Spink, Walter M. - Ajanta. History and Development. Vol. II

AJANTAARGUMENTS ABOUT AJANTA

HANDBOOK OF ORIENTAL STUDIESHANDBUCH DER ORIENTALISTIK

SECTION TWO

INDIAINDIEN

edited by

J. BRONKHORST

VOLUME 18/2

AJANTA: HISTORY AND DEVELOPMENT

ARGUMENTS ABOUT AJANTA

AJANTA: HISTORY ANDDEVELOPMENT

VOLUME TWO


BY

WALTER M. SPINK

BRILLLEIDEN BOSTON

2006

Cover illustration: Ajanta Cave 1, front wall Persian Embassy, detail 477.

This book is printed on acid-free paper.

Library of Congress Cataloging-in-Publication Data

A C.I.P. record for this book is available from the Library of Congress.

ISSN 0169-9377ISBN-13: 978 90 04 15072 0ISBN-10: 90 04 15072 2

Copyright 2006 by Koninklijke Brill NV, Leiden, The NetherlandsKoninklijke Brill NV incorporates the imprints Brill Academic Publishers,

Martinus Nijhoff Publishers and VSP.

All rights reserved. No part of this publication may be reproduced, translated, stored ina retrieval system, or transmitted in any form or by any means, electronic,mechanical, photocopying, recording or otherwise, without prior written

permission from the publisher.

Authorization to photocopy items for internal or personal use is grantedby Brill provided that the appropriate fees are paid directly to

The Copyright Clearance Center, 222 Rosewood Drive, Suite 910Danvers MA 01923, USA.Fees are subject to change.

printed in the netherlands

CONTENTS

VOLUME IIA. ARGUMENTS ABOUT AJANTA

Chapter One A discussion of H. Bakkers The Vakatakas 3Chapter Two Cave 26 as an inaugural monument .......... 22Chapter Three Cave 26s complicated development ............ 54Chapter Four Cohens Possible Histories ........................ 97Chapter Five Scholarly contributions to Maharashtra

Pathik .............................................................. 115

VOLUME IIB. PATTERNS OF PATRONAGE

PART I: PATRONAGE: CONSISTENT VS. COLLAPSING

Ajanta diers from most other Buddhist sites, generally created as community undertakings, in that it was purely elitist, developed by less than a dozen major patrons, who planned it with great care in essentially one great burst of enthusiasm. This enthusiasm turned to desperation immediately after Harisenas death, when the patrons anxiously rushed their shrine images to completion, in order to secure the merit from so doing. During the fteen years or so that the siteourished under the aegis of these proud donors, including the emperor Harisena himself, no outsiders ever could donate a single thing. But once the great patrons had rapidly departed from the collapsing site during the disastrous reign of Harisenas successor, the monks still resident there, along with local devotees, briey sponsored a helter-skelter spate of votive donations, also to make merit while they could. After about 480 this activity probably stopped completely, the craftsmen having gone away. The monks continued to live in some of the caves for perhaps as much as another decade after which the site was totally abandoned, except for the use of a few cells by Saivite sadhus and the like in later centuries.

Chapter Six Patterns of Patronage at Ajanta: Consistent vs. Collapsing .............................. 149

Chapter Seven Locating Intrusions in Time ........................ 158Chapter Eight Could Any Intrusions Date Before

Mid-478? ........................................................ 161Chapter Nine Caves 9 and 10: Their Redecoration

and their Intrusions Excavations: Dead or Alive .................................................. 167

vi contents

Chapter Ten Crises and Cave 1 .......................... 174Chapter Eleven The Breakdown of Patronage in the

Period of Disruption ........................ 192

PART II: PATRONAGE: THE HINAYANA CAVES

WITH EMPHASIS ON THEIR REDECORATION IN VAKATAKA TIMES

Although we can assume that, being so readily available, the old Hinayana viharas and caitya halls would have been used both for residence and worship when excavation work began again at Ajanta in about 462, it appears that no one troubled to redecorate or refurbish them until the new Vakataka phase was well underwayperhaps not until the very last year of Harisenas reign. The fact that these eorts came all too late is evidenced by the manner in which they soon had to be brought to a halt, due to the political situation. After this, intrusive donors took over, lling most of the still available areas with their votive oerings. All of these developmentscontribute to our knowledge of Ajantas turbulent history.

Chapter Twelve Patronage of the Hinayana Caves: Considerations .................................. 199

Chapter Thirteen Cave 10: Redecoration .................... 204Chapter Fourteen Cave 10: Intrusions: Summary ...... 221Chapter Fifteen Cave 10: The Aisle Paintings:

Original and Intrusive .................... 230Chapter Sixteen Cave 10: Faade Intrusions ............ 233Chapter Seventeen Cave 12: .......................................... 235Chapter Eighteen Cave 9: ............................................ 239Chapter Nineteen The Anomalous Painting on

Cave 9s Rear Wall ........................ 245Chapter Twenty Cave 9: Triforium Paintings;

Aisle Wall Paintings ........................ 251Chapter Twenty-One Cave 9: Palimpsests and other

Transformations .............................. 257Chapter Twenty-Two Cave 9: Intrusions on Pillars .......... 259Chapter Twenty-Three Cave 9: Faade Intrusions .............. 262Chapter Twenty-Four Cave 9: Considerations about

Usage .................................................. 266

Appendix Ajantas Inscriptions .............................................. 273

VOLUME IIA


CHAPTER ONE

A DISCUSSION OF H. BAKKERS THE VAKATAKAS

The Vakatakas: A Study in Hindu Iconology, by Hans Bakker (Groningen,1997) is an impressive exposition of important Vakataka sculptures,only recently being given their due, and of their revealing religiousand political context. However, some of Bakkers conclusions arehighly controversial, particularly with regard to Vakataka history inthe late years of the dynasty. I shall attempt to show how a veryclosein fact, approximately year by yearanalysis of the remark-able developments at Ajanta (and related caves) can lead to a revi-sion of such commonly held views. At the same time, this shouldgreatly magnify the image of the great emperor Harisena, who (inmy view) was responsible for the startling, and nal, orescence ofthe great Vakataka empire in central India during his brief reignfrom about 460 through about 477 A.D.

Too long regarded as poor relations of the Guptas, the Vakatakasthemselves, under the powerful Harisena, were in fact the nal spon-sors and guardians of the so-called Golden Age.1 However, whenthe great emperor Harisena died, the world of central India sud-denly was disrupted, and the Golden Age suddenly ended. Thisbecame evident almost immediately at Ajanta, dependent as it wasupon courtly patronage.

By the end of 478, Ajantas long established patronage had totallycollapsed; and by the early 480s, the huge empire itself, whichHarisena, through war, marriage, inheritance and intrigue, had puttogether, had been broken back into the constituent parts which hehad gradually welded together as he extended his domains from theeastern to the western sea.2 And that was the end of the Vakatakas.

1 To parallel this statement, I have taken the liberty of slightly revising the label-ing of the map of India in this period, to give the Vakatakas (at least duringHarisenas reign, equal status with the Guptas.

2 This spread of empire is evident from the listing of territories in his ministers(Varahadevas) Cave 16 inscription, verse 18; see Volume 1; the territories whichhe (using Bakkers term) stood above stretch from the eastern to the western sea.

4 volume iia, chapter one

Their rule did not continue into the early sixth century, as mostscholars believe.3 It ended with the rapid destruction of the empirein the early 480s, during the reign of Harisenas inept successor,Sarvasena III, who had inherited the nally unied empire from hisfather, but disastrously failed to hold his inheritance together.4

The sculptures with which Bakker deals are generally isolatedimages, often found in an architectural context. They were producedexclusivelyin the eastern Vakataka domains from the late fourthcentury into the last half of the fth or (Bakker suggests) the earlypart of the sixth.5 Many of these works have been recently discovered,or at least only recently recognized for what they are. A numberhave been published previously, notably by A. P. Jamkhedkar (1991,8592), who has brought so much of this material to light over thecourse of the last twenty years.6 Earlier, we knew next to nothingabout this fascinating body of work.

Furthermore, Bakker focuses new attention on the Vakataka dynasty,which has been almost as disregarded as these sculptures which itproduced. Clearly it is now due for a scholarly renaissance. As Bakker(1997, 2) succinctly states: One may say that from the middle ofthe sixties (of the fth century) the kingdom of the Vakatakas hascome to be seen as pivotal in the history of India, being essentialfor our understanding of the development of its art, religion and cul-ture; as such it is on a par with the Gupta world, of which it canno longer be considered to be merely a province.

It is gratifying to have the Vakatakas so appropriately raised instatus. All too often they get little credit beyond the astute marriageof the Vakataka king Rudrasena II to the Gupta princess PrabhavatiGupta in the late fourth centurythe stress being put on the Guptarather than the Vakataka connection. In the following discussion, Ishall attempt to further amplify Bakkers assertion of the Vakatakas

3 See, for instance, Khandalavala 1992, 123 .4 For a justication of such dating, and related historical considerations, see Spink

1991B, 7199.5 According to my chronology, the latest Vakataka sculptures in the Nagpur

region would not post-date 477 or possibly 478, when the Vakataka empire wascollapsing.

6 The famous Ganga from Paunar ( Jamkhedkar, 1991), plate 13 (page 92) andthe related Ramayana cycle installed at Vinobha Bhaves ashram appear to be post-Vakataka, probably Visnukundinreecting of course the previously strong politicaland even marital connections while Harisena controlled the empire.

a discussion of h. bakkers the vakatakas 5

importance, in part by arguing that Ajanta, in its great later phase,is in fact a Vakataka site.7 In the past, the credit for Ajantas ourishingwas typically diused, with shares being given to the Guptas, theCalukyas, the Rashtrakutas, and even the Asmakas.8 This has obvi-ously diminished the Vakatakas reputation; whereas to see the wholesite (along with Bagh, Dharasiva, Ghatotkacha, Banoti and AurangabadCaves 1, 3, and 4A) as all having been accomplished in less thantwenty years under Vakataka patronage, dramatically denes the cul-tural potency of the Vakataka empire which sponsored it.

When such a reinterpretation of Ajantawhich reveals the highcharacter of Vakataka culture in its various art formsis combinedwith Bakkers presentation of the previously undiscovered and/or little understood Vakataka material from Eastern Vidarbha, the reputa-tion of the too-neglected dynasty will better come to receive its due.9

Since his concern is to elucidate the manner in which art isembedded in a social and cultural contextthat is, to deal withiconology (Bakker 1997, 3)the information that such sources pro-vide can clearly add to ones understanding. This is especially thecase with the (Western) Vakataka site at Ajanta, which provides anillustrated history of Buddhist patronage, wherein so many social,political, and economic forces are at play.

Admittedly, Bakkers titleThe Vakatakasis misleading, for thisimplies that the book will give a relatively equal coverage to theEastern (Nandivardana) and the Western (Vatsagulma) branches ofthe house. However, it is essentially about the former. He has madea deliberate decision to leave out the specic Buddhist evidence . . . .one will not nd, for instance, an art historical assessment of the

7 Even from 478 to 480, when the Asmakas had declared their independencefrom their Vakataka overlords, Sarvasena was still the legal ruler, for he had notyet been defeated (in fact, killed) during the Asmaka insurrection.

8 The Asmakas were in fact among the inaugurators of the site in around 462,and still later, from 475 on, had taken it over completely, although still (from theVakataka point of view) nominally feudatories of the Vakatakas, even after theirrebellion in 478.

9 Bakker provides a generous selection of the little known Vakataka sculptures,along with very useful maps, although one regrets the lack of views of related struc-tures, such as the Ramagiri temples; nor does Bakker illustrate any of the impor-tant sites (such as they are), despite including a useful chapter describing the particularsites from which his illustrations are drawn. He also provides an excellent bibliog-raphy, a revised edition (unfortunately not translated) of the important Kevala-Narasimha Temple Inscription, and an (arguable) Outline of Vakataka Chronology.


Ajanta Caves . . . (Bakker 1997, 3). Although he breaks this ruleoccasionally, the focus of his book is on the Hindu materials fromthe Eastern Vakataka realm. These sculptures he analyzes with eru-dition, drawing upon an impressive body of textual sources and expli-cating their religious and political connections in an exemplary way,while leaving enough room for scholarly contention.

Bakkers second strategy is to utilize textual and archaeologicalsources in combination as much as possible. (Bakker 1997, 3) Buthere too he underplays his options by siding with convention, andgiving short shrift to the Visrutacarita (the eighth ucchvasa of DandinsDasakumaracarita) the one textual source which I am convinced canclarify Vakataka history most fully. In contrast, Bakker insists: Atbest (the Visrutacarita) could be seen as a kind of Alexander Romance,valuable in itself, though no historian would use it as his primarysource . . . (Bakker 1997, 37)

If I am correctand admittedly in this I am in total disagree-ment with most scholarsthe Visrutacarita provides us with clear evi-dence that the Vakataka dynasty came to an end under the Western(Vatsagulma) branch, not the Eastern (Nandivardana) branch. (Spink1991B, 7192) Furthermore, it ended far earlier (in the early 480s)than scholars normally suppose.10 At the same time Bakker (1997,5), claiming the limits set by (his) own competence, intentionallyavoids getting involved in exploring the rich evidence from Ajanta,which conrms what we speak of as the Short Chronology, oncewe see it as a dig and analyze its various strata.11 These twosourcesAjanta, and the Visrutacaritawhich should neither be solelyleft to art historians (Bakker 1997, 41), nor for reading as a mereAlexander Romanceare the essential keys to late Vakataka his-tory, and we exclude them at our scholarly peril.

Admittedly, my friend Karl Khandalavala (now sadly deceased),speaking for many, repeatedly warned that there is no end to Dr.Spinks make-believe theories (Khandalavala 1991, 123) and it istrue that my chimerical views by no means ring of established

10 Although I owe much to Mirashis fundamental study, Historical Evidence inDandins Dasakumaracarita (Mirashi 1945) my conclusions are based upon a sig-nicantly dierent chronology, resulting in a very dierent view of history.

11 See Spink 1991A (The Archaeology of Ajanta) for a discussion of the sitesrecurring crises and revealing strata with which they can be linked.


truth.12 I have nonetheless shown, utilizing my revised chronology(Spink, 1992, 23) how the Visrutacarita tallies remarkably with bothtextual and archaeological evidence; and I also show how an analy-sis of Ajantas development forces us to recognize the painful intru-sions of the aggressive Asmakas into the dynastic history of Harisenaand his successor Sarvasena III, whose great house they nallydestroyed.

It is too much to think that I can prove the remarkable relevanceof Ajanta just by writing about it. Readers should go there and treatthe caves with the same physical and mental precision that is typi-cally involved in the excavation of any major archaeological site.However, I can at least present something of the revealing com-plexity of its development. It is one thing to say with Bakker (1997,89) that we do not subscribe to (the) opinion that all Mahayanacaves (at Ajanta) were excavated in twenty years or less or, withWilliams, to prefer to think of Vakataka patronage at Ajanta asextending roughly from 460505 . . . (with a probable) absolute rangeof c. 460 to 530 (Williams 1982, 182, 186). But how could oneexplain to the established courtly patrons, all anxious to get theirdonations completed while they were still alive, why such work wouldhave to take so long, especially since all of the major caves at thesite, with the exception of two (Caves 14, 28), were started in thesame single initial burst of activity in the early 460s?13 Therefore,reversing the usual objection: How could it have been done soquickly? one should try to answer the more relevant question, giventhe intense commitment of its eager patrons, most of whom werealready of a certain age: Why did it take so long? How could ithave taken nearly twenty years!14

Ajanta is unique in its complexity, and in the manner in whichits complexity can be elucidated. Rather than believing that its assess-ment is only partially relevant for the political history (of the

12 For various critiques of my theories by K. Khandalavala, see Khandalavala1990; 1991A; 1991B.

13 A few very small and very unnished caves, placed at a higher level wherethere was still room, were also started very late, probably all in 477 (See TimeChart). One other, barely undertaken in 477, was at a lower level, between Caves21 and 23; unfortunately, it was recently mostly lled up with cement.

14 As I suggest in Volume I, fteen (or even fourteen) years might be a closerapproximation for the period of the sites consistent developmentexcluding thetwo or three years occupied by its subsequent Period of Disruption.


Vakatakas) (Bakker 1997, 40) we may nd that Ajantas analysis isthe most signicant key of all to that assessment. Indeed, it is myconviction that a proper analysis of Ajantas development, togetherwith a related reading of the Visrutacarita, must transform our wholeview of Indian history in the last half of the fth century, and willalso have a great impact upon our understanding of the chaoticdevelopments of the early sixth. Instead of seeing the great vessel ofthe Golden Age slowly being tarnished as the Guptas, starting withSkandagupta, gradually failed in power and authority in the last halfof the fth century, we must see the vessel polished to its greatestsplendor by the Vakatakas, during the reign of the great Harisena,before being shattered irreparably by his sudden and tragic death inc. 477.15 It was left to his successorsthe multiple dynasties whichmade up Harisenas foldto pick up the pieces. Indeed, they werethe pieces, now become separate again, which he had earlier forgedtogether in the constructing of his empire.

In the discussion below, I will make continual references to Bakkersopinions, even though it will be evident that I am trying to sell myown goods. Bakkers study is excellent in so many ways that I amconcerned lest the weight of its authority may damage the more del-icate structure of truth, even in those areas where (in my opinion)it is clearly wrong, notably in his reconstruction of late Vakatakahistory as well as in his brief forays into Ajantas development. SoI shall present my alternative view of these matters. But my inter-est is not so much to merely present the conclusions as to convincethe reader of the possibilities of analyzing Ajantas development invery precise detail. Such an analysis can lead to a new synthesiswhich goes well beyond the study of the caves themselves.

I should add a word in justication of my very explicit year-by-year dating of the developments in the caves. Ajantas brief devel-opment from c. 462 to c. 480 is so crowded both with forms andwith transformations that, more than any other site in the world, itcan be picked apart (in order to be put together), with what many

15 Obviously this exaltation of the long-neglected emperor Harisena is hardly partof current views of Indian history. Romila Thapar, the doyen of historians, doesnot even mention Harisena in her new edition of The Early History of India. Bycontrast, I have asserted that Harisena may have been the most successful ruler inthe whole world in the 460s and 470s. Such a statement may perhaps be seen asa devotees overstatement; but so far, no one has named a rival.


will see as a hardly believable precision.16 In fact, as I explain else-where, we must see my so-called short chronology as slightly exible,as if all of my dates are inscribed on an elastic band which couldbe stretched (or perhaps better, could be contracted) by a year ortwo at either end as well as at certain points between.17

If I am correct, the sequential relationships on this continuum,because of the insistently revealing nature of the sites interconnecteddevelopments, should hardly change, even if one should decide tomake a slight adjustment of its outer limits or of some of its inter-mediate sections: the (war-produced) Hiatus, for example, may wellhave occupied slightly more or slightly less than the three years(472474) which I have arbitrarily assigned to it on the basis of thevarious developments occupied within this brief period. (See TimeChart) Furthermore, to make these developments more usefullygraphicto make a clear distinction, for instance, between theevents of the year just before his death and the year just afterHarisena has to die, rather arbitrarily, on Dec. 31, 477, ratherthan on some indeterminate day in October or November. By thesame token, the Asmakas have to be expelled from the region, ratherdecisively, on Dec. 31, 468, while the sites lesser patrons, alreadyaected by the implications of this event, started eciently rushingtheir shrine Buddhas to completion on New Years Day, in 469 (seeTime Chart). The reader must forgive this obviously imprecise pre-cision, for the simple reason that it allows us to divide the sitesdramatic development into more workable units.

What I hope to show, in this explanation of a small portion ofAjantas myriad features, is that until we yield to the demands ofthe literally solid evidence that the site presents, we are not goingto get very far along the complicated pathway which can lead us toa satisfactory understanding of its amazing history. Ideally (and actu-ally) too, we must go to the caves themselves for our illustrations.The motifs and features which must be studied, as one moves toward

16 For a justication of such dating, and related historical considerations, seeSpink 1991A, 7199. See also Volume I: Chapters 2, 4.

17 Although my own approach has certain parallels with that of Philippe Stern,the concentration on single motifs which typify his own work and that of his asso-ciates, is diametrically opposed to my method, which involves a consideration ofthe interconnected myriad motifs which compose both the separate caves and thesite of which they are a part. See Schastok 2000, 714.


conclusions, are so many, so varied, and so interconnected, that nobook could properly picture or describe them suciently to satisfyall of the demands of a proper investigation.18 Therefore, the illus-trations included in volume four can eventually only be seen moreas an aide memoire than as sucient evidentiary material.

As noted above, Bakkers main focus is upon the Hindu sculpturefrom the eastern (Nandivardana) branch of the Vakatakas instead ofupon Buddhist monuments of the western (Vatsagulma) branch. How-ever, he also notes the inuence of these eastern productions uponthe generally later developmentschiey at Ajantasponsored bythe great emperor Harisena of the dynastys Vatsagulma branch.Clearly the evidence at Ajanta supports his view (Bakker 1997, 44)that a stream of artisans moving from Nandivardana to Vatsagulmawas certainly only one of the creative forces that contributed to thedevelopment of the Caves. I would only insist that this was onlyone of many streams, the totality of which was responsible for Ajantasrich complexityeven confusionof style. But certainly, as Harisenabecame more and more powerful, and ultimately became the soleheir to Vakataka power, workmen would have been increasinglydrawn from the major cities and sites under his control; and surelyNandivardana would have been most particularly represented.

Far more concerned than most scholars about the role that cul-tural forces play in nourishing such monuments as Ajanta, Bakkerrightly stresses the role of the merchants traveling upon the ourishingtrade routes, who would have played a signicant part in support-ing the costly Buddhist undertakings at Ajanta. To this I would addthe suggestion that the popularity in the caves of representations ofAvalokitesvara as Lord of Travelers would enhance Bakkers assump-tion; indeed, such images typically appear at moments of great polit-ical and/or military stress, but particularly from mid-478 to 480forthese were times when there must have been real danger upon theroads.19

18 For Khandalavalas fervent rejection of this fallacious methodology for dat-ing the caves, see Khandalavala 1991, 96.

19 There are also two examples in the caves of the local king, whose patronagesuddenly ended in 471 due to the Asmaka attack. One painted image is on theshort front wall of the porch of Cave 17, at the left; the other on the rear of theright porch pilaster capital in Cave 20; but both appear to have been done beforethe situation became critical.


One must agree too with Bakkers perceptive suggestion that thesewestern productions reect a newly self-indulgent world, with aneconomy apparently based more on luxury than responsibility, stillunaware and increasingly unprepared forso I would saythe delugewhich was soon to come. The fate of Ajantas local feudatory king,who covered the world with stupas and viharas (Cave 17 inscrip-tion) but was not prepared for the war which was to destroy him,was only reected on a larger scale by the fate of the great emperorshouse hardly a decade later. Considering the obsessive energy whichcould produce such a vast undertaking as Ajanta in less than twentyyearseven if not a single cave was ever nishedBakker is surelyright in seeing it as a heavy drain upon the economic resources ofthe kingdom (Bakker 1997, 44). But there may have been anotherthreat too in the western realm, where the emperor Harisena him-self, surely also devoted to the Buddha, allowed his courtiers toimprove their balance of merit and demerit by spending parts oftheir earnings or spoil on lavishly decorated Buddhist monasteries.(Bakker 1997, 45) Indeed, Buddhism is hardly a faith that preparesone for a particularly committed defense against aggression.

The fact (at least I see it as a fact) that when Harisena suddenlydied, his many courtiers, including his own Prime Minister, imme-diately and single-mindedly turned inward, to complete the shrineimages deep within their caves, only enhances Bakkers point. At thesame time he is surely wrong when he asserts that the king him-self seems to have taken only a marginal interest in these by-prod-ucts of his policy (Bakker 1997, 45). I would assert just the opposite:that Harisena was clearly committed to this great eorteven if per-haps only for reasons of politics and prestige. What cannot be denied(again this is my view) is that it was Harisena himself who wasresponsible for the donation of Cave 1, the most sumptuous rock-cut vihara ever made in India (Spink 1981, 144157). And the beau-tiful paintings with which he lled it, all focused on themes involvingthe character of kingship, do indeed exalt the virtues of sacrice andof generosity, rather than the more practical values of heroic aggres-sion. In fact it must have been with his approval, and perhaps withimperial support, that Ajanta burst so dramatically upon the scenein the early 460s; and his support must have been a factor in theroughly contemporaneous and equally rapid developments at Bagh,which provided a protable sanctuary for Ajanta workmen duringAjantas troubled Recession and Hiatus from early 469 through 474.


Harisenas religious aliation is a matter which has by no meansbeen settled. However, it is hard to believe that the donor of thegreatest vihara in India was not in some sense a Buddhist, despitehis Saivite ancestry. It is relevant to point out, with Bakker (1997,58) that The Vatsagulma kings call themselves Dharmamaharaja,without being more explicit about their religious orientation . . . Thetitle . . . may therefore be interpreted as a general one indicating an ecumenical attitude towards the religions represented in theirkingdom.

A similar suggestion of the perhaps fragile boundaries between thetwo religions is the fact that Harisena is given the title Haritiputrain the Thalner inscription, dated to his regnal year 3 (463, accord-ing to my chronology).20 This somewhat ambiguous title hardly provesthat he was a Buddhist, just as the reference in the Thalner inscrip-tion (c. 463) that he gave gifts to Brahmins (Mirashi 1982, 7885),nor the Visrutacaritas statement that he (maintained in order) thefour castes (Kale 1966, 349), would hardly prove that he was aHindu. The Visrutacaritas various references to the Saivite focus ofreligious activity in Mahismati, where Harisenas second son wasviceroy, and to which his two grandchildren were taken for safety,are more compelling. However, even Maharaja Subandhu (PrinceVisruta of the account), whose adventurous conquest of Harisenasgranddaughter took place in various Saivite contexts, proudly takescredit for repairing the magnicent (Buddhist) caves at Bagh. (Mirashi1955, 1921) It seems evident that this impressive complex wouldnever have been started without the support of the regional viceroy;and this in turn implies the encouragement of the emperor himself.

The Bagh copper plate inscription (approximately 486 C.E.),21 ofwhich I will include only the most relevant portion, is of great inter-est in showing the rulers active support of a Buddhist establishment;and it also revealingly describes much about the arrangement of themonastery and the activities still going on there vigorously, someyears after Ajantas total collapse.22

20 For the Hariti Shrine in Cave 2 see Volume V, Cave 2.21 The date on the Bagh plate is lost, but for the dating of the closely related

Barwani inscription to the Gupta era rather than (following Mirashi 1955, 1719)to the Early Kalacuri era.

22 See Volume I for the political benets of such generosity. See also Bagh: AStudy (Spink 197677, 5384).


(Line 4) Be it known to you that for the increase of the religiousmerit of my parents and myself, this village has been granted byme . . . in order that it may be used for (defraying the expenses of )perfume, frankincense, owers and oerings as well as for main-taining an alms-house, for repairing broken and rent portions (of thevihara) and for providing the community of Venerable Monks com-ing from (all) the four quarters, with clothing, food, nursing of thesick, beds, seats as well as medicine in the Monastery called Kalayana(the Abode of Art) to be constructed by Dattataka, as long as themoon, the sun, the oceans, planets, constellations and the earth wouldendure. (Mirashi 1955, 19)

A.K. Narain, writing at length of the ecumenical attitudes of theperiod, emphasizes the fact that both the Gupta kings and most ofthe contemporary kings and chiefs of India personally belonged toone or the other of Brahmanical religions or sects. . . . It is also true,however, that these kings were liberal in their outlook, and not onlytolerated faiths other than their own but patronized them, whetherthey belonged to the Brahmanical fold or to the non-Brahmanical.(Narain 1983, 48)

If we assume that the attitude of Harisena and other Vakatakakings was ecumenical, one is still left with the curious fact that nearlyall of the known Vakataka works from the eastern area are Hindu,whereas all of the known works (generally slightly later) from thewestern area are Buddhist.23 The latter sites would include Ajanta,Bagh, Aurangabad Caves 1, 3, 4A, the Ghatotkacha vihara, the caveat Banoti (except for sixth century revisions), various panels from thestupas in the Kanheri cemetery, and just possibly the cave at Lonad.24

At the very least can we not suppose that Harisena managed to sub-scribe to both Hinduism and Buddhism, partaking of the politicaladvantages of such ecumenicalism? This was certainly the case withHarisenas Prime Minister, Varahadeva, who in his Cave 16 inscrip-tion (regarded) the sacred law as his (only companion, . . . beingextremely devoted to (the Buddha), the teacher of the world; while

23 Although not mentioned in Bakkers study, there are a few Buddhist nds fromthe eastern Vakataka region, such as the bronze standing Buddhas from Hamlapuri(near Ramtek) and the Phophnar Buddhas. Jamkhedkar (1991, 8791) dates themvariously from the late 4th to early 6th, but precision is dicult with such rela-tively isolated gures.

24 For a brief discussion of these related sites, see Volume 1, Chapter 14.


in his Ghatotkacha inscription his forbears are praised for per-forming religious duties as enjoined by the Vedas and Smritis.25

The Vakataka feudatory Upendragupta, in his Cave 17 inscription,shows an apparently total commitment to Buddhism, while the closeconnections of the monk Buddhabhadra, patron of Cave 26, withthe Asmaka court, suggests an ecumenical attitude at high levelsthere also.

It is relevant to note that the (mostly) Jain caves at Dharasiva,which are located farther to the south, also appear to belong to theperiod of Harisenas reign. They lie in ancient Kuntala which, accord-ing to both the Ajanta Cave 16 inscription and the Visrutacarita laywithin his sphere of inuence or control. Although started by themid-60s, they appear to have been largely roughed out during AjantasRecession cum Hiatus (469 through 471), when many displaced arti-sans from the great site would have been available. Their proxim-ity to Asmaka was probably responsible for the decline of activityduring the following conagration. Perhaps for this reason the devel-oped images in their long unnished shrines reect those of the latestphase of consistent patronage at Ajanta (i.e. 477478).26 The inuenceof Buddhist (specically Ajanta) precedents on both of these mainimages, and on the character of the caves in general, is noteworthy.

Bakker is certainly correct in seeing the Eastern Vakataka pro-ductions as being an inuential source for the sculptural work atAjanta, even though they were at some remove in time and space.But this inuence is by and large limited, as far as we now know,to material such as the much earlier nds from Mandhal, most ofwhich were produced well before the mid-fth century. However, aconnection can be seen in the heavy corporeal forms of a numberof the sites earlier sculptureson the base of the porch doorway ofCave Lower 6, on the faade of Cave 7, on the sculptured ceilingforms of the front aisle of Cave 16, and various other examples,while some of the small but massive gures on the fallen faadeof Bagh Cave 3 have similar characteristics. But none of these exam-ples can be dated prior to 465 at the earliest, for it took a few yearsbefore Ajantas caves were ready for any such carved decoration,

25 Ajanta Cave 16 inscription v. 21, and Ghatotkacha inscription v. 6. However,in neither inscription is Varahadeva himself (or his emperor Harisena) said to havemastered the Vedas.

26 See Volume I: Dharasiva Caves.


which in the rst few years was generally not applied to the mon-uments until they were essentially roughed out.27 Even certain imagesfrom Ramagiri, and the remarkable nds at Mansar, which areindeed linked closely with Ajanta, may show the inuence of Ajanta,from around 470 or a bit later, rather than vice versa, as I shallsuggest below.28 As for the sculptures from Paunar, they cannot bedated until at least 500 A.D., as Bakker himself agrees, by whichtime I would ascribe them to some other patronageprobably Visnu-kundinas noted above. But here too there is a lingering Vakatakaconnection.29

Bakker is predisposed to his view of Vakataka history by his assumptionwhich I shall attempt to counterthat the Eastern(Nandivardana) branch of the Vakatakas ultimately absorbed theWestern (Vatsagulma) branch in the fth century. But even so, heis ready to admit (1997, 44) that the artistic explosion at Ajantacannot be suciently explained as being nourished by the ruralsmall-scale economy of Nandivardana, even though the EasternVakatakas must have contributed something to Ajantas development.It was clearly fed by a far wider world, including the inner universeof its insistently elite patronage. To give a sense of the power of thispatronage, it is remarkable (if my conclusions are correct) that untilthe sites controls literally fell apart in 479480, not a single lesserdevotee ever was allowed to donate a single image at the burgeon-ing site; and these restrictions extended even to the years betweenthe beginning of 469 and the end of 474, when at rst most andthen nally all of the caves were temporarily abandoned.30

This elitist focus of Vakataka patronage is the more surprisingwhen we realize that earlier Buddhist undertakings, including Ajantasown Hinayana nucleus, were generally community eorts. As the

27 See below. Cave 26 was almost entirely reamed out before any such detailingwas done. This may not have been true of every planner; for instance, the carv-ings on Cave 7s faade may have been done shortly after it was exposed, althoughthe intended medallions on the ceilings of the porch projections were merely reserved,and therefore never got completed due to the subsequent interruption of work. One,at the far right, was merely outlined in red, while another, in the next bay, forwhich projecting stone had been reserved, also never got completed.

28 See Bakker plates XXXVI A and B; XXXV C.29 For familial connections with Harisenas family, see Volume I, In Defense of

Dandin.30 See Time Chart. Clearly all of the temporarily abandoned caves were pro-

tected in some way during the Recession and Hiatus.


rst great campaign to emulate, and soon dramatically surpass, theimpressive cave sites of the Satavahana period, and reecting soclearly in its paintings the very world of leisure and auence thatenabled (its) donors to be magnanimous! (Bakker 1997, 44) Ajantadrew its energy as much from pride and the desire for prestige asfrom mere piety. Nonetheless the absolute compulsiveness with whichthese rich donors rushed to complete their images whenever omi-nous clouds hung over the ravine, and the high praise that theyshared with the Buddha in their self-aggrandizing inscriptions, showhow deeply their faith functioned as a motive force.

At the same time the piety of Ajantas patrons was rivalled by thefervor of their desire to outdo both the past and the presenta goalclearly evident in the constant changes in the sites development.These insistent transformations soon led the site from the didentproductions of the early 460s toward the demanding esthetic withwhich both the patrons and the artists had been for years familiar,even if they had to overcome the resistance of the unfamiliar rockbefore it could be properly expressed. It took only a few years ofsocial pressurefor everyone was watching everyone else, concernedboth about pay and prestigeto renounce the embarrassing sim-plicity of the early pillars, doorways, and decorative details, and tostraighten out the crooked walls and the angled ceilings of the ear-lier caves.31 You can tell the earlier caves from the later merely byassessing the alignments of their walls and ceilings and cells. Compare,for instance, Cave 17, excavated in the 460s, with Cave 23, exca-vated (except for certain faade features) in the 470s.

These continually elaborated changes in style and iconographyand plans and functional features would quickly turn the more devel-oped caves under development in the late 460s and the 470s intoforms which, clothed in the brilliance of Indras crown . . . cannotbe even imagined by little-souled men. (Cave 17 inscription, verse25) Created to emulate nothing less than the palaces of the lord ofgods (Cave 16 inscription, verse 27), these memorials on the moun-tains (Cave 26 inscription, verse 8) gradually came to look moreand more like the surely luxurious structural mansions of the citieswhich many of the artisans must have been building all their lives.

31 See Volume I, Cave 4, for discussion of why ceilings and walls are generallyout of alignment in the earlier caves. Also Volume V: Cave 4.


Sadly, these beautiful residences are preserved now only in Ajantaspaintings, as in the startlingly convincing palace scenes in Cave 1sJanaka Jataka, where the elegance and authority of what once musthave been can still be seen today. Indeed, in the whole book ofIndias ancient architectural history, only this chapter from the worldof the Vakatakas has been so remarkably preserved, in living color,as it were.

This dramatic evolution from the archaic provincialism of thesite in the early 460s to the highly developed urbanity of a decadehencethe seeming equivalent of which would take so many cen-turies to achieve in the world of Greececan be explained by thefact that the desired stylistic end was already known from the start.Instead of being slowly searched out, it in fact already existed in theresplendent world to which Ajantas courtly patrons must havebelonged. From the moment the rst Vakataka excavators didentlyfaced Ajantas recalcitrant scarp, the urgent process of bringing thesite up-to-date began, urged on by the highly productive rivalriesbetween the artists eager for the best jobs (and, during times of stress,for any jobs at all), and between their various contending sponsors,eager to best honor the Buddha and to outdo their neighbors withthe supremacy of their achievement.

The insistent progression toward what in fact was already knowncan be followed in the treatment of doorways, windows, shrines, andcells, but is most easily revealed in a comparison of the plain octag-onal (Hinayana) pillars of the early 460s, with the elaborate andmulti-facetted high-based columns being carved throughout the siteand at Aurangabad a mere half-decade later. Thus little by little,having started so willingly under the inuence and aid of the past,within a mere half decade they were already emulating in stoneIndras celestial mansionsan achievement attained with what mayseem to be a staggering swiftness, but in reality was only a conquestof what, as the heirs to a long tradition, they already knew. Facedwith Ajantas soaring clis, and the potential palaces within, theirtask was ultimately one of translation.

However, there is surely no reason to fault what the artists andthe patrons achieved with such rapidity. When Bakker (1997, 89)suggests that the quality of (the Ajanta) sculptures on the wholedoes not reach the sophisticated subtlety and the fresh originality ofthat of the Eastern Vakataka kingdom it is hard to believe that heis thinking of the larger body of Ajantas sculptured forms, whether


images or ornamental motifs. He seems here to have been inuencedby James Harle, who (had) a very marvelous store of merit (Cave17 inscription verse 22) yet at one point lost a bit of it by callingthe Ajanta Buddha images generally uninspired and rather lumpy.(Harle 1974, 24)32

The truth is that the lack of sophisticated subtletyif not offresh originalityfar more appropriately characterizes the sculp-tures of the rst half of the fth century in the Eastern Vakatakadomains. These are the sculptures that comprise so much of the fas-cinating material in Bakkers book. With their compacted clumsiness,and with heads sometimes sprouting from their legs and shouldersand competing for position elsewhere, they do indeed embody aremarkable energy which seems eager for a more subtle expression.Just as Giottos ponderosities in some ways help to explain evenBotticelli, the forms from Mandhal undoubtedly provide some of theinner forces that are ultimately transformed into Ajantas renements.

On the other hand, the characterization of the Mandhal gures,with their links to a folk tradition, can by no means apply to theunique gure of Siva (or a Saivite attendant) from Mansar, whichBakker believes can best be dated to the reign of Pravarasena II andassigns to the rst half of the fth century.33 This is much too earlyin my opinion, since the beautifully modelled form has such strik-ing connections with various yaksha images of the 470s at Ajanta,as Bakker himself notes. Like Jamkhedkar (1991, 90 and 91, Fig. 10)I would date both this image, along with other superb gures fromMansar, and the stylistically related Bhararaksha gures fromNagardhan to 470 or slightly later.34 Genetically linked to its pastit can be said to be somewhat in the tradition of the Mandhalimages ( Jamkhedkar 1991, 90)it has, as if preparing for the future,equally telling connections with some of the ne carvings from theUccakalpa sites (such as Nachna Kuthara and Bhumara) dating some-what after Ajanta, as the early heirs of Ajantas fall.35 And still fur-

32 Harle 1974, p. 24. It is true that a number of the shrine Buddhas are overlyconstricted by the outgrown limitations of the central image block, but Bakkerslarger generalization regarding all of the sculptures at the site is surely unjustied.

33 Bakker 1997, Plate XXXVII. Mirashi assigns Pravarasena IIs reign to 42257;I suggest c. 410c. 445 (Spink, 1992, A Revised Vakataka Chronology. However,I date the image later, as explained herein.

34 Bakker 1997, Plate XXX, A and B.35 For such glorious residue from Ajantas fall, see discussion of the situation

of the Vakataka fall in Volume I.


ther in the potential so solidly embedded here one can sense theprofound authority of Early Kalacuri sculpture, soon to be realizedrst in the urgent forcefulness of the sculptures at Jogesvari andMandapeswar, and then in the ultimate renements of Elephanta.Beyond that, the royal line continues in the beautiful but alreadydoomed renements of Cave 21 at Ellora, after which and aroundwhich the promise is already wearing out, as if waiting for the harshinfusion of a new strengththe taut energywhich will after anothercentury be part of the regions conquest from the south.

The startlingly subtle image from Mansar, one of the still too littleknown masterworks from that staggering site, instead of suggestingthe impact of the East upon the West, as Bakker contends, suggestsjust the opposite, in my view. If my proposed new Vakataka chronol-ogy is correct, Harisena, now having taken over the Nandivardanabranch (instead of vice versa) and having unied the empire, cameinto control of the Eastern Vakataka domains sometime in the mid-470s. The same course of inuence from West to Eastrather thanvice versacan explain the stylistic connections between many East-ern Vakataka forms at Ramagiri and the far more burgeoning devel-opments at Ajanta.36 Just as artists almost surely sought out work at Bagh, Dharasiva, and elsewhere during Ajantas Recession andHiatus (469 through 474: see Time Chart), they surely went to theNandivardana region as well.

Bakker and I dier fundamentally in our interpretation of theimplications of the shift of cultural energy between the two parts ofthe empire. Whereas he and most other scholars see the easternbranch of the dynasty as ultimately transcendent and continuing inpower until the end of the fth century, I have tried to show thatby the mid 470s the holdings of Prithivisena II, the last king of theNandivardana branch, had already been taken over (through eitherinheritance or conquest) by the Vatsagulma branch of the long-divided family. It was Harisenanot Prithivisena IIwho nallyunied the once bifurcated Vakataka dynasty, and was responsiblefor its greatest political, military, and artistic achievements. Both theVisrutacarita, which sees Harisena (= Punyavarman) as the single rulerover vast domains, and the Ajanta Cave 16 inscription (verse 18)which makes it clear that he controlled central India from sea to

36 Bakker Plates XXXV A and B; XXXVI A and B, for Ramagiri.


sea support this point.37 Thus it was under Harisena himself, in the460s and 470s that the Vakataka imperium achieved its greatestheightsheights from which, only a few years after Harisenas death,it was to so tragically and nally fall.

The implications of this view are as immense as they are con-ventionally rejected. For if this is true, Harisena was the last greatruler of Indias so-called Golden Age. Directly or indirectly, he con-trolled the whole of central India from sea to sea, at a time whenthe vaunted Guptas were already in decline. And Ajanta, in itscrescendo of activity, not only mirrors the character of that age atthe highest point of its development, but nally and suddenly mir-rors its destruction as well, recording, in the stone, the traumaticshock of Harisenas great empires shattering during the reign of hisinept successor, Sarvasena III (c. 478c. 483).38 At this point, pacethe history books, India entered an age of darkness from which itdid not recover until the various lesser kings whom Harisena hadonce dominated were able to climb up out of the ruins of theVakataka imperium and make their own new way into the world.

To those who would read Vakataka history in the old way, thissounds oensively speculative, as if it could meet with the approvalof Dandin alone. But the evidence of art history demands this newview of Vakataka history and equally explains the remarkable char-acter and course of Ajantas patronage. In fact I would contend thatart history turns out to be the one ultimately necessary key to ourunderstanding of late Vakataka developments. It is not art history,with its plethora of data, but the unrooted extension of convention,which is speculative. Surprising as it may seem, one can not possi-bly make sense of the political and military situation during and justafter Harisenas reign, nor properly test the validity of the newVakataka chronology, without applying the disciplines of art historical

37 Although past scholars have generally supplied the missing verb in the inscrip-tion as conquered, Bakker (1997, 35) suggests stands above; this seems prefer-able, since Harisena must have developed his empire through marriage and politicalalliances or inheritance (of Prithivisenas domains) as well as war. However, it shouldbe noted that Bakker uses stands above in the sense of outdoing or putting toshame all the surrounding kings, rather than, as I would suggest, transcendingthem and quite literally incorporating (at least to some degree) their authority byhis own. For fuller discussion, see Volume I.

38 See Spink 1991, pp. 7199; see also discussion of Sarvasenas accession andsituation in Volume I: Chapter 4 In Defense of Dandin.


analysis. To do this we must turn, leaving our preconceptions behind,very specically to the Vakataka donations at Ajanta, Bagh, Banoti,Ghatotkacha, Aurangabad Caves 1, 3, 4A, and Dharasiva; for it maywell be deep in those caves that the answers lie.

CHAPTER TWO

CAVE 26 AS AN INAUGURAL MONUMENT

Nowhere are such investigations more relevant than for reaching anunderstanding of the growth of the vast Cave 26 complex, com-prising the main caitya hall and its four wings. Bakker (1997, 41)rightly points out that the monk Buddhabhadras dedicatory inscrip-tion, with its fulsome praise of the Asmaka rulers and ministers,shows no word of Harisenas overlordship nor of any other rulerapart from that of Asmaka . . .. In his view (1997, 41), this stronglysuggests that this complex was excavated after Asmakas had gainedfull control over the region, i.e. after Harisenas death. It is withthis seemingly reasonable conclusion that I must so strongly disagree.

It is certainly true that the Vakatakas are signicantly omitted inthe Cave 26 record. I would also agree that the Asmakas had gainedfull control over the region at the time that it was written. Furthermore,as far as historical and epigraphical evidence goes (Bakker 1997,41) one can agree that it would be hard to locate the cave in rela-tion to the Vakataka developments at the site. But art history canand does locate Cave 26 convincingly and solidly in the mainstreamof the sites Vakataka patronage. It explicates, in its stage by stagedevelopment, the crucial (and often disruptive) role that the Asmakafeudatories played from the very start in the vagaries of Ajantaspatronage, which nally ended with the abandonment of the cave,the disruption of the site, and ultimately the destruction of the wholeempire. The Asmakas, as feudatories, were as involved in the sitesdevelopment as they were implicated in the ominous course of lateVakataka history; and Cave 26, along with its inscription, far fromdenying this involvement, fundamentally helps to explain it.

By the time that the Cave 26 dedicatory inscription was written,in mid-478, the great caitya hall complex was (not surprisingly) essen-tially done; and the Vakatakas were essentially done for. The inscrip-tion in eect predicts Ajantas collapsea collapse that was almostinevitable as soon as the aggressive Asmakas so rapidly asserted theircontrol over the site shortly after Harisenas death in 477. Indeed, itis in part the evidence of this inscription and its revealing art historical

cave 26 as an inaugural monument 23

context that limits the sites patronage activity to less than twentyyears, for it denes Ajantas doom, by writingbetween the linesthat the great emperor was dead. The evidence that clearly provesthat Cave 26 was not created after Harisenas death, as Bakkerhypothesizes, but was in fact one of the sites initial undertakings(being started in 462), is by and large obscured by its lavish over-lay of late (475478) decoration and by the many changes in planmade during the course of its excavation.

But the evidence is surely there, for the cover-up is by no meanstotal. We can list some of the complexs indisputably early featureshere to show that it was one of Ajantas inaugural undertakings. Infact, with its four wings, it was the most ambitious one of all.Furthermore, as will be shown later, all of the other caves in thesites western extremity (i.e. Caves 2124, and 28) were also theproduct of Asmaka patronage; and, perhaps remarkably, their devel-opment appears to have been under the control of Buddhabhadrahimself, as I shall try to show in this chapter and the next.

It should be mentioned that these early features in the Cave 26complex are often obscured, masked by changes made to keep upwith the times, or to improve functional features, or to consciouslycover forms that were troublingly obsolete.1 As in archaeology, oneoften has to dig down through various layers of change, to uncoverthe past; and whereas something of the process and of the ndscan be detailed here, in the end one must go to the site for a num-ber of seasons, in order to properly excavate the caves.

After providing evidence here that the Cave 26 complex is, atthe core one of the very earliest undertakings at Ajanta, I shall thenprovide a resume of its development (and will explain its long periodof abandonment) during the troubled course of the sites history. Itmight only be mentioned that the whole site follows a similar pat-tern, wherein the development of the caves is controlled by politi-cal realities. In nearly every case (other than the very late and oftenminor undertakings Caves 3, 14, 22, 23A, 24A, and 28) all of thevarious caves were started very early, and then were either totally

1 Most earlier caves were blocked out before any decoration started. The Cave26 complex is a striking example; when it was rst abandoned in 468, it had nota bit of imagery or decorative carving on it. In striking contrast, note the moresophisticated development of later caves, such as Cave 24.

24 volume iia, chapter two

interrupted (i.e., Caves 1720; 29) or (in the case of all of the othercaves) suered a period of abandonment before work on them waseventually taken up again, in their patrons ultimately abortive attemptsto complete them. (See Time Chart in Volume IV)

Thus Ajantas Vakataka development, throughout what might, hadtradition not wrought its havoc, be called its lengthy history of nearlytwenty years (instead of half that time), is a history of crises. Andthese crises are particularly revealing; for it is at these points in timethat we nd breaks in development, as artists and planners andpatrons respond dierently to suddenly changing times. When a cityis burned down, or ooded, or bombed, the pre-crisis evidence andthe post-crisis evidence tell a story that hardly need be written down.And it is the same with a site such as Ajanta, or indeed the otherrelated Vakataka sites. When the Asmakas are ejected from the site,and then when they in turn eject the local king; or when the greatemperor Harisena, a partial incarnation, as it were, of the God ofJustice (Kale 1966, 349) sent the empire into shock by his untimelydeath; or when, as a consequence, the region is direly threatened;and then when even the helter-skelter patronage at the dying sitecollapses: these events leave their marks deep in the stone, and thosemarks can eventually be seen to be the letters and the words withwhich a dramatic story can be eventually composed. And these eventsare written not onlyor even so muchin the grandest of the caves;they are even more poignantly exposed in the tragic mini-historiesof those caves which were abruptly abandoned, or had to struggle(leaving evidence of their dicult passage) over the aws and gapsin the sites development to nally reach their patrons obsessivegoal.2 This goal was of course, and particularly by this time, theachievement of the merit that would be theirs, if only, in the devel-oping storm, they could get their shrine Buddhas brought, in boththe inner and the outer sense, to realization.

2 Such gaps or traumas can aect the course of a scholars research also, andoften the positive benet from such events is not immediately realized. A personalreection may be of interest. Some decades ago, at the unfortunate instance of asuggestion from a high foreign ocial in Delhi who thought the major paintedcaves were in jeopardy, those important caves (1, 2, 16, 17) were completely closedfor conservation, only visitors of some political signicance, along with sta (andoccasionally their families) being allowed in while the work, unhappily and insucientlyit turns out, proceeded. I, as a suciently young and interested scholar was banned


Cave 26s Early Features: An Analysis

If one agrees with Bakker (1997, 41) and many other scholars thatCave XXVI . . . may be of a later date than many or most of theother caves at the site, then my own view of the caves developmentmust be incorrect. And if this is incorrect, the whole Short Chro-nology must also be wrong. For this reason a proper understand-ing of Cave 26s developmenta justication of its position as aninaugural monument, as well one whose history ultimately speaks ofthe fall of empireis essential.

K. Khandalavala (1991, 102) expressed the case for the opposi-tion most decisively. The inscription in Cave 26 is a bed-rock ofany sensible chronology of the Ajanta Caves. . . . All the far-fetchedattempts, with no evidence in support, which have been made byDr. Spink to surmount what the inscription in Cave 26 states canonly be regarded as exercises in sheer futility for no amount of inge-nuity can ever displace the fact that Cave 26 could only have beenexcavated after the fall of the Vakataka dominion.

To better explain my own position, I will make a listing of someof the Cave 26 complexs never-noticed early forms and features,and link them with the sites overall development. As a result of thiswe will be better able to reconstruct the remarkable history of thewhole site which the complicated development of Buddhabhadrascave so tantalizingly mirrors.

from the caves for that same period, despite the strenuous eorts of the then DGAM.N. Deshpande, widely known as the doyen of cave studies, to grant me access.For nearly ten years, his pleas to higher authorities, which one wishes to believepained him, were of no avail, despite the innocence, other than scholarly, of myinvolvements. Thus, since I of course had no intention of giving up my work atthe site, I was forced to study the many unnished and broken cavesthe veryones which P. Stern, in his faulted Les Colonnes indiennes dAjanta . . . excludesfrom consideration because they are broken and unnished. But forced to under-take my work in these less studied areas, I eventually discovered that hereandnot in the more fully nished caveswas where the most revealing secrets lay. Thiswas because, due to their variously abandoned states, they far more clearly showedthe telling breaks in the sites development; and it was by an analysis of these breaksand their signicance that I rst constructed the stage by stage pattern of devel-opment which underlies my conclusions still today. I hardly think that thanks arenecessary for the productive pain which I occasioned during this period, but it isintriguing, if ironic, that, both sadly and happily, my presumably necessary, or atleast instructive, exclusion had such a positive conclusion.


Astylar Plans

The astylar plan of Cave 26s four wings is revealing as a notablyearly feature, showing the early dependence on Hinayana prototypesat the site.3 As in most Hinayana caves, each wings central hall, asplanned, was to be surrounded by monks cells on left, rear, andright; but this notably simple arrangement, representing the plan ofall four of Cave 26s wings, undergoes major later adjustments inall cases, as conventions at the site develop. Cave 8 and Cave 15,both inaugural excavations, are also astylar, whereas both Cave Lower6 and Cave 11 were similarly planned, but were found to requireinterior pillars to hold up their ceilings.4

Porch End Development

All of Cave 26s wings were conceived without porch end cells, aretardatory (Hinayana) feature found throughout the site but wiselyabandoned after 465 in order to add cells at these convenient loca-tions in all caves. Beginning in 465 and 466, simple single cells werecut wherever possible in these previously wasted areas, rapidlybecoming conventional (and invariable) features for the next coupleof years. In 467 cells with pillared vestibules come (also invariably)into fashion, replacing the single cells whenever possible, just as thesingle cells, whenever possible, had been cut into the still-earlier plainend walls of the porches.

The surprising adjustments made to Cave 26s left wing, underthe dictates of rapidly changing taste, show this process of transfor-mation. Originally this wing must have been planned to preciselymirror its counterpart on the right, preserving the careful and con-ventional symmetry of the whole complexs faade arrangement. (Seeoriginal plan of Cave 26 complex) But as we can see in the earlier-

3 In fact Hinayana viharas in general are astylar. M.K. Dhavalikar has suggestedthat Ajantas Cave 15 is early fth century for this reason. Dhavalikar (1981, 136;also 197071). However, the wings of the Cave 26 complex are also astylar and,along with the plans of Cave 8 and Cave 20, show that this feature is commonlyfound early in Harisenas reign, in the 460s.

4 For the added pillars in Caves Lower 6 and 11, see Volume V, in the dis-cussion of those caves.


dened right wing, this involved a porch too shallow to allow theaddition of porch-end cells, when (in 466, or even 465) such cellscame into fashion; and in any case, the excavators had already cuta cistern at the right end of the right wings little porch. Such apositioning for the cistern would never have been done if a cell hadbeen intended for this location. The cistern could easily have beencut in some area where it would not take up this now-valuable space.

By the same token, there was no possibility of putting a cell atthe other (left) end, an area where the needs of the abutting porchend of the main hall surely had priority. Single cells must have beenplanned for (and perhaps even begun in) these porch-ends of themain cave by 465 or 466, but if so they were being converted tothe more complex and suddenly fashionable pillared cell-complexesby 467 or 468. At this point, like the rest of the cave, they wouldhave been only roughly dened by the time of the Recession, beingcompleted with their elaborate fronts only in the later phase of work on the cave, starting in 475. (See Dening Features chart inVolume IV.)

Since the lower left wing of the caitya complex had not beenstarted as early as the right, in 465 or 466 the caitya complexs plan-ners, to accord with new conventions, made the decision to shift thestill not fully revealed court of the caitya hall farther to the left (west).This was done in order to make room in the set-back porch of theleft wing (Cave 26LW) for a single cell at either endspecically(even if surprisingly) to accord with the alwaysdemanding new con-ventions. But in 467/468 the still merely (if at all)roughed-out rightporch cell of Cave 26LW was convertedagain to accord withchanging conventionsinto the new cell complex, in which a pil-lared vestibule fronts the residence cell behind.5 That the vestibuleof the complex pillared cell at the right of the porch of Cave 26LWwas previously under excavation as a cell is revealed by its unchar-acteristically deep dimensions, and by the equally uncharacteristicthinness of the fronting pillars. These thin pillars, now long sincebroken away, were, in fact, cut out of the previously dened frontwall of the now-converted cell! The present cement restorations aresomewhat thicker than they should be; this was an understandablemistake on the part of the restorers, who did not understand that

5 The left porch cell was apparently cut, but has since fallen into the ravine.


they had originally been shaped from the typically thin early wallsof the simple cell, and did not look to the still-intact pilasters forguidance.

Such a fundamental change, from plain porch end, to the addi-tion of single cells, and then to pillared cell complexes, all takingplace during the course of a few years of excavation, provides insightinto the exibility of the planners as well as the force of the rule oftaste, or of convention, at the site. There are a number of tellingexamples showing that such changes, if possible, were considered asconventional. The complex cell L1 in Cave Upper 6, and the cellat porch right (PR) in Cave 27 were both started as simple cellsbefore the Recession (in fact prior to 467), and then upgraded tocomplex types after 475, when work on them nally continued.Indeed, the present porch-end complexes of Cave 26 itself were rstintended as single cells; otherwise Cell PR of the Left Wing couldnever have has its present format. Indeed, the present vestibules (cell-like in their dimensions rather than more conventionally shallow)must have been reserved while more crucial work went on in thegreat cave. Then, when they were nally cut, they were turned intopillared complexes. The very thickness of their fronting pillars wouldseem to prove that earlier penetration was never made, and thatwhat we see today are pillars created after 475, and (both on theleft and on the right) extra cells now added for extra residencespace, at that same time. The door ttings, at this late date, wouldof course be D mode, even though throughout most of the wingsthe cells were penetrated earlier, in the A mode, and then converted,after 475, to the practical later type of tting.

The conversion of Cell R1 in Cave Upper 6 presents a specialcase since, due to lack of time, it remained in its pre-Recession stateuntil an intrusive donor took it over in 479, and turned it into hisown personal shrine. Although surely, as in its counterpart Cell L1,its front (originally intended for a simple doorway to a single cell)had been widened in preparation for the denition of the new (nevercompleted) vestibule pillars, the intruder merely used this space asan unconventionally wide entrance to his new shrine, at the sametime cutting his sacred image into the back wall of the previouslyintended residence cell; something very similar occurred in Cave 6sright rear cell, with its intrusive image; it had been started, proba-bly in 468, as another pillared complex, which never got completedby the original patron even after 475.


Although Caves 21 and 23 now have typically late pillared cellcomplexes in their porches, we can tell from the too-narrow area ofscarp between them that single cells must have been originally plannedhere when the two caves were rst laid outthe earlier (Cave 21)in c. 465 or c. 466.6 Hardly a year later, because the deeper cellcomplexes had just come into fashion, the space between Caves 23and 24 had to be very signicantly adjusteda phenomenon of greatinterest, to be discussed later.

Pillars Sitting Directly on Floor

Although the standard convention which was developed at Ajantain the Vakataka period was to support the pillars in the cave witha monolithic oor beam, over which visitors often trip today, it wouldappear that in the earliest excavations the pillars were set directlyon the oor, just as they were in the Hinayana caitya halls, whichhad such a strong inuence on the earliest Vakataka excavations.This primitive characteristicparalleling the cutting of the earliestcell doorways without planning for door-ttingsis notably evidentin the multipillared Cave Lower 6, even though at the upper levelthe pillars support the beams which in turn support the ceiling.Cave 11s pillars were also at rst planned without bases, even thoughcomplex bases were added when the old dormitory was convertedto a shrine.

Although the use of supporting oor beams came into use veryearly, and of course persisted, they are signicantly absent in thevery early Right Wing of the Cave 26 complex, which can be seenas yet another conrmation of the very early date at which it wasundertaken. By contrast, the Left Wing, under excavation a few yearslater, shows the beams under the porch pillars. The very distinctionbetween such features in the two wingsfor they were originallyplanned as parallel excavationsreveals the rapid developmentalchanges which the whole complex was undergoing.

6 It is very unlikely that Cave 21 was underway prior to 465, prior to the timewhen porch end cells were in vogue; but even if this was the case, the space require-ment between the two caves would have been the same because the area requiredby interior cells would be equivalent to that taken up by single cells at the porchends.


Cave 25, the upper right wing of the Cave 26 complex, was prob-ably started before 26RW (the lower right wing), for work wouldhave progressed generally downward (as well as from front to rear)when the whole complex was being exposed. It almost certainly wasplanned in the simplest manner, with its porch pillars connectingdirectly with the oor. Indeed, that is what we see today, even thoughin 478 much of the long abandoned and unnished porch was re-cut when, apparently, it was being expediently converted into ananomalous shrine area.7 We can see that the pillars, originally planned(and only roughed out) with heavy octagonal shafts, as would havebeen expected in the early 460s, were later being recut in a moremodern style when all Asmaka patronage was interrupted late in478. The oor between them, though surely still rough when theselate revisions started, probably had already been enough dened, inthe early phase of work on the cave, so that it was not practical tonow add a more up-to-date oor beam.

It seems to be true, also, from old photographs, that the front pil-lars of the two porticoes of Cave 7one of the earliest excavationsstartedhad no supporting monolithic base-beams; the fact thatthey were reconstructed recently in this way would seem to conrmthis, since reconstructions generally (or at least ideally) respect thetraces of old forms. However, more rearward pillars do have beams,probably because they were exposed slightly later, and thus conformto the new convention. It is not surprising for conventions to changevery rapidly.8

From Shrineless to Shrines

Just as in all of the caves at the site which were underway duringthe rst four years or so of activity, the very early four wings of theCave 26 complex were planned without any image shrines, againfollowing Hinayana precedents. These shrines, which in all of theearliest Vakataka caves such as these were always added features,consequently were not without problems of placement; we will later

7 See Volume I, Chapter 12.8 The porch end complexes of Cave 7perhaps the rst created at the site and

dating to 467also have pillars with no supporting beams beneath, even thoughsuch beams are conventional later on.


discuss the adjustments which had to be made to accommodate suchpreviously unplanned-for images.9 The reason that the shrine Buddhasdid not choose to take up residence at the site until 466, by whichtime nearly all of the sites Vakataka viharas were already under-way, is still unexplained, and should be a challenge to Buddhologists.10

If the viharas at a site as important as Ajanta were originally plannedwithout shrinesthat is, as mere dormitoriesdoes this reect thecontemporary situation in structural viharas elsewhere as well? Wasthis crucial feature in fact one of Ajantas innovations?

Although shrines were planned for by 466, at least in the majorcaves, it seems likely that the rst separate Buddha images were notnished and dedicated until sometime during the rst half of 469.All four of the viharas in which we nd the earliest Buddhas at thesiteCaves 6L, 7, 11, 15had been penetrated to varying degreesin 468. This surely reected the revolutionary changes taking placein the important Vakataka caitya halls, Cave 19 and 26, where theidea of fronting the stupas with images had already developed, eventhough they would not actually be carved until slightly later thanthe Buddha images in the viharas Lower 6, 7, 11, and 15, all nishedand dedicated in early 469. The image in Cave 19 was probablyrevealed by 469, whereas, its carving interrupted by the Recession,the Cave 26 image was not revealed until 477. The fact that a stupa(soon abandoned) was started before the image was carved in Cave11, would further seem to reect the transition seen in the caityahalls.11

The reason that we know that shrine images had to be planned(at least in the major caves) much earlier than they were actuallycarved is from the widening of the intercolumniation between thecentral pillars of the front aisle, to provide an axial focus for theshrine. However, such widening never appears in the earliest exca-vations which have pillared halls, namely Caves 2, Lower 6 andUpper 6. This shows that, like the various astylar halls, they wereoriginally planned (always before 466) with no thought at all ofshrines.

9 See Volume I, Chapter 12 for comments on the four wings of the Cave 26complex.

10 Schopen (1990, 181217) discusses the Buddha as resident in the caves, butdoes not explain why the conception did not appear at the site until about 466.

11 See Volume V: Cave 11; the abandoned stupa appears to have begun, or atleast conceived, in 467.


Since many of Ajantas earliest excavations (8, 15, 20, 25, 26RW,26LW, 27) are astylar, they provide no evidence in this regard, whilethe pillars in the small Cave 11 were created as ceiling-supportssome years before the shrine was even conceived. Although muchof Cave U6s interior was excavated after 475, the front aisle colon-nade was apparently dened (thus setting the pattern) very early.The nearby cells, L1 and R1, which would have been started some-what later, were almost certainly in the A mode and thus were notdened after 467. The present windows, for instance, like those inCave 5, are late (unnished) conversions from very early verticaltypes like those of Caves 15 and 20.

The idea of adding a shrine to the viharas was probably one ofthe many concepts transmitted to Ajanta from Bagh, where the fri-able nature of the relatively low sandstone scarp apparently disal-lowed the making of the expected caitya hall. Therefore, althoughthe earliest of the Bagh viharas, Cave 2, (which has no axial focus)had apparently been started as a mere dormitory, it could be con-verted to a caitya hall by the addition of a shrine at the rear.And being the very rst such converted vihara, at least in centralIndia, it is hardly surprising that the resident caitya was a stupa,not a Buddha image, following the long established convention forhalls of worship.12

Shortly thereafter, at Ajanta, when the idea of adding shrinesprobably at rst only to the largest cavesdeveloped, it was prob-ably also stupas, not Buddhas, that were originally intended for them.Of course, in these privileged major caves, such as 4, 16, 17, andthe small (but royal!) Cave 20, the original plans (without shrines)had to be adjusted to provide for this important (indeed revolu-tionary) new feature. In these larger and more important caves thenew conception must have developed by no later than 466, beforethe pillars of the front aisle had been started, or while they werestill so rough that their central intercolumniations could be widened,as new conventions required.13 Only Cave 1, begun much later (c. 466) could be excavated as planned, at least in this regard. Butit too may well have been rst conceived with a stupa as its focus,

12 See Spink 197677, 5382. Therein I considered Bagh Cave 4 as earlier thanCave 2, which I now feel is not correct. See also Volume I, Chapter 3.

13 This is discussed at length in Volume V: Caves 16, 17, etc.


a year or more before the idea of the image took precedence. Indeed,its image was not even started until 475, although its shrine hadbeen penetrated earlier.

In the larger and more important viharas, such as Caves 1, 16,17, and 4, the plans would have been adjusted well before the exca-vators had reached the shrine area; and when, in the course of exca-vation, they were starting on the shrine itself, Buddhas instead ofstupas had become the caitya of choice. So their past has beenobscured: indeed, were it not for the obsolete central blocks fromwhich all of the earliest ones were cut (or were intended to be cut)we would not know, seeing them today, that their Buddha shrinesreplace once-intended stupa shrines, and that the stupa shrinesthemselves would have been cut where residence cells were onceintended. King Upendraguptas beautiful Cave 20 seems to be thesingle exception to the rule that the smaller excavations were some-what slower in developing the shrine concept. Probably because ofits priority royal patronage, Cave 20 was the rst small cave tohave a shrine added to its originally simpler conception. We knowthat its shrine area (if not the shrine proper) was planned very earlyin 466 or 467because its shrine antechamber projects out into theastylar hall, in order to support a bad ceiling aw; and the hall con-tains many very early features, the most striking being its primi-tive doorways, doorsteps, and door ttings. However, its position ofprivilege did not alleviate its obvious nancial diculties as theRecession progressed, for both its image and its shrine doorway hadto be rushed to a most expedient completion when Upendraguptastroubles, increasing through 469 and 470, caused him to have toabandon his connection with the site late in 471.14

In lesser caves (other than the royal Cave 20) it would appearthat it took two or three years for the idea of adding a shrine tohave its impact, for even though in the smaller caves it took lesstime to reach the rear of the cave in the course of excavation, shrineswere never begun in them either until late 467 or (more probably)468. Perhaps at rst, the addition of shrines was reserved only forthe more prestigious patrons and their caves; although the idea soonbecame universal.

14 See Volume V: Cave 20 for the anomalous projecting antechamber.


The earliest cave in which an image was actually completed maybe Cave 11, where the shrine was expediently made from a con-verted cell; signicantly, it contains a monolithic stupa fronted by aBuddha image. However, the Buddha image, which was probablynot even conceived when the stupa was started, soon become thesole focus, and the stupa is paid no heed whatsoever. Not only doesthe Buddha block the view of the stupathe halo, in particular,seems to have been made particularly large just for this purposebut the Buddha alone was worked on and carefully nished in 469.It seems likely that the combined stupa/Buddha, before the victoryof the latter, had been conceived following the lead of the inuentialplan for a similar combination in Buddhabhadras great hall of wor-ship where, for the rst time in any such caitya hall, an image wasto front the stupa.15

Of course, for reasons of ritual and tradition, one could hardlyabandon the stupas in the caitya halls, as was so abruptly donewith the stupa in Cave 11. However, the developments both in theshrine of Cave 11 and in the purely Buddha oriented shrines ofCaves Lower 6, 7, and 15 reveal the irrelevance of the stupa in thevihara setting. Indeed, the same could be said for the ritual of cir-cumambulation, for in spite of what scholars generally assume, thereis not the slightest reason to think that the ritual of circumambula-tion was ever practiced or even intended in any of the Vakatakasviharas.16

With the interest in stupas as a focus gone, the patrons and plan-ners of the viharas were increasingly burdened by the outmodedconvention of the centralized shrine format. This made sense,indeed was necessary, when the shrines were conceived for stupas,but when the stupa concept (never truly realized) was rejected infavor of Buddha images the old format imposed unacceptable restraints,particularly as the Buddha groupings expanded. This is painfully evident in the constrictions imposed by the old format on the developed image groups of Caves 1 and 4.17 The ultimate goal was to spread out the image group against the rear wall, as in the

15 For the reasons that the shrine of Cave 11 was probably the rst and earli-est started, but at rst with a stupa, see Volume V: Cave 11.

16 For circumambulation, see Volume I, Chapter 15.17 See Volume V: Cave 1, Cave 4.


anomalous Cave 20; but this ideal was not legitimately achieveduntil after 475, in the shrines of Caves 2, 21, Upper 6, and 26L.18

At the same time, it is logical to believe, that just as in Bagh Cave2 and probably Bagh Cave 4, the patrons and planners rst con-ceived of their shrines as having stupas only. Indeed, it may well bethe case that when the Vakataka caitya halls Caves 26 and 19 wererst plannedvery much under the inuence of the Hinayana Caves10 and 9 respectivelythat they too were going to have stupas aloneas their focus, and that the idea of fronting Buddha images onlydeveloped, both at Ajanta and Bagh, as excitement mounted at thesites, and as new inuences, including the desire for image worship,ooded in. As already noted, the rst vihara shrine undertaken, inCave 11, with its unnished stupa fronted by the completed Buddhaimage, apparently appropriated the same already-conceived combi-nation. But the stupa was abandoned, barely half nished, probablyby 468, and by early 469, when it was hastily dedicated, its Buddhaimage, like those in Caves Lower 6, 7, and 15, had become the solefocus.

The insistent switch to Buddha images surely reects a trend towardimage worship throughout the subcontinent in this general period.But in the case of these rst shrines, forced by circumstance (in early469) to be expediently rushed to the point where their Buddha imagescould be dedicated, time and money may have played some part inthe decision to focus on the images alone.

However, stupas had long been recommended by tradition andconvention; and the fact that, in early 469, the Prime Minister puta relief stupa as a kind of substitute altar in his Ghatotkacha vihara,when he felt constrained to discontinue the excavation of the stillincomplete cave in 468, suggests that the preferences were still some-what in balance, as they so clearly were in the shrine of Cave 11.But by 468, the Buddha image had clearly won out, being the soleimage made in the shrines of Lower 6, 7, 15 (and now in Cave11)even though they surely had all been originally conceived forstupas, with a characteristic central placement.19

18 For Cave 20 see Ch. 13.19 The seated Buddha on the Ghatotkacha stupa is a later intrusion, while the

supporting yaksha below was carved along with the stupa.


Similarly, as we shall see, the development of the plans in thelower wings of Cave 26 proves that by 468 (after which these Asmakacaves had to be abandoned for the next six years) these viharas werebeing adjusted (26RW) or extended (26LW) to include shrines.Needless to say these cavesoriginally mere dormitorieshad notbeen planned with shrines in mind, a fact which tested the archi-tects creativity now that he had to add them. These newly intro-duced shrines may well have been conceived for stupas in 468, butat least in Cave 26RW and 26LW the places were still in the processof being prepared for them, when the Asmakas were ousted.20 Then,by the time these lower wings were nally completed, in 478, theywere provided with typically developed Buddha images, just as wewould expect. Since both of the complexs upper wings (Cave 25and 27) had probably been at least temporarily abandoned in 466,it is hardly surprising that early shrines were never started in them;the shrines later planned (in 478, when time was quickly runningout) were, understandably, never nished.

The discussion above only tells us how and when shrines, andultimately shrine images, rst made their appearance; it does not tellus why this remarkable development took place or why the eighteenmere dormitories already underway before 466, were suddenly turnedinto places of worship.21 Judging from the spectrum of the sitesdierent painting styles, as well as its rich body of sculptural andarchitectural motifs, Ajanta drew its forms and features, and its ideas,from many dierent parts of the Indian subcontinent. In this regardthe established Buddhist sites in both the northwest and the south-east deserve special attention, both being linked to Ajanta by theroutes of trade as well as by their strong Buddhist traditions. Forinstance, the bhadrasana pose of the Buddha, perhaps unknown incentral India and wider regions beyond until its appearance at Ajanta,appears much earlier in Andhradesa, while the frontal projectionframing the Buddh

Documents

Spink, Walter M. - Ajanta. History and Development. Vol. II