R. ADRIAENSEN, H. T. BAKKER, AND H. ISAACSON

TOWARDS A CR IT ICAL EDIT ION OF

THE SKANDAPURAIYA

It has long been doubted whether the well-known Skandapurdna ever formed a single textual whole before it was published as such by the Venkatesvara Press in AD 1910. No manuscripts are known that contain what could be considered as 'the complete Skandapur~na,' at least not one that conforms reasonably with the printed editions, which consist of a collection of khand, as. However, our knowledge of the early history of a text called Skandapurdn. a has been revolutionized by a large, though incomplete, Nepalese palm-leaf manuscript, 1 the importance of which was already realized by Haraprasad Shastri, but which, for one reason or another, so far has escaped serious attention of western Purfina scholars. We refer to this MS as $1.

This MS was used by Bhattarfii for his edition of the text under the title of Skandapurdnasya Ambik~khand. ah. (Kathmandu 1988). It is to Bhat.tarfii's credit to have noted the existence of two other Nepalese palm-leaf manu- scripts of this text, 2 of an age comparable to $1, and to have provided us with the editio princeps9 However, a closer look at his edition reveals serious shortcomings, which makes the edition unsatisfactory in many respects. For that reason a team (consisting of the present authors) has been formed at the Institute of Indian Studies (University of Groningen) to prepare a critical edition of this oldest version of the Skandapurdn. a (hereafter SP), using the three palm-leaf MSS from the National Archives.

A later recension of this text refers to itself as the Ambikdkhanda of the Skandapurdn. a. We have so far collected four MSS of this recension, one of which was also used by Bhattarfii, hence the title of his edition? A third recension, calling itself Revdkhan. d.a of the Skandapurdna (entirely different from the printed Revdkhand. a), is known to us from a MS in the Asiatic Society in Calcutta. 4 This MS differs significantly from the Ambikdkhanda and in some respects is closer to the Nepalese palm-leaf, or S recension. It proves to be invaluable for establishing the critical text, especially when the palm-leaf MSS fail and we otherwise would have been dependent on the A MSS, as indeed Bhattarfii was, viz., on A3. Neither the palm-leaf recension, nor the Ambikd and Revd recensions of this text have much in common with the printed Skandapurdn. a as far as we have been able to ascertain. This may be illustrated by three chapters dealing with Vfirfinasi. These are

Indo-Iranian Journal 3"7: 325--331, 1994. 9 1994 Kluwer Academic Publishers. Printed in the Netherlands.

326 R. ADRIAENSEN, H. T. BAKKER, AND H. ISAACSON

totally different from the Kddfkhanda, but one of them appears to be largely identical with a long passage in the 77rthavivecanakdn. d.a of Laksmidhara's K.rtyakalpataru, who ascribes this quotation to an, until now unidentified, Skandapurdn. a (TVK pp. 36--45, 130--5 ~- SP 29).

The palm-leaf manuscript $1 is in fact dated, as was first pointed out by M. Witzel in his On the Archetype of Patahjali's Mahdbhds.ya, in which he briefly refers to this very MS, though solely from the point of view of its pal~eographical interest. Witzel (in: IIJ 29 (1986), 256 n. 9) reads the date as 252 and says that this is MS.nadeva [= Arp. guvarman] sam. vat, equivalent to AD 810. We read with Bhattarfti the date as 234 [MS_nadeva samvat], which would indeed correspond to AD 810. 5 It is the oldest dated Nepalese manuscript known to us, and probably the oldest surviving manuscript of a Puranic text.

The antiquity of this Nepalese manuscript and the fact -- rightly em- phasized by Shastri -- that it refers to itself in the colophons simply as Skandapurdna, with no mention of khan. d. as, already suggest that this text is particularly worthy of notice and study. Another source-which casts light on its history is formed by citations from the Skandapurdna in early Nibandha works.

It seems that the practice of calling newly composed texts khan. d. as of the Skandapurdn. a might have started in the 12 th century. Laksmidhara still only quotes from an undivided Skandapurdn. a, which can for the most part be identified with the text we are presently editing. On the authority of Ball~lasena's Ddnasdgara (AD 1169--70) we know that in the 12th century three texts (khand. as) circulated, which told the stories of Rev~, Avanti and Paun.dra, and which were apparently assigned to the Skandapurdn. a. 6 However, the 'undivided' Skandapurdn. a still well-known and extant~ these were considered less authoritative and consequently BallSlasena preferred not to use them. Instead he quotes rather extensively in the Ddnasdgara from the SP. Out of the 204 pddas quoted only four could not be traced.

The general pattern that arises from an investigation of the Nibandha literature with respect to some khandas is that, when such a khanda had only recently been composed, Nibandha authors were generally reluctant to endorse their claims of being part of the Skandapur~n.a. To avoid commit- ment they often quoted from the khand, a by name without assigning it to the Skandapurdn. a, or they quoted verses from them assigning these to the Skandapurdn. a without mentioning that they actually had derived them from khan. d.a-texts. From Hazra's work it appears that not all of them occur in the printed editions. Hazra's conjecture that these, at least partly, belonged to Pur~nas that have been lost or replaced has proved to be right in a

TOWARDS A CRITICAL EDITION OF THE SKANDAPURANA 327

number of cases, as for instance the Skandapurdna at issue demonstrates9 However, the great number of unidentified verses found in the early Nibandha literature is also consistent with the hypothesis advanced earlier by Bakker, viz., that what eventually made its way into generally acclaimed Puranic compilations, was a redactor's choice out of textual materials locally produced9 It is through philological research based on manuscripts that this selection and the criteria by which it operated -- i.e., the genetic principles of Purfina literature as a whole -- can be brought alight9 No structuralistic analysis, taking printed texts for granted, will ever delve so deep]

From an examination of the testimonies it has appeared that the SP was known to Nibandha authors working in the north-eastern quarter of the sub-continent (Laksmidhara, Ballfilasena, Candegvara). The preponderance of the north-east tallies with the provenance of the SP manuscripts: the palm-leaf MSS are Nepalese, the 'Revdkhan. d.a' MS belongs to a Bengali tradition, and the Ambikdkhanda MSS we have examined all seem to be descended from a single MS in Bengali script9 It has also emerged that the SP was hardly, if at all, known in the Deccan in the 13th century and thereafter. Two other texts took its place, the Ndgara- and Prabhdsakhand. a, next to several, initially, unattached khandas.

It is evident that with the three palm-leaf MSS mentioned we have a most important early source of the cultural and religious developments in North India at our disposal9 They testify to a flourishing gaiva community in which lengthy, original Purfina texts were composed, of no mean literary merit to boot.

With the proclaimed intention of relating the birth and deeds of Skanda, the text gives us a cycle of gaiva mythology, which deviates in several respects from myths of ~iva known so far. In dealing with this god's various deeds and manifestations the Purfina appears to be quite original. With occasional touches of undeniable humour, the text delineates a godhead fond of disguises and tricks which often bewilder the other gods, and who is perhaps most frequently described as laughing. His darker and more destructive side is not neglected though, and the text also leaves no doubt that it is he who is ultimately the real creator of and power in the universe9 The text, entirely permeated by ascetic and yogic values, of which giva is made', the supreme personification, contains numerous references to 'Pfigupata-yoga' or a Pfigupata vrata and concludes with a section on Pfigupata yoga. In addition, there is some 'smr.ti-type' material, which is quoted fairly extensively by early Nibandha authors9

We have seen that external evidence, viz. the date accepted for MS SI,

328 R. ADRIAENSEN, H. T. BAKKER, AND H. ISAACSON

gives us a terminus ante quem in the early ninth century. Let us now see what internal indications there are which might be of use for establishing the age of the text, and in particular any which might suggest a tentative terminus post quem for its composition or compilation.

The SP explicitly refers to the Mahdbhdrata (Bhdratdkhydnam, SP 1.9a and 1.1 la), with the contents of which it was familiar, as appears not only from some verses that it borrowed, 8 but also from the myths and legends both texts share. These often make the impression of having been reworked to suit sectarian gaiva purposes. The text also contains versions of some myths which were almost certainly developed only some time after the materials found in the Mahdbhdrata. For instance, SP 60--68 present a developed form of the cycle of myths describing the Devi's slaying of the asuras Sunda and Nisunda, gumbha and Nigumbha and finally Mahisa.

The text presupposes and refers to the existence of PurS_nic texts, though there is reason to think that these may not yet have been very numerous. No names or numbers of Puffmas are given, nor is there mention of the concept of Mahfi- and Upa-purgnas. In fact, interestingly, in the places where the word purdna is used to refer to a text different from the SP itself, it is generally found in the singular and with no name or qualification added.

Another class of texts which was known to the SP is that of the Tantras. Though it is itself comparatively free of Tmatric and ~kta elements, there are several more or less clear references to Tantric texts -- particularly noteworthy is the occurrence of an Otherwise unknown list of mdtrtantras.

Such observations as these lead us to feel that it is very improbable that the text could have arisen earlier than ca. AD 500, and we are inclined to tentatively situate it in the seventh or eighth century.

As the preceding may have demonstrated, the publication of Bhattarfii's edition of the SP was a major event in Purfina studies. Or rather, it should have been. For reasons which are not completely clear to us, the edition seems to have gone entirely unnoticed. In the five years since its publi- cation, we have seen no monographs or articles, by Western or Indian scholars, which make use of the material it provides. Nor was the book reviewed in any journal available to us. We hope that this short note will lead Purfina specialists to take more account of Bhat.tarfii's edition, but in addition to this, we are jointly preparing a new and more critical edition of the text. The editing of Purfinas is however a somewhat controversial activity and one which very few Western indologists have undertaken.

TOWARDS A CR IT ICAL EDIT ION OF THE SKANDAPURANA 329

Several distinguished scholars have indeed expressed serious doubts about the possibility or value of critical editions of Purfinas. It is perhaps desirable to state here clearly how we expect to be able to improve on the editio princeps and say something of the principles which underly our work.

First of all, a new edition is worth undertaking because we now have at our disposal a wider range of manuscript materials than Bhattarfii was able to employ. Admittedly these do not include further Nepalese palm-leaf manuscripts, but in rotographs of $1 preserved in Oxford we have a record of an earlier and better state of the most important manuscript. 9 It was also possible in many cases to improve on Bhattarfii's reading of the Nepalese manuscripts. And for those portions of the text where all the palm-leaf manuscripts have suffered loss or damage of folios, the situation is much improved by having R and three other A manuscripts, whereas Bhattarfii was forced to base his text on A 3 alone. We are also able to take into consideration the testimonia and parallels that have been uncovered, of which Bhattarfi was apparently unaware, as he makes no mention of them.

The manuscript basis of our edition is thus broader than that of Bhattarfii's. In addition, the policies and principles we follow differ from those of the earlier edition. Like Bhattarfii, we attempt to some extent a reconstruction. Even the best manuscript -- which we judge to be S 1 -- gives a 1Lext which is quite unsatisfactory and indeed which can, by itself, sometimes hardly be interpreted. All the manuscripts have in our opinion suffered to a fairly large extent from purely involuntary scribal errors. We thus attempt a 'critical' edition, in the sense that we try to identify readings which have arisen as a result of such scribal errors and to correct them as far as possible. On the other hand, we are equally aware of the fact that many of the divergencies of the manuscripts are the result of deliberate change, a process that elsewhere has been described as 'composition in transmission, '1~ and an editor must realize that such processes were prob- ably already at work in a stage of the transmission of the text which predates all our manuscript witnesses. Given that a Puranic text has a certain inherent fluidity, it may be unwise to be solely intent on the recon- struction of the original. Furthermore, we are in basic agreement with those who emphasize that all versions of a (Puranic) text should ideally be taken into account, rather than concentrating solely on a single version which can never be anything but a hypothetical reconstruction.

We aim in the first place to edit the Nepalese recension, as far as the damaged and incomplete palm-leaf manuscripts allow, giving $1 the status of base-text when its testimony is available. This choice is justifiable not

330 R. ADRIAENSEN, H. T. BAKKER, AND H. ISAACSON

merely because S 1 is most probably older than $2 and $3, but also because, where $1 differs substantially from the other palm-leaf manuscripts, intrinsic and transcriptional probability often suggest that its readings are earlier.

The case is different for the portions of the text where we lack the palm- leaf manuscripts as witness. A comparison with the R and A recensions in cases where the Nepalese manuscripts are available shows convincingly that when they are not present it is impossible to reconstruct the Nepalese readings from R and A with any degree of certainty. Rather than omit these portions from our edition -- which would have disturbed the continuity of the story-line and deprived the reader of much interesting material -- we have decided in such cases to give an edited text based on the R and A manuscripts, set in a smaller type.

The main drawbacks of Bhattarfii's editorial practices are the following. First of all, he gives A 3 too much weight, to our mind, sometimes following it even where a good reading is available in the Nepalese manuscripts. This occurs not only on the level of individual words and readings but also on that of entire verses; in several places Bhat.tar~i prints verses present only in A 3. This would be less serious, were it not for another failing: the critical apparatus is markedly deficient in reporting accurately what the basis for the adopted reading is and what the testimony of the palm-leaf manuscripts actually is.

Though we thus attempt to give the text of the 'Nepalese' recension as far as the manuscripts allow, it is also our goal to present the divergencies of the other two recensions as clearly as possible in the apparatus. The manuscript situation allows a unique opportunity to study the process of transmission, involving on the one hand simple scribal corruptions, insignifi- cant variations and transpositions, and on the other hand major additions and 'recomposition.' By introducing a separate layer of apparatus for each of the recensions, we hope to facilitate the study of these aspects. It is our aim to find an acceptable balance between clarity and brevity in presenting a large amount of detailed information about the readings of the manu- scripts. Indologists may be little accustomed to a layout such as ours, with its multiple levels of apparatus, but we believe that once the trouble is taken to learn our conventions, the text together with its apparatus will be found to provide a mass of material for linguistic, textual and cultural studies.

In order to facilitate the use and accessibility of the text we give a synopsis of the contents of each adhydya in English. The edition is planned in approximately 10 volumes, the first one of which, comprising the adhygtyas 1 to 25 and an extensive prolegomenon is due to appear in 1995. Each volume will contain a register of proper names. Moreover, the

TOWARDS A CRIT ICAL EDIT ION OF THE SKANDAPURfitNA 331

Sanskrit text is entered into the computer in such a way that eventual ly a

pdda- and word- index can be produced. We shall make these indexes

avai lable in electronic form by putt ing them on a publ ic server. In addit ion

the colour photographs of MS S~ will be made avai lable on CD-ROM, on

which, possibly, also the indexes may find a place. It will be clear by now

that the project will spread over a considerable length of time, but we hope

that the last volume will appear before the year 2015.

NOTES

1 For this MS see Haraprasad Shastri, A Catalogue of Palm-leaf and Selected Paper MSS belonging to the Durbar Library, Nepal (to which has been added a Historical Introduction by C. Bendall), Calcutta 1905, p. lii, reprinted with a concordance by Reinhold Gr/inendahl, Stuttgart 1989, [Verzeichnis der orientalischen Handschriften in Deutschland, Supplement- band 31]; T. R. Gambier-Parry, A Catalogue of Photographs of Sanskrit MSS purchased for the administrators of the Max Miiller Memorial Fund, Oxford/London 1930, pp. 22--25; Brhatsitcipatram [Index of Old MSS in the National Archives, Nepal. By Bfibu Krsna garman] Vol. 8 p. 278 MS No. dvi. 229. The MS has been microfilmed by the N'G~cIPP: Reel No. B11/4. 2 Brhatsftcfpatram vol. 8, 278 No. pra. 831 (Bhattarfii's siglum ka, our siglum $2); vol. 8, 292 No. ca. 2260 (Bhattarfff's siglum ga, our siglt//n $3). 3 The MS used by Bha't'tarfii (his siglum gha, our siglum A3) is in the collection of the Asiatic Society of Benga] (Calcutta), MS 972 (= Haraprasad Shastri, A Descriptive Catalogue of Sanskrit Manuscripts in the Government Collection under the care of the Asiatic Society of Bengal, Calcutta 1928, No. 3921). The other three MSS are in the Asiatic Society of Bengal (Calcutta) MS 4554 (= Shastri 1928, No. 3922) in Bengali script (our siglum A1); in the India Office Library (London) MS 662--663 (our siglum A2) , and in the Library of the Sansklit College (Varanasi) MS 14311 (our siglum A4). They are all undated. A dated, but so far not obtained MS of the Ambikdkhanda recension is in the collection of the Sanskrit College in Calcutta (samvat 1955). Still another, possibly important MS, is in the collection of the Dacca University (No. 3376). 4 MS 3656 (= Shastri 1928, No. 3909). 5 On Mfinadeva or Amguvarman era see in the first place Petech, Mediaeval History of Nepal (c. 750--1482), Roma 1984, 12. 6 Ddnasdgara, edit. by Bh. Bhattacharya, Calcutta 1953, p. 7. 7 Bakker, Ayodhyd, Groningen 1986 II, XXIV--XXVII; Bakker in: xxiii9 Deutscher Orien- talistentag. Ausgewiihlte Vortrtige, Stuttgart 1989, 338. Whether or not some redactors of khandas may have had 'an agenda very like that of many Indologists' (Doniger, Purdna Perennis, Albany 1993, 60f.), for those in search of such a hidden 'agenda' of the Panrfinikas there seems to be no other way than to scrutinize the sources, that is the manuscripts, no matter how cumbersome, unspectacular and time-consuming that may be. s MBh.9.44.25ff. ~ SPBh 164.64ff. 9 See above note 1. 10 Bakker 1989, 331f. See n. 7 above9

University o f Groningen