] ABpc ; vy (y - ?)s Bac ; vy y s ntus C; vy pau santan D a a a a a a a . . . a 123a nihsvabh vo ] Bpc ; nisvabh vo A; ni vabh v Bac C; nimbabh vo D a a s a a a 123b kriy ] ABC; kriyo a . pc ac D 123c niskalo ] B ; niskalo AD; niskal B C vikalpas ] ABC; vikalpan D 123d . . a a arupo guna ] ABpc ; amk (n - aste?)na Bac ; amk n(y/p) strena C; a(nk ?)po guna D a 124b . . a . . . a. . . advaitapada ] ABpc D; advaita( - ?)da Bac ; advaitadbheda C 124c yogin m ] em.; yoginy m AD; a. a. yoginy+ m+ B; yogin C a. a gamyo sau ] ABpc ; (m ny - ?) Bac ; m ny (sau?) C; gamy sau D a a a a a a 124d j narupo ] ABpc ; j nak ya Bac C; j na(n k )err. +ru+po D a a a a mah ya e ] corr.; mah yase A; a s a mah (ya[se]err. )corr. B; mah maya C; mah yame D a a a 125a avasthah ] ABpc ; ava( - ?) Bac ; avastha C; . avastheh D 125c jyotirupam ] em.; jyotirupa ABpc ; (jo?)tibh (p/y)a Bac ; jotibh ya C; skotirupa a a . . . D 125d sarv nugraha ] ABpc D; saccntagraha Bac C a 126a hy avakta ] ABpc ; hrvakta Bac C; . h(y a?)vyakta D rup ] ABpc D; k ( - ?) Bac ; k s C a a. 126b amanasko ] A; anavastho B; ananasth C; anenasko D a mah tmanah ] A; mah ( - - h) Bpc ; mah ( rayah?) Bac ; mah (a/ ra?)yah a a as a s . . . . C; mah nsanah D a 126c saktir ] AD; sakti+r+ B; sakti(h?) C 126d svabh votth ] ABpc D; a a . . svabh v sth Bac C a a a akrtrim ] Bpc ; akrttim ABac D; akrrttim C a a a 127a jyotsn rup ] ABpc ; a a . . . ac pc . sant k yo B ; ksant k yo C; (sko?)nsn rupo D a a a svarupena ] AB D; svakosena Bac C 127b . a a . . . sph tikasyeva ] em.; sph tikasyaiva ] ABpc D; sph tikasyava Bac C a. a. a. ra mayah ] Bpc ; rasmayah AD; ra( s . . - ?)mayah Bac ; rammayah C . .

311 a tasyecch nirgat saktir j narup manonman a a a pravartate nir bh s avadhuteti s smrt | a a a a .a prabodhayati s nant bindun dau ksanena tu a a a . .



kundal krtisamsth n svar dau samvyavasthit | a a a a .. a . . . caturbh gavibhakt s caturbh gavibh jit a a a a a a 129

evam kundalin saktih svaraih soda abhih sthit | a . .. . . . . s . catuskapathakopet pacavyoma-alamkrt a . .a . 130

a evam pacavidh s tu saktir ady manonman | a a . nav ksaravidh nena puna caiva praj yate a . a s a 131

svaravyajanasamyukt pac saksarasamyut | a a . a . . ::::::::: a avadhut mah devi navabhedair vyavasthit a a 132

s atra devyo tha dutya ca yoginyocchusmam tar h | a a. . samast n srjate devi sivecch m anuvarttin a . a 133

iti mah bhairave tantre a dv da as hasrake picumate nav ksaravidh ne a s a a . a sambandhapatalah prathamah . . .
A: f. 4v 127c tasyecch nirgat ] ABpc ; tasya dv ra mah Bac C; tasyacch nirgat D a a a a a a saktir ] B; sakti ACD a 127d rup ] Bpc ; rupo AD; (k y ?) Bac ; k y C a a a a 128a pravartate ] ABpc D; prava(n/t - a?) Bac ; pravattana C 128b avadhuteti s ] ABpc D; avadhut(a - m ?) Bac ; avadhutabhim C a a a 128c pc s nanta ] AB D; s ran (tu?) Bac ; s ran tu C a a a 128d ksanena ] ABpc D; ksa( - )na Bac ; ksarena . . . . . C 129a samsth n ] ABpc D; samtth n Bac ; mam(tth ?)n C a a a a a a 129c vibhakt s ] AD; a a . . . a vibha( - a s ?) Bac ; vibhajy n C a a 129d vibh jit ] AD; vibh+ +jit B; vibhamjit C a a a a 130a . a saktih ] ACD; sakti+h+ B 130b sthit ] em.; sthit h ABCD a a. 130c catuskapathakopet ] ABpc ; a . . . catu( - - ?)thako( - )t Bac ; catusthayathakosat C; catuskapathako( - - )erasure D a 131a paca ] ABpc D; .. . a . ac a (enpa?) B ; pa(nya?) C vidh s ] AB; vidh n n C; vidh m D a a a a a a 131b ady ] ABC; adyo . . D 131d puna caiva ] ABpc D; puna( - va?) Bac ; pranayaiva C s praj yate ] ABpc D; praja( a - )ya( - ) Bac ; prajaptayat C 132a samyukt ] AD; sam(pu/pra - a?) Bpc ; sampreksy Bac C a . . . . a 132b pac sa ] corr.; pac s AD; pa c (s )err. B; pa c s C a a a s a a s a a samyut ] conj.; s mpratam a a . . ABCD 132d bhedair ] AB; bhedai C; bhe(dy)air D 133a devyo tha ] ABpc ; devy tha Bac CD a s dutya ] AB; dutya C; d( - ty)a D s s 133b yoginyocchusma ] ABpc ; yoginy cchusma Bac C; yo( a . . pc - i - ?)cchusma D m tarah ] A ; m tar h Aac D; m tar Bpc ; m tar(ah?) Bac ; (sobhar ?) a a a. a a a a . . . C 133c samast n ] AD; samast +n+ B; samast C a a a 133d sivecch m anuvartin ] A; sive( - ama a -?)vartin Bpc ; sivak mena varttin Bac C; sivecch sa( - ?)varttin D a a


Colophon: bhairave tantre ] A; bhairavatamtre BCD picumate ] ABpc D; pi( - ?)mate Bac ; bi(ndu). mata C vidh ne ] AD; vidh na BC a a sambandha ] A; sambaddhah BC; savatta D pa. talah prathamah ] ACD; patalah 1 B . . . . .


ath tah sampravaksy mi aghoryarcanam uttamam | a . . . a nir c ro yad mantr avadhutatanuh sthitah a a a . . a. tad tu kurute puj m yogen m sivasya ca | a s a . a avadhut tu s saktir nir c rapadah sivah a a a . .

etesam tu vidhim j tv tato mantr prasidhyati | . a a . . atas tesam pravaksy mi mantroddh ram anukram t a a . . . a sukl mbaradharo mantr suklagandh nulepanah | a a . bhuprade e sucau divye divyapuspair alamkrte s . . .

s tatra devya ca dutya ca yoginyo m taras tath | s a a uddhareta sad pr jah kap lsapurahsar h a a . a . a. 5

adimam tu dvityasya prathamam tu vyavasthitam | . . esa dev smrt rakt bindumastakayojit a a .a .

Codices: ABD A: f. 4v

1b aghoryarcanam ] corr.; aghory rcanam AD; agh( - ?)rccanam Bpc ; a(ghosy mcca?)na Bac a a. uttamam ] ABpc D; suttamah Bac 1d tanuh ] Bpc ; tanu ] ABac D 2a tu ] AD; tucorr.? . . a B 2c avadhut ] ABpc D; avadhuta Bac s ] AD; +s + B a a saktir ] B; sakti AD 2d . padah ] conj.; parah ABD 3c tesam ] AD; tayoh B 3d anukram t ] ABpc D; a( - ?)kram t Bac a a . . . . 4a dharo ] AD; dha(ro)corr.? B 4b nulepanah ] ABpc D; nulepanam Bac 4c bhu ] ABpc D; ( . . ac pc ac s - ?) B 4c divye ] conj.; divyair ABD 5a devya ] B ; devy s AB D s a dutya ca ] ADpc ; s (tusram ca)err. +dutya ca+ B; sudra( - ?) Dac 5c sad pr jah ] A; mah pr ja+h+ B; sad pr ja a a . a a a a . . D 5d kap lsa ] AB; kap lmsa D a a. purahsar h ] Bpc ; purahsar m A; (parapanam?) Bac ; . a. . a . . . purahsar h D 6a adimam tu ] ABpc ; adisam tu Bac ; adimantr(a?) D 6b vyavasthitam ] ABpc ; . a. . . ac pc ac corr. vyavasthitah B D ; vyavasthit D a 6c esa ] AD; (e) sa B rakt ] AD; ra(ksa)err. +kt + B a a . . . . corr. 6d mastaka ] AD; ma(sta) ka B


314 . trtyasya trtyam tu svaraik da abhusitam | a s . . . esa dev smrt ghor kar lti ca vi rut a a s a .a . 7

adimam tu dvityena urdhvan dena yojitam | a . ekavimsat par yonir dvityasvarayojit a a . :::::: 8

astavimsa tr*tyena vimsamam tu tath punah| a . . . .. . . dv trimsaikunatrimsena urdhvan dena yojitam a . a .


hamsa esa dvityena pranav disamanvitam | . . a . etad guhyam may proktam mantrabhedavyavasthitam a . . ath tah sampravaksy mi yogen m tu laksanam | a . s a . . . a . . pranav dinamask ram vidy m samyojya yatnatah a . a. . a . . esa te pratham prokt krostuk ca mahoday | a a a . .. sv h k r ntasamyuktam dvity yogin smrt 12 a a aa a . . .a . a . . a humk rena trty tu vausatk re caturthik | a . . a pacam vasatk rena phatk re sasthik bhavet . . a . . a . .. a
A: ff. 4v*5r . . . 7b bhusitam ] ABpc ; bhusitah Bac D 7c esa ] ABpc D; eso Bac ghor ] AD; (gho?)r B a a 7d . . kar lti ] AB; karti D a 8a adimam tu ] ABpc D; adimamtra Bac 8b n dena ] AB; n de(na?) a a . . D yojitam ] ABpc ; yojita(h?) Bac ; yojita D 8c vimsat par ] corr.; vinsat (p)ar A; vim( asva)r a a a . . . s B; vimsat pan D a yonir ] corr.; yoni AB; yo(ni?) D 8d dvitya ] A; dvityam B; dvitye . . D svara ] AD; (pu)err. +sva+ra B 9a astavimsa ] Bpc astavinsa A; astavim( -?) Bac ; as.. . .. .. . . tavinsa D 9b vimsamam tu ] corr.; vimsaman tu A; vim( - - - ?) B; vimsamantra D 9c . . . . . . dv trimsaikunatrimsena ] conj.; dv trimsakonatrmsena AD; dv trimsakonatrimsena B a . a . a . 9d yo. .. . jitam ] ABpc ; yojitah Bac D 10a hamsa ] ABpc D; hamse Bac 10b pranav di ] AD; praa . . . . na(v di)corr. B a samanvitam ] ABpc ; samanvitah Bac D 10c etad guhyam ] AD; etan tu hrn B . . . . proktam ] em.; prokt m AD; prokt B a. a 10d mantra ] AB; se( ?)tu D vyavasthitam ] ABpc ; . vyavasthitah Bac D 11a ath tah ] ABpc ; ath tam Bac ; ath ta D a . a . a sampravaksy mi ] AB; . . . a samprava(ksy - - ?) D 11b laksanam ] ABpc ; laksanah Bac ; laksana(m?) D 11c namask ram ] AB; a . . . . . . . . . . . na samsk ram D a . 11d vidy m ] B; vidy AD a. a samyojya yatnatah ] A; samyojay+e+t tatah B; . . . . . samyoksayen natah D 12a esa ] em.; esa ABD 12b krostuk ] A; kro(stu)corr. k B; krostrek D . . . . . .. .. .. 12c k r nta ] ABpc ; k r tra Bac ; k r tu D aa aa aa samyuktam ] D; sa(m)yukta(m?) A; samyu - . . . . . pc ac . . ?) B ; (yampraksa?) B 13a hum ] A; hu(m?) B; hum D trty ] ABac ; trty m Bac ; trtyo . . . a . a. . . D 13b vausatk re ] ABpc ; vausatk rah Bac ; vau(sa - ?)k re D a cathurthik ] Bpc ; cathurthak a a . . a . . a . . AD; catu(rthak ?) Bac a 13c vasat ] ABpc D; vasat Bac 13d phatk re ] ABpc ; humk ra Bac ; . a . a . . . hatk re D . a




315 sad yoginyah sam khy t aghory ngavinihsrt h | a aa a . . . a. . . ath to m tar h vaksye tantre ucchusmasambhave a a a. . . pranave tu sthito devo binduke tu mahe var | s . ek re tu sthit brahm ak re caiva vaisnav a a a ..



ak re caiva kaum r k re ca vivasvat | a a a ik re v sav dev sv k re caiva candik a a a a .. a



h k rena par saktir et yasy vinirgat h | a a . a a a a. m taras te may prokt y ge ucchusmapujite a a a a . . y s eva may prokt m t na caiva puran | a a a a a r. . tasyedam kathitam sarvam yam j tv n vasdati . . . . a a a 18 17

iti mah bhairave mulatantre dv da as hasrake picumate a a s a nav ksaravidh ne mantroddh rapatalah dvityah a . a a . . . 2

A: f. 5r 14a sad yoginyah ] AD; sa( - ?)oginyah Bpc ; sa( - - ?)sinya Bac 14b aghoryanga ] corr.; aghory a . . . . . . A; aghory ( - ?) Bpc ; agh ( - ?)rddha Bac ; aghory (t ca?) D nga a a a. vinihsrt h ] corr.; vini rt h ABac ; s. a . . . a. vinihsrt h Bac ; vini rt D s. a 14c ath to ] AD; ath (t ?)corr. Bpc ; ath (tt ?) Bac a a a a a m tar h ] em.; m taa a. a . . a . pc r m AD; m taram B ; so(tta?)r m Bac a. a a. 14d tantre ] AD; tantra B ucchusma ] A; ucch(usma)corr. . . . B; ucchumma D 15a sthito ] em.; sthit ABD a devo ] em.; deva ABD 15b binduke ] ABpc D; a bindukan Bac 15c ek re ] ABpc D; ek r(an?) Bac a a brahm ] AD; br hm B a 15d ak re ] AD; pc a onk re B ; ( nk ra?) Bac a a 16a ak re ] AB; ak ( - ?)re D a a 16b k re ] A; ik r+e+ B; ik re D a a a 16c ik re ] AD; k re B a a 16d sv k re ] AB; svok re D a a a 17a saktir ] em.; sakti ABD 17b et yasy ] conj.; etayasya AD; (e?)corr. te yasya B a a 17c m taras te ] AD; m ta(ras te)corr. B a a 17d a y ge ] A; yoga B; yoge D a pujite ] ABpc D; pujit Bac 18a y s ] AB; yo n D a a a may ] AB; a err. pc ma(yo) D prokt ] AB D; pre Bac a 18b m t na ] ABpc ; m (t - - ?) Bac ; ( - ?)trnam D a r. a s 18c . .. . sarvam ] BD; sarvv m A a. 18d yam ] A; (ma+j+?) B; ya D n vasdati ] Bpc ; n vadati A; nava( a a s . . ?)dati Bac ; n va(m?)dati D a Colophon: mah bhairave ] AD; mah bhairava B a a vidh ne ] AD; a . s vidh nam B a .


devy uv ca a cchommak h kdrsa deva kul n m s dhakasya ca | a . . a a. a praj yate yath bhr t bhagin v vi esatah a a aa a s . . cary yuktasya deve a yath j syanti yoginh | a s a a . paraspara ca vr nam ekatantrasam sray m | a. a a a a al p rthe mah deva kathayasva prabh satah a a. . bhairava uv ca a srnu devi pravaksy mi cchomak n m tu laksanam | . . a a. . a . . yena vij yate bhr t bhagin v mahe vari a aa a s 101 100 99

j tv ca yoginm mantr sivecch codit tmav n | a a . a a a s dhakas tu tato dady d v cikam mudralaksanam a a a . . . 102

potangety abhiv danam pratipotange pratyabhiv danam | a a . yoginn m tu vr nam n rsety abhiv danam | a. a . . a . a
Codices: ABC A: f. 234r

99a cchommak h ] em.; cchommak ABC a. a kdrsa ] em.; kdrso ABC . . 100a deve a ] B; des vesa AC 100b yoginh ] em.; yogin AB; yogini C . 100c vr nam ] corr.; vr nam AC; vr a. a. . a a a namm B 100d sam sray m ] conj.; sam srayam ABC a a a 100e al p rthe ] ABpc C 100f . . . kathayasva ] corr.; kathaya va ABC s prabh satah ] AB; prabh sata C a. a. 101d mahe vari ] em.; mas . he var ABC s 102a yoginm ] em.; yogin ABC . mantr ] AB; mantri C 102b codit tmaa v n ] B; codit tmav m AC a a a. 102c dady d ] B; dady AC a a 103a potangety ] AC; potange(tt?)y B 103b pratipotange ] AB; pratipotanga C pratyabhiv danam ] B; praty bhiv danam ABC a a a . 103d n rsety ] AB; n risebhy C a . a .


317 pratin rsa abdena procyate prativ danam a . s a 103

ek ngulidar an t sv gatam dv bhy m susv gatam | a s a a a a. a . kosthapravisten ngusthena ksemamudr vidhyate a .. .. a .. . siram dar ayate y tu v rt m s tu samhate | s a a a. a .


t m di am vksya d tavy mudr de agam tu y | a. s . . a a a s a a


anguly samsprset p dam kathit tu na samsayah a a . a . . . . s lal tam dar ayed y tu kutra y syasi -m- adi et | a. . s a a


vksya suryam sprsed vaktram yathestam krtitam bhavet . . . . . .. . sikh m dar ayate y tu kutra suto si suvrate | a. s a anusmrtya bhavam lingam yonim sprstv siv tmakam . . . . ... a a vaktram dar ayate y tu gotram te kdrsam giri | s a . . . . smrtv devm sprsed b hum v mam v mena p nin | . a a. a . a . . a . a v m c ras tu me gotram saktayo vardham naj h a a a a a. . 108



da anam dar ayed y tu kim purvam te niketanam | s s a . . . anusmrtya sivam so hi sprsate -m- udaram priye | . . . . ::::::::::::: m yodar d idam pr ptam dvityam tu siv srayam a a a . a . . karnam dar ayate y tu kim srutam tu sam di et | s a a s . . . . nabham sam*vksya hasta ca pacasrotasam gamam | a . . .
A: f. 234r*234v B: 104ab missing; 104cd in lower margin, possibly by original scribe. C: skips from 103c (pratin ( a?). . . ) to 105c (. . . smrse p dam) as a . C: skips 107b109a. At 112b, it skips back to . 107b and resumes, copying 109b112a a second time (C1 = 1st reading, C2 = 2nd reading). 104c kostha ] B; kosta A 105b v rt m s tu samhate ] conj.; v rtt sa tu mahyate AB a a. a a a 105d .. .. de agam tu y ] B; de agam nuy A s a a s a a 105e samsprset ] B; sa sprset A; smrse C 105f . . . . s kathit ] em.; kathit s ABC a a 106b y syasi ] AB; y sya i C a a s adi et ] corr.; adiset ABC 106c suryam ] em.; surya ABC sprsed ] B; sprsed AC 106c vaktram ] AB; vaktum C 106d . . . . . yathestam ] AC; yathestham B 107a sikh m ] em.; sikh ABC a. a 107b suto si ] conj. (Isaac.. . .. . son); sutro si ABC 107d yonim ] AB; y ni C a 108b giri ] C; giri+h+ A; girih B 108c . . . devm ] em.; devi ABC . sprsed ] corr.; sprsed ABC 108d v mam v mena ] AB; v sam v sena C a a . a . . . a 108e me ] AB; sa C 109a dar ayed ] B; dar aye A; dar a C s s s 109d sprsate ] B; sprsate AC1 C2 . . -m- udaram ] conj.; sodaram AB; sodanam C1 ; saudaram C2 109e m yodar d ] em.; m y dar d a a a a a . . . . ABC idam ] AB; ida C1 ; itam C2 110a karnam ] B; karna AC 110b sam di et ] corr.; sama s . . . . . adiset ABC


318 visrt ngulikam krtv dar ayec chatap nin s a. a . a . . a jihv m dar ayate y tu rasitum s samhate | a. s a . a mrtujayam smaritv tu dar ayta kamandalum a s . . .. grv m dar ayate y tu supriyo si mah tmana | a. s a a . atmane tu par m murtim smrtv tv atyanta me priy a. a . a skandham dar ayate y tu svasth nam kutra c di et | s a a . a s . svagotrasy srayam j tv sa di am avalokayet a s . a a





b hum samsprsate y tu bhr t si mama suvrate | a a aa . . . v mahastasya samspar ad bhaginti sam di et a s a s . 114

v m ngulim yad vaktre praksipyan tu pradar ayet | a a a s . . tad s bhojanam svestam pr rthayed vrapungavam a a a . .. . a ten pi pacabhut tmam yuktam sarv rthasamyutam | a a . . . n n vidham rasam vaktre smartavyam tu nav tmakam | a a a . . . a a bhuktv trpt tu s bhutv v mena parivartate a . a a 116 115

hrdayam sprsate y tu hann meta mah vratam | a a a . . . tryoda anga ivam smrtv atmalingam anusprset s s . . a . stanam nirksate v mam sprsate v yad priye | . a a a . . .
A: f. 234v 110e visrt ngulikam ] corr.; vi rt mgulikam ABC2 ; vi rt mgulike C1 s. a . s. a . 111b samhate ] ABC2 pc ; . a . . samhato C1 111c smaritv ] ABC2 ; smatitv C1 a a 111d kamandalum ] AB; khamandalu .. .. C1 ; khamandalum C2 112a grv m ] AB; grv C1 C2 a. a 112b supriyo si ] AB; kim sruta .. . . kutra sutro si C1 ; supriy si C2 a mah tmana ] em.; mah tmanah ABC2 ; suvrate C1 a a 112c . . murtim ] AB; murtti C 112d tv atyanta me ] conj.; tudyanta me ABC priy ] C; priyo AB a 113a skandham ] em.; skandha AB; skadha C 113b sth nam ] em.; sth na (unmetrical) ABC a . a . s kutra ] AB; ku(ru?) C adi et ] corr.; adiset ABC 113c gotrasy srayam ] corr.; gotrasy srayam a a . . AB; g trasy srayam C a a 113d sa ] conj.; s ABC a di am ] corr.; dis m AB; dis C s a a 113d . avalokayet ] AB; savaloktayet C 114a samsprsate ] B; samsprsate A; sam(smr)sate C 114c sam. . . . . . . spar ad ] B; samspars d AC s a 114d sam di et ] corr.; sam diset ABC a s a 115a v m ngulim ] em.; v a a a . . m nguli ABC a 115c bhojanam ] em.; bhojana AB; (h?)jana C 115d pr rthayed ] AB; pr sa a. . thamyed C 116c vidham rasam ] AB; vidharasam C 117a hrdayam ] AB; hrdamya C . . . . . . . . . 117b hann meta ] AB; ha(tt/nn) mete C a a 117d lingam ] AB; linga C anusprset ] corr.; anu. . sprset AB; tu sprset C 118a nirksate ] B; nirksyate A; nirojyate C . . 118b sprsate ] corr.; sprsate . . . . AB; sprte C v ] AB; v C a .


319 pras rya s dhako vaktram putro ham te prabh sitam a a a. . . jatharam sprsate y tu raksitavyo si suvrate | a . . . . s rasena pranamam tu kartavyam tv atmaraksanam a . . . . .



s n bhim samsprsate y tu mah mel pam adi et | a a a a . . . madhyade e mah vra kulasapt da asya pi s a a s 120

katim samsprsam n tu y m di am c valokayet | a. s . a . . . . a a ::::::::::: tatkulasya sam khy ti melakam tu na samsayah a a . . . guhyam samsprsate y tu s tu putra krt tmav n | a a a . . . .a manas cintya svam y gam tvatpras d t krtam bhavet a a a . . . a . . urum samsprsate y tu ksnaham s sam di et | a a s . . . . . a anusmrtya tu manth nam tasy dehe niyojayet | a . a . tanniyog t suvi r nt manth nsavimardan t a sa a a a j nu samsprsate y tu kriy ksuna tu s bhavet | a a a . . a . .
:::::::::: :::




krocabjam tatocc rya aksasutram tu samsprset | . a . . . . laksaj p d vimucyeta kriy ksunam tu yojayet a . . . . a a :::::::::: ::::: 124

a. jangh m ca sprsate y tu s priy n tu niyacchati | a a a . vimocayet tato mustim v mahastasya mocane .. . a

A: f. 234v 118d putro ] AB; putrom C prabh sitam ] AB; prabh site C a. a. 119a jatharam ] AB; jathara . . . . . C 119b raksitavyo ] AB; raksitapo C 119d raksanam ] conj.; laksanam AB; taksayam C . . . . . . . . . 120a n bhim ] em.; n mbhim B; n bhi AC a a a 120b mah mel pam ] BCApc ; mah mel pakam Aac a a a a . . s adi et ] BC; adiset A 120c de e ] AB; de a C s s 121a samsprsam n ] em.; samsprsam nas a . . a a . . ABC 121b y m di am c valokayet ] conj.; y n di is(v/c) avalokayet A; y n di a c valokayet B; a. s . a a s a s s a y n di isv avalokayat C a s 122a guhyam ] AB; guhya C samsprsate ] corr.; samsprsate ABC . . . . . . . 122b krt tmav n ] B; krt tmav m AC a a. 123a urum ] em.; uru AC; (u)rum B samsprsate ] B; .a .a . . samsprsate AC 123b sam di et ] B; sam diset AC a s a 123d tasy ] em.; tasya ABC a dehe . . niyojayet ] AB; dehaniyojayat C 123e suvi r nt ] B; suvisr nt A; suvibhr nt C sa a a a a a 123f manth nsavimardan t ] AB; manth ni amardan t C a a a s a 124a j nu ] AC; j num B a a samsprsate ] B; . . . samsprsate A; samprte C 124b kriy ksuna ] conj.; kriy ksunan ] AB; kriy (ks - ?)nan C a . . a . . a . s a . . . . . bhavet ] A; so bhavet B; sobhayet C 124c krocabjam ] AB; kr cav(ryya) C . a 124d samsprset ] B; . . samsprset AC 124e laksaj p d ] AB; laksan t (py/pp)ad C a 124f kriy ksunam ] conj.; kriy ksua . . . a . . . . a a . a. a na AB; kriy ksun n C a . a 125a jangh m ] corr.; jamgh AC; jangh B a ca sprsate ] em.; tu sprsate . . . . B; ca sprsated AC .


320 p dam samsprsate y tu padabhramsam tu s di et | a . a a s . . . . na sth tavyam tad tena tasmin sth ne vipa cit | a a a s a . pranavam tu samucc rya gantavyam n nya me gatih a . . . a . nakham p dasya y devi sprstv y ti par nmukh | a a . a ... a a khecaratv cirenaiva* kathate s dhakasya tu | a a . tatah prabhrti so py evam nityam vai samyato bhavet . . . . . atmap datalam y tu samutksipya pradar ayet | a s . a . p t lasiddhir vrasya kathate s cirena tu aa a . 128 127 126

a ak se mocayen mustim dhunate ca svakam tanum | . .. . tad tu svargav sn m melak myas tu samyutam a a a. a . .


urdhvam samvksya y pa c d di alokanam acaret | a s a s . . . caturnam melakam s tu kathayec c rdhar tratah a a . . . a . 130

nitambasthau tu y hastau krtv prahasate muhuh | a . a . melakam satkasamghasya dvyardhay me ti s kathet a a . . . . ::::::::::::::::::::::::::::: n s gre tu yad hastau krtv c layate siram | a a a . a a . navakasya tath khy ti melakam tu mah vane a a a . 132 131

a adhomukh tu y bhutv bhumilekhanam arabhet | a p t lac rinn m tu melakam m trmandire a a a . a . . a.

A: ff. 234v*235r 126a samsprsate ] corr.; samsprsate AB; samsprsave C 126b bhramsam ] B; bhramsan A; . . . . . . . . . bhramsam C tu s di et ] em.; tu m diset AC; sam di et B a s a a s 126c tad ] conj.; yad ABC a a 126d . . tasmin ] corr.; tasmim ABC vipa cit ] em.; vipa citah B; vipascitah AC s a s 126e samucc rya ] AB; a . . . samuc rya C a 126f gantavyam ] AB; samntavyam C me ] AB; se C 126f gatih ] em.; gati . . . . ABC 127b par nmukh ] corr.; par nmukh AB; par nmukhi C a a a 128a talam y tu ] AB; ta( . a -)C 128b samutksipya ] B; samuksipya A; samuksi C pradar ayet ] AB; pradar aya(mt) C s s . . . . a a a 128c siddhir ] em.; siddhi ABC 129a ak se ] corr.; ak se AB; ak sa C mocayen ] A; mocayet BC 129c svargav sn m ] conj.; sarvav sn m ABC a a. a a . 129d melak myas ] AC; melak s(y/p)an B a a 130a urdhvam ] B; urddham A; urddha C samvksya ] AB; samvjya C . 130b pa c d ] B; pa c s a s a . . . . . A; ( - ) c C s a 130c caturnam ] B; caturna AC 130d kathayec ] corr.; kathaye ABC 131a . . nitambasthau ] AB; nitambe sthau C 131b muhuh ] AB; muhu C 131d dvyardhay me ti s a a . kathet ] conj.; dya(dva/ddha)y medimekathe A; dyadvay medi(m/s)ekathe B; yaddhay sedimekathe a a a C 132d melakam tu mah vane ] AB; malakam tu mah bale C a a 133a y ] AC; yo B a 133b . . bhumi ] em.; bhuti ABC


321 svajihv lokanam y tu krtv pa c t prakampate | a . a . a s a jal ntarv sinn m tu melakam kathate tu s a a a. a . 134

a a a p d n murdhaparyantam krtv hastaprakampanam | . . a


s y s siv ditattvasth tatstham melakam adi et a a a a .



so pi mudr patih pujya tath manth nabhairavam | a a a . bhakty paryatanam kury d yath tantraprabh sitam a a a a. . . namo stu digbhyo devebhyah purvasiddhavin yak n | a a . dattv rgham paray bhakty tato mel pakam bhavet| a a a a . . tats m nyam mah devi sarvakaly nasampadam a a a a. . . . 137 136

s dhakasya pravaksy mi maunasthasya yad bhavet | a a . a aksunam tu mah bh ge srnusvek gram nas a a . . . a a a . . . 138

tarjany ngusthak gre tu puspamudr prakrtit | a .. a a a . mulaparvabhram ngusthe pr rthitam tu vilepanam a .. a . utt nahaste sollole dhupamudr su obhane | a a s


s adhomukhaprac lena anguln argham adi et a


a a kumbhamustir jalam vindy d dhup ng rordhvag ngulh | a a . . . .:: :: drgbhram n m rjanam viddhi jihv lolopalepanam 141 a a a . . ::::::::::::::::::::::
A: f. 235r a a 134b prakampate ] AB; prakalpayet C 134c jal ntar ] corr.; jal nta ABC a a 135a a p d n ] em.; a a p d AB; a p da C a a 135b prakampanam ] em.; prakampane AB; prakalpate C 135c s y ] AB; c C a a 135d adi et ] B; adiset AC tatstham ] em.; tastham AB; tam C 136a . . . patih pujya ] em.; patipujyam (unmetrical) AC; patih pujyam B 136c bhakty ] AB; bhaksy C a . . . . . a 137a namo stu ] AB; nam stu C a digbhyo ] AB; digbh C a 137a devebhyah ] Bpc ; deveb. hya Bac AC 137b siddha ] AC; siddhi B vin yak n ] corr.; vin yak m ABC a a a a. 137c argham ] AB; arghas C 137d tato ] AB; teto C 137e s m nyam ] em.; s m nyo AB; a a a a . . s m ny C a a a 137f kaly na ] B; kaly na AC a. a sampadam ] AB; sampadah C 138b . . . . .. yad bhavet ] AB; pad (- -)vet C a a 138c aksunam ] AB; aksanan C 139a angusthak ] B; a . . . . . .. a angustak AC 140a haste ] conj.; hasta ABC sollole ] conj.; sollola ABC 140b su obs hane ] corr.; susobhane A; su obhabe B; su obhan C s s a 140c prac lena ] AB; prac rena C a a . 140d s anguln argham ] em.; anguln rgham ABC a adi et ] B; adiset AC 141a mustir ] em.; musti .. .. ABC vindy d ] B; vimdy A; vidy C a a 141b ordhvag ngulh ] em.; orddhag mgul AB; a . a. . a ardhagoguli C 141c bhram n ] A; bhram t B; bhram C a a a m rjanam ] AB; m rjjana C a a . 141d jihv lolopalepanam ] conj.; jihv llolopalepane ABC a a

322 hastam trendhan n m tu j nukurparasamgam t | a a a. a a .


naivedya ca vij ny d utt n dhomukham karam a a a a . kanyas prasrt mustir adhovaktr tu sastrik | a a a .a ..


s mustim badhv khag lok t khadgapr rthanam adi et a a a a . .. . : tiryakpras dade iny darbhamudr prakrtit | a s a a a mrg til m yav n n tu sukar prasrtih punah a a . a. . :: . . 144


dh ny n m bilvapadm n m sarv ngulipras ranat | a a a. a a. a a . bilvamadhyasrt ngulya dvitye prasrtam talam . a . . 145

pac ngula*sam gr s tu urdhvavaktr phal tmakam | a a a a a musty rdh tmasamudr tu rjus tiryakprah rane a a a . . .. a :::: samhat ngusthayogena matsyak n tu vinirdi et | a .. a s . proktam lokavisargam tu svan s granirksanat a a . . . . n s ghr yanarupena krsnady madir sav h | a a a . . a a. .. . a v makarnam sprsan devi gudajyamadhu krtitam 148 a . . . .



sirasamspar an t proktam phalajam tu mah savam | s a a . . . ghrt dimadhu dugdh n m sruvayogo yath vratam a a. a .a 149

mustim murdhni viniksep d bhasmamudr var nane | a a .. . . a

A: ff. 235r*235v 142a m trendhan n m ] B; m tre(ndh?) n n A; m trandhan n n C a a a. a a a a a a 142c vij ny d ] B; vij ny a a a a AC 143a kanyas prasrt ] conj.; anyas prakrt ABC a a a a mustir ] em.; musti ABC 143b sas . . .. .. a a trik ] B; sastrik A; mamtrik C a a a 143c alok t ] em.; alok ABC 143d khadga ] AB; kha C . . s adi et ] corr.; adiset ABC 144a de iny ] B; desiny AC s a a 144b prakrtit ] em.; prakrtit m a a. AC; prakrttit h B a. 144c mrg til m ] AB; mrgil m C 144d prasrtih ] conj.; prasat ABC . a. . a. . . punah ] AB; puna B 145b anguli ] AB; angali C 145c srt ] corr.; srt ABC . a . .a 145d dvitye ] AB; dvitiy C a prasrtam ] B; pra rtam AC s. . 146a angulasam gr s ] em.; aa a . . ngulam sam gr (n/t) A; angulam sam gr n BC a a a a 146b vaktr h ] corr.; vaktr ABC a. a 146d . . rjus ] conj.; rju B; riju AC prah rane ] em.; prah ranet A; prah ranat B; prah renet C a . a . a . a . 147a . . samhat ngustha ] B; samghat ngusta AC a .. a .. 147b vinirdi et ] C; vinirddiset AB s 147d nirksa . . . nat ] AB; (ni?)riksanat C 148b madir ] AB; sadir C a a asav h ] em.; asav t ABC a. a 148c . . . karnam sprsan ] em.; karnasprsamn B; karnasprsam AC . 149a sira ] AC; sirah B . 149b . . . . . . . . phalajam ] AB; phalan tu C asavam ] AB; asavam C 149d sruvayogo ] AB; bhruvay g C a a . . 150a mustim ] AB; musti C murdhni ] AB; mu( - - )ni C 150a viniksep d ] AB; viniksep C a .. . .. . . a

323 s prsthasamspar an t proktam yogapattakam adi et s a ... . . .. sphicasamspar an d devi asana ca vinirdi et | s a s . an m madhyam ngusthac lan c c ksasutrakam | a a a .. a a a . kaksasamspar anenaiva smrto bok nako naghe s a . . . murdhn dikatiparyante haste mudr su pacakam | a a . bhrtam kamandalum devi damarum rocan pi ca a . . .. . . . adhomukhaprakampena v mena tu karena tu | a . ghantamudr vinirdista vna vnakrtim karam a .. . .. . . . . et nyam api vr nam sanketam sastracoditam | a a. . . svecchay v prabodhy dau pa c n maunam sam caret a a a s a a . lokasangavirakt tm maun dhy napar yanah| a a a a . . ek ntaratalas tu sidhyate vigat mayah a s a . 155 154 153 152 151 150

anyonyasammatam j tv v ky l pam tathaiva ca | . . a a a aa . cchommak n bh samudr bhir yojayta vicaksanah a a. a . . . 156 55


iti cchomm dhik ras pacapac satimah patalah a a a . . .

A: f. 235v s 150c prstha ] C; prsta AB 150d adi et ] corr.; adiset ABC 151a samspar an d ] B; s a ... ... . .. .. samspar an AC s a 151b vinirdi et ] corr.; vinirddiset ABC s 151c ngustha ] B; ngusta . AC 151d c lan t ] em.; c lan AB; b lan C a a a a a a c ksa ] AB; v ksa C a . a . 151f bok nako a naghe ] AB; v k makoraye C a a 152a murdhn di ] corr.; mudhn di ABC a a 152c bhrtam kama. . ndalum ] conj.; bhrtakamandale ABC 152d damarum ] AB; (daman)um C 153b v mena ] AB; a .. . . .. . . . . v mena C a 153c vinirdista ] em.; vinirdisto ABC 153d vnakrtim karam ] C; vnakrtikaram . . . . . . . .. .. (unmetrical) A; vnakrtih karam B . . . 154a et nyam ] AB; et n(me)m C a a 154c svecchay ] AB; a . svecchayo C 154d pa c n maunam ] AB; pac t mauna C s a a 155a lokasanga ] AB; lokam anga . C 155b maun ] AB; mauli C par yanah ] AB; par yanaih C a . . a . . 155d vigat mayah ] AB; a . vigat maya C a 156a sammatam AB; sam(p/y)atam C 156c cchommak n ] corr.; cchoma . . . . mak m AB; cchosmak m C a. a. 156c mudr bhir ] B; mudr bhi AC a a 156d vicaksanah ] AB; . . . vicaksana C 157a iti ] C; absent in AB Colophon: cchomm dhik ras ] AB; chosm dhik s C a a a a . . pacapac satimah ] em.; trpac satimah ABC a a . . .


bhairava uv ca a ath tah sampravaksy mi chomak n m yath vidhih | a . a a. a . . . a


rupalaksanakarma ca kul c ravicestitam a a . . .. yath vij yate vro yogin v kulodbhav | a a a a

siddh siddhavibh g tu sama cottaras dhakaih | a a a a .

::::::::: ::::::::::::::::::::::::::

y gacary vi esasthais tan me nigadatah srnu a a s . . . .


s . . . n pum n str nik rena bhu sma anam kr dakin | a a . a . a a bhru l m raudrik ghru ca khih sy n m trkulodbhav a a. a . a ak ratritayenaiva siv n m kulaj smrt | a a a. a .a deti damarik prokt hi ca davy var nane a a s a . . a

s s mrtam bru ca v m khyam heti m msam var nane | a . . a a a. a . . yo bh ry bhagin yena mak rotpattir ucyate a a a 5

lik rena smrtam bhaksam vak r t p nakam priye | a . aa a . . . . . him app bh janam kena gena v cyam tu bhojanam a a a . .

Codices: ABD

A: f. 278r

1b chomak n m ] AB; cchomen m D a a. a. yath vidhih ] AD; yath vidhi ] B a a 2a vro ] AB; . (de)err. v(r?)o D 2c vibh g ] conj.; vibh gas ABD a a a 2e vi esasthais ] B; vi esasthai AD s . s . 2f tan me ] AB; tan ma D 3a n pum n ] AB; n ( - ?)um m D a a a a. 3b sma anam ] B; smas nam s . a . AD 3b kr ] AD; (hr?) B 3c bhru ] AB; (gr?)u D 3c raudrik ghru ] em.; raudrik ghru a a . . AB; raudrik( skr ?) D a. a 3d khih sy n ] corr.; khi+h+ sy n AB; khih sy t D a 4a tritayenaiva ] B; . a . . a trtayenaiva AD 4b kulaj smrt ] em.; kulajah smrtam A; kulaj smrtam B; kulajah nmrtam D a a . .a . . . . . . . . 4c deti ] conj.; devi ABD 4d ca davy ] em.; candavy ABD 6c him app ] conj.; hisapp a a . a . . a ABD

325 phak rena vij ny t surat rthe prabh sanam | a . a a a a. . raktam vas tath sukram sesavarnaih kramena tu a a . . . . . . ksak renatitam proktam jak rena vivarjitam | a . . . . a .::.::::: pak r d dhrdayam proktam v sanam sakhisamgame aa . . . a . .

nak rena tu naivedyam puspad nam tasamjay | a . a . a . . .


y canam tu dak rena thak rena pratisthanam a a . a . . ..

kap lam tu tam ity uktam thah spharo vai var nane | a . a . . . . sastriko daksarenaiva pheti p t divandanam aa . . . 10

mudr lankaranam nena cena proktam tu cumbanam | a . . . viruddhakaranam sena jhak rena tu karsanam a . . . . . s pacanan tu dak rena gheti gh tanam adi et | a . a . vyomavr sisamj n m yath k lam ksak rakam a a a . . a . a a. kulasoda akasyokt h svar dy h kramakalpit h | a. a a. a. . . s vre abhairavnam tu amsamja v sapurvakam s . . s a . .




eva caik *ksar h samj h prokt y h samkhyay yut h | a . a. a a. a a. . a. . sv msadev ksavarnadivibh gena suvistar h a. a . a a. . devy uv ca a j t varnatmik h samj h tath ny y h subh vibho | aa a. a a a. a . . a.
A: ff. 278r*278v 15 j t . . . uv ca ] inserted in lower margin in B aa a


7a phak rena ] AD; hak rena B a . a . vij ny t ] corr.; vij ny AD; vij ny m B a a a a a a. 7b surat rthe ] AB; a surat (r - ?) D a 7d varnaih ] AB; varnnah D 8a ksak renatitam ] em.; ksak renaditam ABD a . . . a .. . . .. . . . . 8c dhrdayam ] AB; drdayam D 8d sakhi ] conj.; sukhi ABD 9b tasamjay ] conj.; tu a . . . . . samjay ABD a 10b thah spharo ] em.; tham pharo ABD 10c sastriko ] corr.; sastriko ABD . . . . s 10d pheti ] AB; phati D 12b adi et ] B; adiset AD 12d ksak rakam ] BD; ksak raka(m?) A . . a . a 13a soda akasyokt h ] B; soda akasyokt A; soda akany kt D a. a a a 13b svar dy h ] B; svar dy a a. a a . . s . . s . . s A; svar dyo D a kalpit h ] B; kalpit AD a. a 13c vre a ] B; vresa AD s . bhairav nam ] corr.; bhairavn m A; bhairav n n BD a. a a 13d amsamja ] conj. (Isaacson); hamsamja . . . . . . A; hamsajah B; ham( - ?)ja D sv sa ] B; sv sa AD a a 14a samj h ] corr.; samj h . . . . a. . a. ABD 14b y h ] corr.; y ABD a. a samkhyay ] AD; samkhy y B a a a yut h ] em.; yut a. a . . ABD 14c sv msa ] corr.; sv msa ABD a. a. 14d suvistar h ] corr.; suvistar ABD a. a 15a varnatmik h ] corr.; varnatmik ABD a. a samj h ] em.; samj ABD 15b ny y h ] corr.; ny a a. a . . . a. . a y AD; j y B a a a subh ] corr.; subh ABD a a

326 vada tattvena deve a sarr vayav tmik h s a a a. bhairava uv ca a srnu devi mah bh ge samj y dehasambhav h | . . a a a. . a a . vrayogkul n m tu hit rthaprad yik h a a. a a a. 16 15

sirasamspar anenoktam vandanam prativandanam | s . . . sikh samspar anenaiva lal tena tu sv gatam a . s a. a 17

susv gatam ap ngasya spar an t sampratyate | a a s a . kasm d de ad ih y tah prcchito hi bhruv h sprsan a s a a . . a. . daksottarasab hyam tu spar ad de am tad tmakam | a s s . a . . ::::: n s gram pr gbhavam de am pa c dde am krk tik a a a s . s a s . . a. a . .



sprsan samdar ayen mudr h prativ rtt vidh yinah | s a. a a a . . . gantavyam svosthasamspar at krte taddi i vksanat s . s . . . . .. krtam bhavati su roni pratimudr vidh nakam | s . a a . . ksna svabhujasamspar ad urvoh spar at tu vi ramet s s s . . . . upavi ya j nusamspar at sphicenaiva karomy aham | s a s . na karomi tad y y sprsta jangh tu bh sitam a a a . .. a a.
:::::::: ::::: :: :::::::::::::::::::




A: f. 278v

tmik AB; ( - ?)ik D a a 16b sambhav h ] em.; sambhav ABD a. a 16d prad yik h ] em.; praa a. . . d yik ABD a a 17a sira ] AD; sirah B . ktam ] em.; kt ABD a 17b vandanam ] AB; . . candanam D 17b prativandanam ] AB; prati(c?)andanam D 17c sikh ] B; sikha AD a . . 18a susv gatam ap ngasya ] B; susv gatam rap ngasya AD a a a a 18b spar an t ] AB; spar an n D s a s a . 18c kasm d ] conj.; tasm AD; tasm d B a a a ih y tah ] corr.; ih y ta ABD a a . a a 18d bhruv h ] em.; bruv a. a ABD sprsan ] em.; sprsam ABD 19a ttara ] AD; ttarah B sab hyam ] B; sab hy n a a a . . . . . AD 19b de am ] em.; ve am ABD s . s . 19c n s gram ] em.; n s gr A; n s gr t B; n sa( - ?)r a a a a a a a a a a . D pr gbhavam ] Bpc D; progbhavam A; progbhavam Bac a de am ] corr.; desam AB; de(- s . . . . . ?) D 19d pa c dde am ] em.; pa c d vesam ABD s a s . s a 20a sprsan ] em.; sprsam AD; sprsam B . . . . . . . samdar ayen ] AB; sandar ayet D s s mudr h ] corr.; mudr ABD a. a 20c gantavyam ] AB; garttavyam . . . D svostha ] B; svosta AD samspar at ] B; samspar a AD s s 20d taddi i ] corr.; tadvisi AD; s . . .. .. taddisi B 21a su roni ] B; susroni A; susr ni D s . a. 21c ksna ] em.; ksno ABD spar ad ] B; s . . . . . spar a AD s 21d spar at ] corr.; urvvo spars A; urvo spar a B; urvv mparso D s a s a vi ramet ] B; s visramet A; vi( - r - ?)met D 22a upavi ya ] B; upavisya AD s samspar at ] B; samspar a AD s s . . a a 22c karomi ] conj.; karoti ABD 22d sprsta ] AB; sprsto D jangh tu ] conj.; jangh nu AB; ... ... janghonu D

327 karna askulik nguly srutam te m trmandalam | a a a. . s . . .. tatp r vaspar an t siddham sphutamel pak srayah as s a a a . . .
:::::: :::::::::::::::::::


na caiv mocayen mustim gaganastv hyasamgame | a a . .. .


hrdayam tu sprsed y tu v mahastena bh vit a a a a . . . bhagin s vinirdista s dhak n m phalaprad | a a a. a .. a


v mahast grasamspar an m trmadhye tu n yik a a s a. a a . garudasya tu samspar at dakinti vinirdi et | s . s . . tiryakcaksusvadrsty tu rudradakinilaksanam ... a . . . . sv msade akarenaiva damaritvam tu s di et | a. s a s . . . krk tik karenaiva daksinena var nane a . a. a . . . 27 26


kulam tu s dhak c ry n davy tm nam tu s kathet | a a a a. . a a . a . n s gradrstirodhena urdhva v sena s siv a a s a a a ... 28

samastadehabhangena svahastabhramanena ca | . mi r nam laksanam devi svarupakathane smrtam s a. . . . . . m msam kapolahastena jihv drsty tu matsyakam | a. a ... a . da an nguliyogena bhaksyam bhojyam tu cehitam s a . . . n s spar anayogena gandh prokt tu mudray | a a s a a a



A: f. 278v 30d31a: AD repeat 30d and 31a, reading bhaksam for bhaksyam the second time. After . . . . 30d, B inserts n s spar anayogena bhaksyam bhojyan tu cehitam in the upper margin, resuming with 31a a a s . . 23a saskulik ] corr.; saskulik AB; sa( - )erasure skulik D . a a a 23b srutam ] B; srutan AD 23c . . . spar an t ] B; spar an AD s a s a 23d sphutamel pak srayah ] conj.; sphuta mel pak msrayam ABD a a a a. . . . 24a caiv mocayen ] em.; cev maicayen A; cev mocayen B; cev m caye D a a a a a mustim ] AB; astim .. . .. . D 24b gaganas ] BD; gaga+na+s A 24c sprsed ] B; sprsed AD y ] AB; y n D a a 25a . . vinirddista ] AB; vinirddisto D 25c hast gra ] B; hasta(r )err. gra A; haste gra D a a 26b .. .. . dakinti ] em.; k kinti ABD a 26c tiryakcaksu ] B; tiryakcaksu A; tirya(- cuksu?) D 27a . . . s sv msa ] AB; sv sa D a. a 27b tu s di et ] em.; tu m adiset AD; tu adi et B a s 27c krk tik ] AB; . a. a krkotik D 28a c ry n ] em.; c ry ABD a a. a a 28b davy tm nam ] em.; dapy tm nan ABD . . a . a a . . a a tu s kathet ] AB; tu -m- akathet D a 28c drsti ] AB; ( - sti?) D 28d sv sena s ] B; a a ... .. sv sena s A; sv sana so D a a a 29a samasta ] AD; samastam B bhangena ] AB; bha( - - ) D . 29d svarupa ] AD; svarupam B 30a m msam ] B; m nsam AD a. a kapola ] AB; kap( ?)la D a . . . erasure 30b jihv ] AD; jihv m B a a. drsty ] AB; dr( - ) sty D 30b mastyakam ] em.; matsyak m a. ... a . .. a ABD 31b gandh ] corr.; ga(ndh/tv) A; gatv BD a a a

328 v masrkvinik jihv bhaved v m mrtam tu tam | a a a a a . . . *daksinasrkvinyog d vijeyam daksinamrtam a . . . . . . . vas hastatalaspar an majj kurparadar an t | a s a s a s jathare tu krte haste dakse putratvam adi et . . . 32 31

v me tu duhit prokt pit murdhni nirksayet | a a a a . m t tu kathit s tu v makuksipradar an t aa a a a s a . 33

samgrhtam smrtam n bhau prsthe ksiptam tu laksayet | . . . . . a ... . . . bh ry nitambahastena v mena pati daksine a a a . . 34

. a dutah p dasprsat siddho mitro v mabhujam sprset | a . . . s prakucakagrahenaiva kulatabh vam adi et . a 35

guptam kaksakarenoktam na guptam digalokan t | a . . . . . siddham v m ksisamkoc l lupt c ram tu daksinat a a a . . a a . . . . s dhik rapade v me purvo ksiptam tathottaram | a a a . .
::::::: :::::::::::::::::::::::::::::


daksine tu vij ny n melakam v make kare a a . a . .


evam yoge ivr nam sammatottaras dhakam | s a. . a . . mudr l pam sam khy tam yadanantaravistar t aa . a a . a


anyonyasammatam krtv v ky l p tha v priye | a . . . a a aa a svamudr l payog d v gopayed v ma asanam aa a a a s devy uv ca a

A: ff. 278v*279r 31c srkvinik ] A srkkinik B; srnkinik D a . a 31e srkvin ] corr.; srkvin A; srkkin B; . . . . a . . . srnkin D . 32a tala ] em.; tal ABD a 32b majj ] A; majj m BD a a. 32c jathare tu ] AB; . s ja(th - - ?) D 32d dakse ] AB; daksa D adi et ] B; adiset AD 33b murdhni ] AB; mudhni D . . . . 34b prsthe ] B; prste A; prsteste D 34d daksine ] AB; daksin(a?) D 35a dutah ] corr.; duto ... ... ... .. . . . . ABD sprsat siddho ] conj.; sprsam siddham B; sprsam siddhah AD 35b v ma ] AB; a . . . . . . . v m(e) D a sprset ] B; sprset AD 35c grahenaiva ] B; grahenaiva AD 35d kulatabh vam . . . . a s a a adi et ] conj. (Isaacson); kulatabh vin diset AD; kulath bh van di et B a a a s 36a oktam ] em.; okt m a. . . . ABD 36c samkoc l ] em.; samkocc ABD a a 36d lupt c ram ] AB; luptoc ran D a a . a 37b . . purvo ] AD; purvv D a 37c vij ny n ] B; vij ny AD a a a a 37d v make ] AD; +v make+ B a a 38a yoge i ] corr.; yogesi AB; yogrami D s 39a sammatam ] A samnmatam B; setmatam D . . . . . 39b v ky ] em.; b hy ABD a a a a l p tha ] AB; l po tha D a a a 39d sasanam ] B; s sanam AD a


329 cary yogakriy yog c chivecch s dhakasya tu | a a a a a yad drstiva am y t yoginyo martyasamgat h | a ... s . a a a. . . katham jey h svarupena rupam t s m tath vada a. a . . a a. bhairava uv ca a atah param pravaksy mi yoginn m tu laksanam | a. . . . a . . yena vij tam trena trailokyam va agam bhavet a a . s . . vij yate sudure pi ksetram rge vi esatah | a a s . . . bhutale caiva vartante sivecch sv dhik rik h a a a a. cary yuktasya deve i drster ay nti gocaram | a s ... a tasm j jeyam tu vrena yogen m tu laksanam a . s a . . . . a. tisro rekh lal te tu urdhvasmantam asrit h | a a. gaur campakagandh ca brahmacaryarat sad a a vedaghosapriy nityam aksobhy satyav din | a a a . . dandam kamandalu caiva ajinam yogapattakam .. . .. . .. srucdarbhopavtam tu padma ca likhitam grhe | . . . laksitavy prayatnena brahm nyamsa var nane a a. a . . tadarcanam tu vrena khecaratvajigsay | . . a . :::::::::::: brahm nkulaj devi sv msasiddhiprad yik a. a a. a a 47 46 45 44 43 42 41 40

a gandayoh kupake yasy h kundalagr grake in | a. .. . . . :::::::: s

A: f. 279r 40a kriy ] B; kry AD a 40c va am ] corr.; vasam ABD s . 40d samgat h ] em.; sama. . a . . . gat ] ABD a 40e jey h ] em.; jey ABD a. a 40f rupam ] AB; rupet D vada ] B; vadah AD . . 41d va agam ] B; vasagam AD s 42b ksetra ] AB; ksatra D 42d dhik rik h ] em.; dhik rik a a. a a . . . . a a ABD 43b drster ] em. (Isaacson); drstir ABD ay nti ] AB; ay tti D 43c tasm j ] B; a ... ... tasm AD a jeyam tu ] AB; ( - - - ?) D 44a tisro ] AB; tisr D a 44b smantam . a. asrit h ] em.; smantag srt A; smantag srt B; smattag srt D a . a a .a a . a 44c gandh ] AB; gat