A history of Indian literature, Volume 2, Part 2. Epics and Sanskrit religious literature. Hindu Tantric and Sakta Literature

Citation preview

TEUN GOUDRIAAN and SANJUKTA GUPTA HINDU TANTRIC AND SAKTA LITERATURE

}\HISTORY OF INDIA_N LITERATURE EDITED BY J.AN GONDA

VOLUME II Fasc. 2

1981 OTTO HARRASSOWITZ · WIESBADEN

TEUN GOUDRIAAN and SANJUKTA GUPTA "'

HINDU T ANT RI C AND SAKTA LITERATURE

1981

OTTO HARRASSOWITZ · WIESBADEN

A HISTORY OF INDIAN LITERATURE

Contents of Vol. II

Vol.II: Fasc. 1:

Fesc. 2:

Epics and Sanskrit Religious Literature J. Gonda Medieval Religious Literature in Sanskrit T. Goudriaan and S. Gupta 0. Botto L. Rocher

Hindu Tantric and Sakta Literature Sanskrit Epics PurliJ:).as

{)IP-Kurztitelallfna.hme der Deutschen Bibliothek A hlstory o! lndla.n literature I ed. by Jan Gonda.- Wiesbaden: Hnrrassowitz. NE: Gouda, Jan [Hrsg.] VoL 2. Epics and Sanskrit religious literature. Fasc. 2.-+ Goudriaan Teun: Hindu tantric and SAkta literature

-Goudrlaan. Teun: Hindu tantric and Sakta literature I Teun Goudriaan and Sanjukta. Gupta.- Wiesbaden: Harrassowitz.1981. (A history of Indian literature; Vol. 2, Fasc. 2) ISBN 3-447.02091·1

NE; Gnpta, Sanjnkta:

C Otto Harra.ssowitz, Wiesbaden 1981. Alle Rechte vorbehalten. Photographlsche und photomecbanische Wiedergabe nur mit ausdr!l.cklicher Genehmigung des Verlages. Gesamtherstellung: Frledrich Pustet, Regens'b!ll'g, Printed in Germany. :Sigel: HI L

PREFACE The authors are very grateful for the kind permission given by the authorities of the German-Nepal Manuscript Preservation Project to make use of the manuscript material collected within the scope of the Project; and for the good services of the Staatsbibliothek Preussischer Kulturbesitz at Berlin, and especially of Dr. Dieter George, in preparing microfilm copies from the Project's collections. Sincere thanks are also due to Dr. Michael Witzel, Dr. Karel van Kooy and Drs. J an Schoterman for their activities in procuring or interpreting rare or unique manuscripts.

CONTENTS

V

PREFACE

PART I:

TEUN GOUDRIAAN, Hindu Tantric Literature in Sanskrit CB:.A:PTER I General Characteristics CHAPTER II The Early Tradition .. C:H.!.:PTER Ill Tantras of the Srikula CHAPTER IV Tantras Devoted to Kali and Some Other Goddesses CHAPTER V Some Other Kaulatantras and Tantras on Restricted Subjects CHAPTER VI Va~avatantras and Tantras of Other Sects .. CHAPTER VII Tantras of Magic CHAPTER VIII Digests of Mantrasastra CHAPTER IX Ritual Compilations .. CHAPTER X Monographs on Miscellaneous Subjects CHAPTER XI Some Works of Speculative Contents

IX 1

32 58 75

92 105 112

130 141 157

162

PART II:

SANJUKTA GUPTA, Tantric Sakta Literature in Modern Indian 173

Languages . .

CHAPTER I

Introduction

175

CHAPTER II

Tantric Literature in Bengali

185

CHAPTER Ill

Tantric Literature in Hindi and Related Languages

202

Bibliography

215 219

Index

224

Abbreviations

PART

I

TEUN GOUDRIAAN HINDU TA..."I\ITRIC LITERATURE IN

SA...~SKRIT

CHAPTER

I

GENERAL CHARACTERISTICS

It is self-evident at first sight that the literature of Hindu Tantrism comprises those Hindu texts which proclaim Tantric doctrines and ritual. But there is a problem as to the extent of "Tantrism". What is most often called by this term is a systematic quest for salvation or for spiritual excellence by realizing and fostering the bipolar, bisexual divinity within one's own body. This result is methodically striven after by specific means (kinds of sadhanfi): the recitation of mantras or bijasl; the construction of geometrical cosmic symbols ( marJJ!,ala) ; the making of appropriate gestures (mudra); the assignment or "laying down" (nyasa) of powerful sounds or syllables on the body; the meditation on the deity's concrete manifestation (dhyana); the application of these and other elements in special ritual procedures, to wit Tantric worship (piija), initiation ( dik§a) etc. 2 ; besides, the performance of Km;u;laliniyoga by means of which the microcosmic form of the Sakti (female divine power) present in the body in the form of a fiery tube or serpent is conducted upwards along the yogic nerves towards Siva's mystic residence at some distance above the heada. All these procedures, which can only smnmarily be alluded to here, might be characterized as "Tantric elements", and they occur more or less frequently in · several sects, schools or religions which would not always without more admit to be "Tantric", such as the Paficaratra, Saiva Siddhanta, Buddhism and Jainism. Their literatures have been or will be treated in other volumes of the "History of Indian Literature". The present part restricts itself to those Saiva and Sakta works which are commonly called Tantras or which without any doubt belong to what is agreed on as the "Tantric tradition". This is not the place to go into details of doctrine or ritual practice; the essentials are assumed to be more or less familiar to the

1 J. GOND.A., The Indian Mantra, in: Oriens, 16, 1963, p. 244-297; reprinted in: J. GOND.A., Selected Studies, IV, Leiden 1975, p. 248f.; PADoux, Recherches, p. 293f.; HoENS, in: GUPT.A. a.o., HT, p. 91f.; ZVELEBIL, Kumaratantra, p. 178£.; on mantra and the other "Tantric elements" mentioned here, see also GoND.A., MRL, p. 67-75, with literature. 2 Piija: Gu:PT.A., in: Gu:PT.A. a.o., HT, p. 121f.; Dik~a. see HoENS, in: Gu:PT.A. a.o., HT, p. 71f. and (more general) GoND.A., Change and Continuity, p. 315-462. a A. Av.A.LON, The Serpent Power, Madras41950; S.B. D.A.sGUPT.A., Obscure Cults, p. XXXIV-XL; ELI.A.DE, Yoga, p. 205f.; and many popularizing books.

2

T. Goudriaan . Hindu Tantric Literature in Sanskrit

reader'. But it should be kept in mind that the distinction Saiva: Sakta cannot always be clearly drawn. Sa~m, the belief and worship of_ t~e Supreme Principle as a female force or Sakti (accompamed by a male divme partner, usually Siva or one of his manifestations) which creates, regulates ~n~ destroys the cosmoss, as it were grew into maturity under the cover of SatVIsm_ where Siva holds a similar position. This holds good at least as far as the literary development of Sakta ideology is concerned. Therefore, if the following pages "Will be mainly devoted to Sakta literature, it is unavoidable that some attention is also paid to tantrically oriented works which focus on the worship of Siva or other male gods. There is no clear line of demarcation; both denominations can be distinguished but not separated. Besides, many Tantras and secondary Tantric digests deal with subjects which allow of an eclectic standpoint, for instance the "science of mantras" or magic. These and other factors - and not in the last place the fact that Tantrism is still a living and not unimportant aspect of present-day Hinduism-lend to Tantric literature a significance and variety which is at first sight scarcely discernible6• The study of Tantric literature has often been neglected in the past. Winternitz' standard work devotes less than two pages to the Tantras at the end of the chapter on Pural).ic literature7. The catalogues are still the main source of information on most of the Tantric books in Sanskrit; many works have been edited, but the number of critical editions is very small. The present book is the first non-cataloguing survey claiming a certain degree of generality in scopeS. Systematic further investigation into the field is important and urgent; important because of the intrinsic value of this province of Indian literature as the literary heritage of an extremely influential stream in Indian religious history, even if the contents may be incomprehensible or displeasing to many;

U:

4 Cf. GUPTA a. o., HT; CH.A.KRAVARTI, Tantras; GONDA, Rei. Indiens, II, p. 26-52; works by AB.THUR AvALON alias Sm JoHN Woon:&OFFE; J. Evou, Le yoga tantrique, Paris 1971 (who bases himself mainly on AvALoN) . • 5 For delineations of Saktism, cf. GouniUA.AN, in: GUPTA a.o., HT, p. 6f. and the literature referred to there. For presentations of the bewildering variety of Sakti's (or Devi's) manifestations, we refer to the same work, p. 56 and 64f. (too short) and to PusHPENDRA KUM.AR SHARMA, Sakti Cult in Ancient India, esp. on p. 228f. Cf. also D. C. SmeAR, Tantrasaradh:rta Dhyanamala, in: Journal of Ancient Indian History, VT, 1972-73, p. 186-278, which contains descriptions of deities taken from the Tantrasara by ~~ANANDA and provided with a translation. 6 As "literature" we regard anything which an author deemed worth to be committed to writing, ~d w~ch following generations took the trouble to copy. No standards of aesthetics or literary competence have been applied. 7 WINTE:&NITZ, GIL, I, p. 481f. The references given there are completely outdated by now. 8 A short survey, of use especially for the Nibandhas, is CH.AKRAV.ARTI Tantras P· ?9 f. ~RARATI, Tantric Tradition, is not a reliable guide, BAGCRI, Studies, is re~ stnct.ed m scope. Useful, but written in Sanskrit, is DWIVEDA, Introd. to the NSA. On the early Kashmir tradition, see PANDEY, Abhinavagupta, p. 132f.; 542f.; PAnoux, Recherches, p. 41 f ..; RA.sTOGI, Krama Tantricism.

General Characteristics

3

and urgent, because a great number of manuscripts, neglected by the present generation, is in danger of getting lost9. It is therefore a happy circumstance that in the recent period projects have been set up to preserve these manuscripts or their contents by purchase or microfilming. Some compilers of library catalogues have done pioneer work in describing a great number of Tantric manuscripts. Without detracting from the merits of others, mention should in any case be made of Rajendralala Mitra, of Hara Prasad Sastri (Nepal Cat.) and of Chintaharan Chakravarti (RASB Cat-)1°. These catalogues have been heavily relied on, and profusely been referred to, in this book. Besides these, a great number of texts had already been edited in India before the end of the nineteenth century and the stream of editions went on, although with fluctuating intensity. But almost all of these were uncritical and destined for a restricted public of devotees or admirers. Much better known became the editions in the Tantrik Texts Series set up by Sir John Woodroffe and in most cases prepared by pandits who collaborated with him. Although these editions are far from definitive, W oodroffe was the first European to set himself to make the Tantras better known to (and to defend them against) cultured opinion in the modern worldll. More recently, attempts at critical editions of adequate standard have at last been made12, but this work ·has as yet scarcely begun. The number of translations into Western languages is negligible. In short, the literature of Hindu Tantrism, expressive of ideas and practices often scorned by the orthodoxies of former and present generations, has always been a stepchild of Indology. In this state of affairs, it will be understandable that this book cannot be a final appreciation based upon results reached by generations of investigators. On the contrary, it does not pretend to be more than a preliminary assessment of the material, a kind of repertory which, though imperfect and certainly not complete, can at least give an idea of the general contents, the chief characteristics, the diffuseness and the mere size of this literature, and show the interested student the way to the catalogues, the manuscripts and, if they exist, the editions.1s These sources will therefore be regularly referred to in the notes. 9 A similar note of warning was already sounded by AvALON, Introd. to the Tantrabhidhana, p. IV f. 1o R. Ml:TRA's "Notices on Sanskrit Manuscripts" was inaccessible to the present author, but is profusely paraphrased in KAVIRAJ, TSah.-For the titles of the catalogues consulted, see the Bibliography. 11 On WOODROFFE, see DERRETT, :Maha-Nirval).a, p. 233f. {stern judgment). According to EvoLA, Yoga tantrique, p. 22, Woodroffe used the pseudonym Avalon when he had been assisted by Hindus in the preparation of a book. 12 E.g., DWIVEDA's ed. of the NSA; OA.RLSTEDT's studies on the KT. 13 A recent repertory in Hindi, the Tantrika Sahitya prepared under the supervision of GOPINATH KAVIRAJ, has been made use of intensively. It is not an original work; it consists of an alphabetical list of 4422 Tantric texts provided with references to and extracts from relevant library catalogues (not all collections are cov·ered). The NCO (not utilized by the TSah) also served as a most welcome source of material. For tracing old editions, the IOL-SB proved invaluable.

4

T. Goudriaan ·Hindu Tantric Literature in Sanskrit

In the absence of anything like a clear chronology14-especiallymany "Original Tantras" offer little scope for chronological argumentation-the material has been ordered in the first place according to literary genre or method of presentation. The chief distinction adopted is that between "Original Tantras" and "secondary works", i.e. in most cases nibandhas or digests 15 • The "Original Tantras" are anonymous and written in dialogue form; their Sanskrit is often inferior, their style mediocre, sometimes even awkward; their method of presentation is repetitive, associative and non-systematical. The digests are of known authorship; their language and style are much better; their set-up is more systematical; they abound in references to and quotations from older authorities. The authors were in overwhelming majority Brahmans16 , but other social groups are also represented, among them ruling aristocrats17 • The terminology adopted here should not be taken to imply that the "Original Tantras" are always older or of more fundamental importance than the "secondary" Nibandhas; the reverse is often the case. However, the Original Tantras are given most of the attention in this book, although an attempt has been made not to neglect some other genres ofTantric literature. In some cases, even unimportant or quite modern products have been mentioned in order to present a true picture of the variety of material and the continuing productivity in the field. A secondary principle of division has been that of adherence to a particular deity worshipped as supreme. But here also, it is difficult to be consistent. Many texts are eclectic and treat of several deities together. Tantras of magic have been treated separately. Within each group, the chronological principle has of course been followed as far as possible. Whereas the chronology and early history of the Tantras are beset with problems and therefore still insufficiently known, the texts themselves are in complete agreement about their ultimate origin. The verbal tradition-inseparable from the doctrines communicated in it-was inaugurated by Siva himself, the Supreme Lord and Divine Teacher, although he may act in one of his divine manifestat~ons (Bhairava, Aghora etc.). 1B The first recipient of the tradition is usually Siva's spouse, Devi, or one of her manifestations; in a few-

14 A courageous, but somewhat abortive, attempt at chronology was made by F ARQUR.A.R, RLI. 15 The same division is applied in the RASB Cat., with a third rubric (stotraa and kavacas) added. For Stotras, see GONDA, MRL, p. 232f. For a similar division within ~heT~ school: P~oU::, Recherches, p. 53.-The distinction of Original Tantras ~to Miilatantra or Ntrdesa and Laghutantra or Uddesa is scarcely valid for Tantric literature as a whole (cf. H. NAKAMlJRA, Religions and Philosophies of India Tokyo 1973, IT, p. V-9, referring to M.J. SHENDGE). ' 16 Cf. CRAXRA.VARTI, RASB Cat., p. XXXII. 17 See below, p. 146f. 18 See e.g. the SvT, 8,27; MVT 1, 7; KKV 51 f.; PKS 1,1 and R.!MESVARA's commentary, Introd.; cf. also CHAXRAVARTI, Tantras, p. 25, who refers to BHOJA's Tattvaprakasa.

General Characteristics

5

usually early-texts, Skanda/Kumara acts as such, while occasionally the roles of Siva and Devi are reversed19. In a more speculative context, however, the sacred word is described as a function of the evolution of speech as a pure and primeval stage of cosmic evolution. In such cases, the boundary between legendary history, myth and metaphysics is almost imperceptibly crossed. For Abhinavagupta, the Siva-Devi dialogue is a refiexion of the Supreme on Itself, expressed in the framework of a teacher-pupil relation2o. This discursive relation between guru and 8i§ya is placed within the third (or Madhyama) of the four evolutionary stages of speech21, Opinions come to differ at the point where the second stage of transmission, the descent from the divine to the human plane, is to be described. In mythological language, this stage sets in when Dev'i, though asked for secrecy, cannot keep her vow. The younger Tantras as a rule do not bother much about it. Siva just enlightens Devi, and that is enough. In the older texts, divine emanations and ~~?is are often important intermediaries in the process. In an old source like the 1\TTS, the dialogue Siva-Devi is still embedded in the Pura:r,llc-Agamic situation of a colloquy between two Vedic sages; in the MVT, the secondary transmitter is Kumara who enlightens Narada and others22_ The description of a large chain of intermediary stages without doubt contributed to a feeling of reverence for an ancient and superhuman tradition and it suited a new movement which could not supply a parampara of ancient and respected gurus. Abhinavagupta in his Tantraloka (eh. 36) claims fourteen intermediaries between Bhairava and the Yogins; besides, due to deteriorating circumstances in the Kaliyuga, the tradition went lost until Siva as Sr'ikaQ.~ha (a pivotal figure in this context) caused it to reappear by word of mouth from Durvasas23. The prehistory of the Kula tradition, however, is differently described by the same Tantraloka in eh. 29: in the four world-periods there were four mythical expounders called Khagendra "Lord of Birds", Kiirma "Tortoise", Met?a "Ram" and Macchanda ( =Matsyendra "Lord ofFishes"n. All of these had wives and sons/pupils to whom they transmitted the doctrine. The latter tradition is simpler, less intellectual and certainly closer to popular legend. A few instances from other texts: in the SvT (8,27f.), Siva's invisible form 19 Thus in the lost Kramasadbhava; the KulCT; the RY-UT. Such texts are called Nigamas (cf. RASTOGI, Krama Tantricism, p. 229f.); the term Nigama is, however, often used by Tantric authors as a synonym for Veda, cf. Aw.A.STIII, Rahasya, p. 7f.; S.A.STRY, Sidelights, p. 4f.; DWIVED.A., ATKD, p. 28.-0n the dialogue situation as a literary motif, see below, p. 26. 2o .A:sHIN.A.VAGUPTA, Comm. on Paratrilp.Sika 1, ed. M.R. SHA.sT.Ri, Srinagar 1918 (KSTS, vol. 18), p. 6f., esp. on p. 12 where he quotes SvT 8,31; the passage is quoted by R.AMESV.A.RA on PKS 1,2 (ed. GOS, p. 20). 21 Discussion by KAvmA.r, Saktad~ti, p. 45f. The Madhyama stage is described by PADoux, Recherches, p. 168f. 22 GONDA, MRL, p. 203. 2a Cf. PANDEY, Abhinavagupta, p. 136, who refers to Bhagavadgita 4,1 for the idea.

6

T. Goudriaan · Hindu Tantric Literature in Sanskrit

is the first emitter; over Sadasiva, the tradition reaches Isvara, who proclaims it to the Vidyesvaras and to Srikal).tha from whom it is handed down to Bhairava who in his turn intimated it to Devi; she is the cause of its further spread among gods, demons, sages and human beings. The BY (cha~ter X_:K:~:IX) describes how the revelation originated from the :first creat1ve prmCiple, the Bindu, because of the activity of the Sakti, and how through various intermediaries it came to Srikal).~ha and from him to Isvara who taught it to Devi in the form of 125.000 anU§tubk stanzas. After that, the doctrine was received on earth by "a Brahman from Kuruklj!etra" and an inhabitant of Ogra (presumably UQ.~yana in this context) called Devadatta24. The Deviyamala pretends to transmit an original revelation by ten Rudras as first gurus 25 : "The ten mighty Rudras are known to have been the best of gurus; they wore plaited hair and crown and were always intent on worship of the linga; they constantly enjoyed wine and meat, and consistently performed the handling of mantras; they had intercourse with their female partners and diligently concentrated on the Scriptures".

The MBT (colophon) holds that the Primeval Manifestation was the first cause of the tradition, from which it came down along the Meru towards two pi(has "Seats of Revelation": the Adya- and Vidyapitha, and from there split itself into several schools; the MBT itself represents the Kadibheda. The Srividya school places Kama as Vidyesvara (a form of Siva) at the head; he communicated the doctrine to twelve subordinate Vidyesvaras, i. a. Manu, Indra, Manmatha (Kamaraja, the "real" god Kama), Lopamudra and Durvasas who founded separate lines of tradition. Only the lines of Kamaraja and (for some time at least) of Lopamudra survived into the historical period26. The V edic sages, as has been said, continue to play an important role, but only as transmitters, not as revealers 27 . It should be noted that each mantra possesses its r§i who is often mentioned together with its deity, sakti etc. and assigned to parts of the speaker's body (r§yadinyasaps. In Balinese Saiva Tantrio speculation, a group of five r~is acts as a kind of culture heroes29. 24 BAGCHI, Studies, p. 102f.; see also p. 111 and 106 for the traditions of the JY and the PM.· 25 Quoted by JAYARATRA on TA 28, 391-393. 26 KA..VIRA.J, ed. YR, 2 1963, p. IV f.-Elaboration of the doctrine in KAN 15 141f. Differently, and akin to the doctrine of the SvT: BHA8KARA.RAYA, Setubandha 4 7f. (cf. CHAKRA.VARTI, Tantras, p. 25, n. 26) . • 27 GoUDRIAAN, in GUPTA a.o:, HT, p. 14. There is a list of eight groups of rt!is m the JY, eh.~~ (BAGCHI, Studies, p. 112). Pippalada figures in magical kalpas such a~ the .Pra~angrra.sakalpa; Parasurama in the Tripurarahasya and the PKS ; Durvasas l.S claimed by both the Trika and the Srividya schools. 28 For a Tantric definition of an "r~", see S. C:a:. SIDDI:l:A.NTAB:a:uSAN comm. on Ka.Iitantra 1, 8. A similar role is played by r~is in the Paiicaratra, ~f. BB:ATTACRARYYA, Foreword to the ed. of Jayakhyasa:rphita by E. KRISBNAMACB:ARYA, p. 6 (JayS. 1,42; 1,49; 1, 74). 29 Cf. Q. HooYXAAS, Cosmogony and Creation in Balinese Tradition The Hague 1974, p. 93-128, esp. on p. 115. '

7,

B.

General Characteristics

7

To summarize: it will be clear that the Tantras claim superhuman authority because they present themselves as an inseparable part of divine emanation. Anyway, it was of vital importance for early Tantric groups to create their own parampara. Regrettably enough, it is impossible by now to discern how and how soon the transition from evident human authorship to divine authority was made. Of course the exalted position of the guru must have been an important factor in the acceleration and intensification of the process. What is, after all, a Tantra? The term itself is vague; it is applied in several other provinces of Indian literature to a technical "expose" or "handbook" which in a more or less extensive way deals with a certain subject. Sometimes it means little more than "tradition", "specialism", for instance in the word Bhiltatantra "the lore of exorcism of spirits" or Matrtantra "the set of rules for the worship of Mother goddesses" (NT 16, 76). A similar meaning can be assumed when Tantra- serves as the first member of compound titles: Tantrasadbhava "The Real Essence of Tantra". One could also speak of "the Tantra of Siva" or of Dev1so. Only in the course of tradition seems the term to have undergone a specialization towards denoting individual religious texts of Saiva and especially Sakta denomination. It would on the other hand be a serious error to hold that the whole of Sakta Tantric literature would be called Tantra. The use of titles is much too varied for that, although there has certainly been a tendency at work to increase the incidence of "Tantras" by simply putting the word behind other titles. In this way, Kular:r;tava became Kular:r;tavatantra; Tantraraja became Tantrarajatantra; Vi:r;tasikha was henceforth known as Vi:r;tasikhatantra; and so on. Even digests and commentaries might be subjected to the process: Umananda's Nityotsava is called Nityotsavatantra in some Mss., while K~?emaraja's commentary Uddyota "Elucidation" on the NT obtained the title Uddyotatantra in one Ms. The historical position of the term Tantra in the Tantric tradition is therefore not entirely clear and we may assume that it only gradually came to be closely affiliated with Sakta and Saktioriented Saiva literature31. Besides, "Tantra" is often scarcely distinguished from "Agama". The Bhagavatapura:r;ta (4, 24, 62) alludes to people who are experienced in Veda as well as Tantra and the Sanskrit commentary by Sridhara explains tantra by agama. In some older Tantric sources, the earlier literature of the school is referred to as "Agama"32. Some Saiva Agamas must have been transmitted for some time in the North of India as Tantras, while

30 BY (quoted by J"AY.A.RA.THA on TA 29, 11-13): ye punal:J, Sankare tantre Devitantre ea dikl}ita(i,. 31 When the MVT (1, 9) calls itself a Tantra (Malinivijaye tantre), this refers to the original revelation of 35 million slokas. The preserved summary is called Malinivijayottaram. 32 The Trika referred to the whole body of authoritative anonymous Saiva texts as "Agamasastra". Cf. PADoux, Recherches, p. 42; 57.

8

T. Goudriaan ·Hindu Tantric Literature in Sanskrit

even Paiicaratra Saip.b.itas could be denoted by the latter term33 • On the other hand a few Saiva Tantras are styled Sa:rp.hita (thus the Nihsvasatattvasa:rp.hita). Acco~cr to the Deviyamala as paraphrased by Abhinavagupta in his Tantraloka (28, 3s7) and quoted by Jayaratha on this place, a Sa:rp.hita is ch~r~~terized by a structure made up of four quarters (ptida). This fourfold diVISion has nothing to do with the theoretical division into Jfiana, Kriya, Yoga and Carya34 which plays a part in Saiva Agamic literature but scarcely at all in the texts discussed in this book. As far as nomenclature is concerned, therefore, it is impossible to be dogmatic. Thls is not to deny that there are characteristic differences between the typical Saiva Agaroa or Va~I}.ava Sa:rp.hita text on the one hand, and the typical Tantra on the other. In the Agamas, the emphasis is on those subjects which in the tradition fall under the heading of kriyli (and caryli): construction of temples, installation of images, temple worship, ceremonial bathing, public festivals etc., although there may also be expositions of cosmogony, roantra creation and other theoretical matter. The Tantras most often tend to prefer more esoteric subjects: speculations on the (usually bipolar) nature of the Absolute, cosmogony, the creative nature of sound and word, micro-macrocosmic equivalence, the powers of speech, communication and handling of mantras (very important and frequent), symbolic interpretations of words and namesas, construction of and initiation into rna1J4,alas and worship of deities therein (this ritual aspect of Tantric lore is also very important); many pages are also devoted to prediction, the obtaining of magical proficiency, varieties of Kui).